Fair value betting lines

[1ejp]

I am trying to get a handle on how to attach value to contending horses.

For example, is there a formula to follow on whether a favorite is 6-5 or 2-1?

If it's a 10 horse field and I throw out 5 horses as non-contenders how do I know where to start with the horse I like best? Is it what I am willing to accept as "fair" odds? If I make my top pick 2-1, where do I go from there?

[PICSIX]

Search Steve Fierro on this site. He has created templates that do a good job of creating useful odds lines for your contenders. :ThmbUp:

[ubercapper]

I am trying to get a handle on how to attach value to contending horses.

For example, is there a formula to follow on whether a favorite is 6-5 or 2-1?

If it's a 10 horse field and I throw out 5 horses as non-contenders how do I know where to start with the horse I like best? Is it what I am willing to accept as "fair" odds? If I make my top pick 2-1, where do I go from there?

The late Dick Mitchell taught me this - Ask yourself the question "If this race were run 100 times how many times (or what percentage of the time) would X horse win?" If the answer is 33% then set 2 to 1 odds, if it's 20% then set 4 to 1, etc.. After you have odds for all your contenders, try to get the total to 80%. It's pretty subjective so don't worry if it's not perfect.

Dick also said that in many situations this simple rule set will suffice -
If you have four contenders left after elimination, make them all 4 to 1. If you have five contenders, make them all 5 to 1. If you have three contenders make them all 3 to 1. If you have two, make them both 2 to 1. In most cases (except the 2 to 1 case) you have about an 80% odds line, which gives 20% of the probability to the horses you eliminated.

Good Luck.

[thaskalos]

The late Dick Mitchell taught me this - Ask yourself the question "If this race were run 100 times how many times (or what percentage of the time) would X horse win?" If the answer is 33% then set 2 to 1 odds, if it's 20% then set 4 to 1, etc.. After you have odds for all your contenders, try to get the total to 80%. It's pretty subjective so don't worry if it's not perfect.

Dick also said that in many situations this simple rule set will suffice -
If you have four contenders left after elimination, make them all 4 to 1. If you have five contenders, make them all 5 to 1. If you have three contenders make them all 3 to 1. If you have two, make them both 2 to 1. In most cases (except the 2 to 1 case) you have about an 80% odds line, which gives 20% of the probability to the horses you eliminated.

Good Luck.

Wouldn't all these contenders have to be of equal ABILITY in order for them to be listed at the same odds? When was the last time any of us encountered a race where the five contenders were all of equal ability? Don't some contenders always look better than others? You would really feel comfortable assigning your #1 contender and your #5 contender the same odds of winning the race?

I understand if Mitchell said that "a few" situations may warrant this sort of odds-making...but "many situations? I would have to disagree...

IMO...it takes a certain amount of PRECISION to overcome today's onerous takeouts.

[therussmeister]

The late Dick Mitchell taught me this - Ask yourself the question "If this race were run 100 times how many times (or what percentage of the time) would X horse win?" If the answer is 33% then set 2 to 1 odds, if it's 20% then set 4 to 1, etc.. After you have odds for all your contenders, try to get the total to 80%. It's pretty subjective so don't worry if it's not perfect.

Dick also said that in many situations this simple rule set will suffice -
If you have four contenders left after elimination, make them all 4 to 1. If you have five contenders, make them all 5 to 1. If you have three contenders make them all 3 to 1. If you have two, make them both 2 to 1. In most cases (except the 2 to 1 case) you have about an 80% odds line, which gives 20% of the probability to the horses you eliminated.

Good Luck.
I think it is a waste of time to set odds for each individual horse and then convert to a percent to see if you get a total of 80%. You should instead just set a percentage for all your horses, make adjustments so your totals add up to 80%, then convert to odds. You should also keep records to determine if an 80% odds line is appropriate for your handicapping ability.

[Redboard]

If it's a 10 horse field and I throw out 5 horses as non-contenders how do I know where to start with the horse I like best? Is it what I am willing to accept as "fair" odds? If I make my top pick 2-1, where do I go from there?

If you narrowed down the race to your five selections, look at the tote board and take the best price. You’ll win less often, but you’ll do better in the long run. If they are all 4-1, then take who you think has the best chance to win.

[acorn54]

a very long time ago i came across something about making an odds line. the most profound thing about it, was setting the odds of ALL horses in the FIRST phase of an odds line at equal valure, ie; if a 12-horse field, then each horse starts with and initial probability of 8%, if 8-horses, then each starts with a 12.5% probability.
after that, if i remember correctly you give a weight to each of what you consider as the important factors of handicapping, such as class-speed- early
speed and so forth, and subtract or add as you deem neccessary to each horse in the field keeping the total percentage at the same 100 percent BEFORE take out tax.
hope this helps

[traynor]

I think it is a waste of time to set odds for each individual horse and then convert to a percent to see if you get a total of 80%. You should instead just set a percentage for all your horses, make adjustments so your totals add up to 80%, then convert to odds. You should also keep records to determine if an 80% odds line is appropriate for your handicapping ability.

I agree that the best odds lines will come from one's own records and a cold-blooded analysis of one's own handicapping skills. The relative probability of a given horse winning a given race is exactly that--a probability value relative to the probability values of all the other entries winning that particular race. An accurate assessment of the combinations of probabilites in the exact mix of horses in a given race is way beyond the capability of most (current commercial) software apps (despite what the developers may believe).

In plain English, that means obsessing over the "accuracy" of an odds line is generally misguided. Especially if that odds line is pulled out of the air with a subjective "opinion" with no real research to back it up.

[1ejp]

If you narrowed down the race to your five selections, look at the tote board and take the best price. You’ll win less often, but you’ll do better in the long run. If they are all 4-1, then take who you think has the best chance to win.

This is what I do now. I was wondering if there was a more detailed way of assigning value to each horse.

[Hoofless_Wonder]

Every time the topic of "fair odds line" comes up, it makes me think it's like searching for El Dorado or Shangri-La. It can be derived rather simply, or it can be as complex as landing a robot on the far side of the moon, but in the end the application of it is still extremely difficult. Late odds changes, horses not trying (or rather jockeys/trainers not trying), bad trips, DQs, etc. can all blow a hole in the world's most accurate "fair" odds line pretty quickly.

I rarely bet a horse I think has less than a 25 or 30 percent chance to win, though I prefer that percentage to be much higher. I rarely "dutch" races which a fair odds line implies can be done, 'cause I'm just not that good of handicapper.

What I can do though, is look at the odds of the horse or horses I like, and determine if it's above or below my strike rate for that type of race, and then determine if the race should be bet, or if it's a pass.

And then, like everybody else, I hope for the best....

[therussmeister]

...horses not trying (or rather jockeys/trainers not trying), bad trips, DQs, etc. can all blow a hole in the world's most accurate "fair" odds line pretty quickly.

All these thing can be, and are, incorporated in to an odds line without you even realizing it. If you give a horse a 30% chance of winning you are not saying anything about how it will go about losing 70% of the time, so you really don't have to detail all the different ways a horse can lose in order to make an accurate line.

[Sapio]

I would think that it is nearly impossible to derive an accurate and useable fair odds line without taking into account confounding variables. If you are unaware of variables that are confounded, you are better off not making an odds-line. It will fail for the purpose of betting.

Thomas Sapio

[shank]

i've been making odds lines daily for 20 years.it's WORK which most people involved in racing don't see as a necessity. It's merely ranking the horses in order of preference. Like any task it takes practice over time to feel comfortable with your results. I always start with the worst horses and work my way to the most preferred. It's a work in progress, massaging the values until I get the sense that my odds line represents my feelings about the horses.

[Hoofless_Wonder]

All these thing can be, and are, incorporated in to an odds line without you even realizing it. If you give a horse a 30% chance of winning you are not saying anything about how it will go about losing 70% of the time, so you really don't have to detail all the different ways a horse can lose in order to make an accurate line.

When you say "accurate line", do you mean a line that can be demonstrated or proven to be truly accurate in predicting the outcome of the race for all the horses, or simply accurate in the sense it can be used to make a profit by zeroing in on the value horses? I see those as two different things, but maybe they're not.

For example, in today's Fountain of Youth, I liked the :8: dog, Almost Famous, and gave him about a 20 percent chance to win. Going off at 13-1, and with my average at 14-15% picking the winner in a 3YO/stakes/allowance race, it was as easy bet to make, since I liked his chances about twice as much as any other horse in the race.

So, he gets steadied at the start, doesn' sniff the lead and is done early. The :4: horse breaks alertly, and gamely hangs on for the win. If the race is run 100 times, I don't have a clue, really, whether the :8: wins 20 times, 40 times, or never. But I do know that eventually, another horse I like about as much will come along at those odds, and win for me.

In other words, since the race is only run once, I don't really worry about the accuracy of the odds line for each horse in the race, but I do need to keep in mind the odds of my favored selection versus the long-term strike rate.

[traynor]

I would think that it is nearly impossible to derive an accurate and useable fair odds line without taking into account confounding variables. If you are unaware of variables that are confounded, you are better off not making an odds-line. It will fail for the purpose of betting.

Thomas Sapio

Quite true. And every horse in every race carries a set of confounding (and largely ignored) variables that affect the outcome.

[Overlay]

I am trying to get a handle on how to attach value to contending horses.

For example, is there a formula to follow on whether a favorite is 6-5 or 2-1?

If it's a 10 horse field and I throw out 5 horses as non-contenders how do I know where to start with the horse I like best? Is it what I am willing to accept as "fair" odds? If I make my top pick 2-1, where do I go from there?
For me, using weighted statistical probabilities to determine the winning chances and fair odds of each of the horses in a field takes no longer than developing a line for only the "contenders". I've found it to be more consistently accurate than qualitative or intuitive evaluation, as well as helpful in making sure that there aren't significant positive aspects among the "non-contenders" that I might overlook entirely if I discarded them out-of-hand. (Not to mention missing out on the boxcar mutuels that "non-contenders" can pay on those occasions when they do come in.)

[traynor]

For me, using weighted statistical probabilities to determine the winning chances and fair odds of each of the horses in a field takes no longer than developing a line for only the "contenders". I've found it to be more consistently accurate than qualitative or intuitive evaluation, as well as helpful in making sure that there aren't significant positive aspects among the "non-contenders" that I might overlook entirely if I discarded them out-of-hand. (Not to mention missing out on the boxcar mutuels that "non-contenders" can pay on those occasions when they do come in.)

Great advice!

[therussmeister]

When you say "accurate line", do you mean a line that can be demonstrated or proven to be truly accurate in predicting the outcome of the race for all the horses, or simply accurate in the sense it can be used to make a profit by zeroing in on the value horses? I see those as two different things, but maybe they're not.

Accurate in the sense it can be used to make a profit by zeroing in on the value horses. It is impossible to demonstrate any line is accurate in predicting the probability of winning for all horses.

For the record, I bet very few win bets, so I do not make an odds line, but I used to years ago.

[Hoofless_Wonder]

Accurate in the sense it can be used to make a profit by zeroing in on the value horses. It is impossible to demonstrate any line is accurate in predicting the probability of winning for all horses.

For the record, I bet very few win bets, so I do not make an odds line, but I used to years ago.

That makes sense - trying to exploit the overlays, and it does really seem to come down to "horse is worth a play" or the "horse is not worth a play". If you're not making win bets very often though, then one needs a more complex odds line - for each horse, and for each finish position.

The part I struggle with is that an "accurate" odds line represents smoothing to me - introducing more stats into the process just seems to drive the result closer to the public odds. I prefer to approach it with the chaos of changing the variables weights on the fly, not that it's always successful. When I find a race I want to play, it usually looks something like this:

Horse M/L Tote Fair Odds My odds
*****************************
#1 8-5 6-5 2-1 7-5
#2 2-1 2-1 2-1 4-1
#3 3-1 5-2 3-1 5-2
#4 5-1 7-1 10-1 10-1
#5 10-1 12-1 25-1 3-1

The trainer of the :5: headed to the finish line bar after saddling his horse, and bought a beer in anticipation of the winner's circle photo in a few minutes. This lowered my odds for that horse, as I've weighted that confounding variable very high. After the race, of course, the :5: horse runs poorly and I find out the the trainer was just thirsty. D'oh! :bang:

[thaskalos]

I view the topic of betting lines the same as I do the rhetoric of our two main political parties.

I am impressed by the theory being presented...but I am just never satisfied when I see this theory put to practice.

[Sapio]

I view the topic of betting lines the same as I do the rhetoric of our two main political parties.

I am impressed by the theory being presented...but I am just never satisfied when I see this theory put to practice.

Hi thaskalos,

I disgree. I've spent nearly 12 hours per day studying the posts by TrifectaMike. His posts on making odds-lines are amazing. In fact, he posted oddsline for several tracks and they were very very good.

I'm truly amazed at how much great information is contained on this site. One could easily write a book that has not been written.

I just wish I knew more about Bayesian statistics. I know just enough to understand TM's posts, but implementation is my weak point. I hope to remedy this weakness by 2015.

Thomas Sapio

[thaskalos]

Hi thaskalos,

I disgree. I've spent nearly 12 hours per day studying the posts by TrifectaMike. His posts on making odds-lines are amazing. In fact, he posted oddsline for several tracks and they were very very good.

I'm truly amazed at how much great information is contained on this site. One could easily write a book that has not been written.

I just wish I knew more about Bayesian statistics. I know just enough to understand TM's posts, but implementation is my weak point. I hope to remedy this weakness by 2015.

Thomas Sapio
Hi Thomas,

I too have been impressed by TrifectaMike's knowledge and intellect...and I have deeply regretted the fact that I lack the necessary mathematical acumen to fully grasp what he is trying to convey. Had I been aware during my school years of the pivotal role that mathematics would play in my subsequent fascination with gambling...I would have, no doubt, paid a lot more attention during math class.

The shortcomings in creating an effective odds-line are clearly my own...

[Maximillion]

Im a firm believer in basically having a plan in place....irrespective of the toteboard.
Im not concerned at all (mostly) with how the public bets the race....for better or worse.

[Magister Ludi]

I've spent nearly 12 hours per day studying the posts by TrifectaMike. His posts on making odds-lines are amazing.
Thomas Sapio

That is the wisest endeavor that you can undertake on this site! Dr. Beav always maintained that someone aspiring to become a proficient trader in the racetrack betting market should learn all that he can about odds. I also believe that it's far more fruitful to handicap the handicappers than to handicap the horses.

[Cratos]

That is the wisest endeavor that you can undertake on this site! Dr. Beav always maintained that someone aspiring to become a proficient trader in the racetrack betting market should learn all that he can about odds. I also believe that it's far more fruitful to handicap the handicappers than to handicap the horses.

A very astute observation from one of my favorite posters who along with TM doesn't share their wealth of statistical/mathematical acumen nearly or frequently enough on this forum.

[pondman]

The late Dick Mitchell taught me this - Ask yourself the question "If this race were run 100 times how many times (or what percentage of the time) would X horse win?" If the answer is 33% then set 2 to 1 odds, if it's 20% then set 4 to 1, etc.. After you have odds for all your contenders, try to get the total to 80%. It's pretty subjective so don't worry if it's not perfect.


80% is a little high for contenders. If your contenders are chosen from one method, I'd suggest 69%. I believe you will be disappointed with the over confidence.

[raybo]

80% is a little high for contenders. If your contenders are chosen from one method, I'd suggest 69%. I believe you will be disappointed with the over confidence.

I think this depends mostly on how many contenders you have, and field size. In a full field (10-12 horses), if you have 4 contenders then 65-75% is probably about right. With 3 contenders my experience is that 50-64% is about right, with 2 contenders 35-49%, and only 1 contender 25-34%. I think that, over the long term, the above ranges can be achieved by a good handicapper or automated system, and also produce a positive ROI.

One might do better on the ranges (closer to what the public does), but the ROI will suffer, as you will be sitting on more low priced horses (as the public endlessly proves).

[Some_One]

For example, in today's Fountain of Youth, I liked the :8: dog, Almost Famous, and gave him about a 20 percent chance to win. Going off at 13-1, and with my average at 14-15% picking the winner in a 3YO/stakes/allowance race, it was as easy bet to make, since I liked his chances about twice as much as any other horse in the race.


If you have a 13-1 horse at 4-1, then is a 99% chance your line is wrong. The numbers show that the toteboard is pretty damn efficient as it is.

[Overlay]

If you have a 13-1 horse at 4-1, then is a 99% chance your line is wrong. The numbers show that the toteboard is pretty damn efficient as it is.
I would agree with the ongoing aggregate/overall efficiency of the toteboard, but I think that there is still room for significant errors by the public on a race-to-race basis, especially if I can identify one or more reasons why the odds are out of line, and satisfactorily explain or discount those reasons (which basing my composite odds assignment on individual handicapping elements helps me to do).

[Sapio]

A problem with fair oddsline is that you can have numerous lines. One must have a way to resolve which is the most accurate.

Since I come from a signal processing/information systems background. I would test the lines using information theory.

I would measure information content of the winner as suggested by the fair oddsline.

Thomas Sapio

[Fox]

Hi thaskalos,

I disgree. I've spent nearly 12 hours per day studying the posts by TrifectaMike. His posts on making odds-lines are amazing. In fact, he posted oddsline for several tracks and they were very very good.

I'm truly amazed at how much great information is contained on this site. One could easily write a book that has not been written.

I just wish I knew more about Bayesian statistics. I know just enough to understand TM's posts, but implementation is my weak point. I hope to remedy this weakness by 2015.

Thomas Sapio

Anyone know if Trifecta Mike is still around? I notice he hasn't posted in several months.

[classhandicapper]

I view the topic of betting lines the same as I do the rhetoric of our two main political parties.

I am impressed by the theory being presented...but I am just never satisfied when I see this theory put to practice.

In theory making an odds line is definitely the correct way to play.

In practice your odds line will be dependent on the figures and information you use as well as how you weigh it all. Give the same guy 2 different sets of figures or information and he'll come up with 2 different odds lines. That doesn't inspire a lot of confidence.

My personal experience suggests that overlays leap off the pages at you when you have 1 or 2 specific insights into the race that you're pretty sure are not being weighed properly by the public rather than by your ability to weigh loads of information properly and create a line on every horse.

I might have a very general line like "this horse looks like he deserves to be the favorite in the 8/5 to 5/2 range, but I know his trainer is horrible in this spot. So I'd like to bet against him". Then as long he's the favorite, I'll follow through on the bet against. But I won't nitpick and create an exact line on him and every other horse.

[overthehill]

I donT make a line per se but I do try to pick the contenders and put some odds on them. I am curious though as to whether you adjust your odds for what you observe on the track. In the fountain of youth, i reduced the likelihood of the pletcher horses winning by a lot after seeing the speed track bias. In the end i only felt comfortable using the pletcher horses in a box triple with first two finishers hoping that the speed bias would work in their favor and against top billing. that didnt work.

On sunday at gp after the longshot step up from mdn claiming ranks won in allowance company. I upgraded the pletcher horse who ran second to him from 2-1 to even money. I thought the 3-1 he went off at was a huge overlay and can only assume that the bettors didnt adjust for the result of the previous race.

I m also curious as to whether you adjust your line base on early flashes on the tote board or what the morning line maker does.

[overthehill]

i dont know how you can explain the discount. I have given up trying. I have seen horses come back in the same class off a win pay $50. I have seen horse Dqued in the same class after winning come back and pay $16. granted this only happened once but there was a horse who debuted in a msw at monmouth run second or third as the 6/5 favorite and come back in a msw in kentucky and pay over $30 to win! Years ago in a new york maiden race there was a horse who in his previous race first on turf was checked and stopped three times and still finished a close up third. his next time out he was listed as the 4-1 second choice from the 12 post. he wired the field won by 6 lengths and paid $26. It seem like whenever i come up with some excuse as to why not to play a horse that looks like a big overlay such as a 10 lb bug boy who has not yet won a race the horse wins anyway.

probably the biggest overlay i have ever seen was when genuine risk and
bold n determined met in a stake race in new york. I think the ml was something like 6/5 and 7/5. I didnt bother with it and was shocked to find that bold n determined paid over $16 to win.

[Hoofless_Wonder]

If you have a 13-1 horse at 4-1, then is a 99% chance your line is wrong. The numbers show that the toteboard is pretty damn efficient as it is.

I'd argue that in this case, my odds line had a 100% chance of being wrong. :D A 3YO stakes race in February ain't easy.

I'd also argue, that my odds line was irrelevant, except in the context of identifying a playable horse. 13-1 shots win about 6% of the races, and when they do win, it's a win. Period. (except when the stewards get involved). That would be an odds line of 100%. In the case of my pick, 20%, 2%, 60%, the horse still finished second to last. Was it a good bet? Only time will tell if the horse shows enough talent that the FOY was a "tosser". Part of my problem is trying to relieve the glory of making a big bet on Builtforpleasure in the 1996 FOY. Can't play anything less than 10-1 in that race after beating Unbridled's Song that year. But when I handicap a very contentious race, have some extra value from some underlays, and end up liking a horse twice as much as any other in the race - I make that bet every time if the odds are right.

Toteboard "efficiency" is somewhat misleading IMHO - the favorite only wins 1/3 of the races, and it's a rare race indeed in which the final odds reflect the actual chances of winning. Yes, it's difficult indeed to outpick the toteboard over time in terms of the number of winners, but it's a fool's game for most of us to attempt to use advanced statistical methods to separate a 9% chance horse from an 11% chance horse, or even a 35% chance horse from a 50% chance horse.

Look at all the debate on the various handicapping factors on the forum here. We couldn't even get consensus on defining what the independent, dependent, confounding variables are - let alone try to weight them....

[Sapio]

In theory making an odds line is definitely the correct way to play.

In practice your odds line will be dependent on the figures and information you use as well as how you weigh it all. Give the same guy 2 different sets of figures or information and he'll come up with 2 different odds lines. That doesn't inspire a lot of confidence.


Hi classhandicapper

I don't see different sets of odds line as a problem. In fact, there i nothing special about the odds line infered by the tote-board. They are all subjective.

The critical part is in the evaluation of the odds lines against an objective criteria other than predicted winners vs actual winners.

Thomas Sapio

[classhandicapper]

Hi classhandicapper

I don't see different sets of odds line as a problem. In fact, there i nothing special about the odds line infered by the tote-board. They are all subjective.

The critical part is in the evaluation of the odds lines against an objective criteria other than predicted winners vs actual winners.

Thomas Sapio


I know what you are saying.

I guess it's a style issue or personal quirk of mine. My view is that there's an underlying truth to this game. I'm always trying to get closer to it. Part of that is having accurate information. One of the problems I always had along those lines was looking at different sets of figures. After awhile, when you see enough differences (some of which are huge), you start to question every play because you don't know if the horse you are considering betting is an actual overlay or you are simply looking at an inaccurate figure.

I feel the same way in certain other situations unrelated to figures.

One example is on turf. I'm not much of a pedigree guy. So if I am considering playing a turf race with several horses trying turf for the first time, I know I can't value them really well. Instead of it being an accuracy issue, in this case it's an information gap.

My game is constantly evolving, but one thing that remains steady is that I do best when I have a horse that I think is better/worse that it looks on paper for a very specific reason as opposed to trying to weigh all the information properly.

If I have an insight like that, I could even play that turf race above with some confidence even if I couldn't value "some" of the horses really well.

[raybo]

If one has a fairly consistent contender selection method, and one knows what the hit rate of his top, 2nd, 3rd, 4th contenders are, at different field sizes, and race types, couldn't one calculate the win probabilities of each of those selections, convert it to odds and then fudge it a point or 2 to get a fair odds line?

Yeah, I know, most people don't do a good job of record keeping, and therefore couldn't create a very good line, but suppose someone actually did perform that very important task, record keeping, just for grins.

[Poindexter]

Every time this subject is brought up there is often a chorus about how odds lines cannot be accurate(like there is a more accurate way of determining whether a horse offers value)……The purpose of an odds line is to make you analyze a race determine how much chance each horse has and right down what his fair price is. Is that a lot of work? Yes. Is it worth the work? That’s subjective. Are there alternatives? Of course. They are just similar (just far less accurate) ways of getting to the same information (Which horse is offering enough value to be worthy of a play).

Will a personal odds line be 100% accurate? Of Course it won’t. That is why the people who write on the subject usually demand a significant premium. If you are betting a horse you make 3-1 at 9/2 you do not have to win 25% of the time (as you project) you only have to win 18% of the time to break even. Whether that margin of error is enough to get a specific handicapper into the profit zone is dependent of that handicappers skill set. You are delusional if you think just because you make a line and bet value you are guaranteed to make money. That being said, I think you will likely improve your results over the long run no matter who you are and over time perhaps you can cross into the profit zone. Once again it depends on your personal skill set.

By the way it was brought up that if you make a horse 3-1 and he is 14-1 there is no way the horse has in actuality a 25% chance of winning. That is probably true with 99%+ of all handicappers(and yes I am certainly in the 99% group). So if you do go through the process of making a line, you do need to make modifications to your line. I think as a rule, you should not expect that the public will give you over double your fair odds line. So in the above example re-evaluate and figure out what is going on. Maybe you have info that most do not. For example you are at a master of detecting stiff jobs and in your opinion this horse was stiffed 2 straight times, or the horses form is clouded because he has been making huge wide moves on heavily speed biased tracks and today the track is fair or better yet have an outside closer bias. Maybe you are master at watching horses warm up and after look lame his last three starts today he looks like a stakes horse and that is why you give him a high chance of winning. If you are privy to info most are not stick to your guns. On the other hand, maybe the crowd has info you do not have. Horse has gone lame, is not cranked up, is out for a tightner, has benefitted from easy pace situations in the last 2 starts and you do not know anything about pace handicapping, you missed something in your handicapping that everyone sees but you………………….

Bottom line is that unless you have a strong reason to stick to your guns, you will at the very least have to change the line on this horse to 7-1 (half of the public line). This will save you from betting too much. If there are some partial reasons why you think the horse is fooling the public maybe 6-1 or 5-1……

Another point of concern is when you have 3 horses offering value in the same race. If this happens, you screwed up, plain in simple (I am talking fields of 13 or less). Either pass the race of make a tiny exotic play with your “value” and leave it at that. My experience tells me that 3 horses offering value in one race means I will virtually always lose. Another point of concern is the dead board horse. A horse that is say morning line 3-1 opens at 7 and never gets down below say 9-2. The no early action can be a bad sign, and the fact that the public who all sees 3-1 on the program never finds a reason to bet him is another bad sign. Be careful with these guys. I will usually chop my bet in half on these types of horses. I am sure as time goes on you will or have recognized other situations where you have to be careful or adjust your line. It is all part of the learning curve. Little tweaks like these will help your bottom line.

Another thing you should pay attention to is horse you do not like that the public likes. For instance you make a horse 30-1 and the public has him at 6-1. That is pretty significant. Further tweaks with these types of horses are likely needed to make your line more accurate.

[PaceAdvantage]

Good stuff Poindexter...thanks so much for sharing!

This place needs more posts like the above. :ThmbUp:

[lamboguy]

an odds line is only as good as the input one uses to make it. i have one that is based strictly on pace., its not bad.

but even if one does produce a great odds line, i question if its possible to make money on it alone. how many times has this happened to horse players after a race when they see a horse in the winner's circle and he pays lets say $8 to win. "i would have had that one if i knew he was going to return that much" the problem is that when the horse was in the starting gate he was 2-1 and for what ever reason other horses were bet and this one was ignored. for sure you will get the odd ball race where you will find a horse that meets more than your expectations and you get a generous price. but my observations have been that when i find horses that in my mind has fair value of 7-1 that leave the gate 8-1 they over correct down to 6-1 no matter how big the pools are. the big money is running their computer programs and come up with the same overlay's as i did and now make them underlay's at the window's. its probably better to be on underlay's and hope that your horse is ignored and goes up in price.

[traynor]

an odds line is only as good as the input one uses to make it. i have one that is based strictly on pace., its not bad.

but even if one does produce a great odds line, i question if its possible to make money on it alone. how many times has this happened to horse players after a race when they see a horse in the winner's circle and he pays lets say $8 to win. "i would have had that one if i knew he was going to return that much" the problem is that when the horse was in the starting gate he was 2-1 and for what ever reason other horses were bet and this one was ignored. for sure you will get the odd ball race where you will find a horse that meets more than your expectations and you get a generous price. but my observations have been that when i find horses that in my mind has fair value of 7-1 that leave the gate 8-1 they over correct down to 6-1 no matter how big the pools are. the big money is running their computer programs and come up with the same overlay's as i did and now make them underlay's at the window's. its probably better to be on underlay's and hope that your horse is ignored and goes up in price.

More than a few are making serious money by leveraging the myth of "value wagering." Horses run as horses run, not according to the dictates of the tote board. One should occasionally reward those who pass on betting a mortal lock to bet another horse in the same race that "offers better value." From such foolishness are fortunes being made, like plucking low-hanging fruit within easy reach of most, but only recognized by a few.

[raybo]

If one has a fairly consistent contender selection method, and one knows what the hit rate of his top, 2nd, 3rd, 4th contenders are, at different field sizes, and race types, couldn't one calculate the win probabilities of each of those selections, convert it to odds and then fudge it a point or 2 to get a fair odds line?

Yeah, I know, most people don't do a good job of record keeping, and therefore couldn't create a very good line, but suppose someone actually did perform that very important task, record keeping, just for grins.

Since nobody commented on my quoted post, am I to assume that this will not work?

I agree with PA that Poindexter's post is a very good one, and makes some very good points.

Lambo's post reiterates a common problem that all of us must deal with, late changes in the odds, after we have made our bets. But, his final sentence of; it probably being better to be on underlays and hope that the public ignores it, causing the odds to rise late, makes very little sense to me. If the public, including those who are truly exploiting the public's perception of "value"(whales/syndicates), is so good , then an "underlay" will probably not rise in odds late, enough to make that argument viable.

Traynor's post, which like many of his posts, is more than a bit fuzzy in it's presentation, seems to say the same thing. That being that there are people out there who will see that "preceived" value, and capitalize on it by betting into that value. But, because of the way he worded his post, I'm not really sure what he said. The first part seems to say that "anyone" who bets, with value in mind, is betting a "myth" and will fail long term. The "anyone" in my statement is key, IMO it depends on the skill of the person betting the "value". I tend to believe that seeking "value" is the only way to beat this game, long term, and since Traynor offered no empirical proof of his statement of supposed "truth", then I am left with my own "truth", that value betting is key. Once again Traynor enters into a discussion and offers nothing of value, IMO. His posts scream "Type A" personality, with the classic motto of "Lead, follow, or get the f--- out of the way!!", except he tends to delete the first word in that motto most of the time.:bang:

[lamboguy]

Since nobody commented on my quoted post, am I to assume that this will not work?

I agree with PA that Poindexter's post is a very good one, and makes some very good points.

Lambo's post reiterates a common problem that all of us must deal with, late changes in the odds, after we have made our bets. But, his final sentence of; it probably being better to be on underlays and hope that the public ignores it, causing the odds to rise late, makes very little sense to me. If the public, including those who are truly exploiting the public's perception of "value"(whales/syndicates), is so good , then an "underlay" will probably not rise in odds late, enough to make that argument viable.

Traynor's post, which like many of his posts, is more than a bit fuzzy in it's presentation, seems to say the same thing. That being that there are people out there who will see that "preceived" value, and capitalize on it by betting into that value. But, because of the way he worded his post, I'm not really sure what he said. The first part seems to say that "anyone" who bets, with value in mind, is betting a "myth" and will fail long term. The "anyone" in my statement is key, IMO it depends on the skill of the person betting the "value". I tend to believe that seeking "value" is the only way to beat this game, long term, and since Traynor offered no empirical proof of his statement of supposed "truth", then I am left with my own "truth", that value betting is key. Once again Traynor enters into a discussion and offers nothing of value, IMO. His posts scream "Type A" personality, with the classic motto of "Lead, follow, or get the f--- out of the way!!", except he tends to delete the first word in that motto most of the time.:bang:you obviously don't pay to much attention to Mountaineer. there are plenty of horses there that should be legit 3/5 shots that are in the gate at 1-9 and return $4.80 or more, i have seen them pay as much as $7.00. and Mountaineer does open its meet up tomorrow, Saturday evening March 1.

and please correct me if i have this wrong!

[thaskalos]

Every time this subject is brought up there is often a chorus about how odds lines cannot be accurate(like there is a more accurate way of determining whether a horse offers value)……The purpose of an odds line is to make you analyze a race determine how much chance each horse has and right down what his fair price is. Is that a lot of work? Yes. Is it worth the work? That’s subjective. Are there alternatives? Of course. They are just similar (just far less accurate) ways of getting to the same information (Which horse is offering enough value to be worthy of a play).

Will a personal odds line be 100% accurate? Of Course it won’t. That is why the people who write on the subject usually demand a significant premium. If you are betting a horse you make 3-1 at 9/2 you do not have to win 25% of the time (as you project) you only have to win 18% of the time to break even. Whether that margin of error is enough to get a specific handicapper into the profit zone is dependent of that handicappers skill set. You are delusional if you think just because you make a line and bet value you are guaranteed to make money. That being said, I think you will likely improve your results over the long run no matter who you are and over time perhaps you can cross into the profit zone. Once again it depends on your personal skill set.

By the way it was brought up that if you make a horse 3-1 and he is 14-1 there is no way the horse has in actuality a 25% chance of winning. That is probably true with 99%+ of all handicappers(and yes I am certainly in the 99% group). So if you do go through the process of making a line, you do need to make modifications to your line. I think as a rule, you should not expect that the public will give you over double your fair odds line. So in the above example re-evaluate and figure out what is going on. Maybe you have info that most do not. For example you are at a master of detecting stiff jobs and in your opinion this horse was stiffed 2 straight times, or the horses form is clouded because he has been making huge wide moves on heavily speed biased tracks and today the track is fair or better yet have an outside closer bias. Maybe you are master at watching horses warm up and after look lame his last three starts today he looks like a stakes horse and that is why you give him a high chance of winning. If you are privy to info most are not stick to your guns. On the other hand, maybe the crowd has info you do not have. Horse has gone lame, is not cranked up, is out for a tightner, has benefitted from easy pace situations in the last 2 starts and you do not know anything about pace handicapping, you missed something in your handicapping that everyone sees but you………………….

Bottom line is that unless you have a strong reason to stick to your guns, you will at the very least have to change the line on this horse to 7-1 (half of the public line). This will save you from betting too much. If there are some partial reasons why you think the horse is fooling the public maybe 6-1 or 5-1……

Another point of concern is when you have 3 horses offering value in the same race. If this happens, you screwed up, plain in simple (I am talking fields of 13 or less). Either pass the race of make a tiny exotic play with your “value” and leave it at that. My experience tells me that 3 horses offering value in one race means I will virtually always lose. Another point of concern is the dead board horse. A horse that is say morning line 3-1 opens at 7 and never gets down below say 9-2. The no early action can be a bad sign, and the fact that the public who all sees 3-1 on the program never finds a reason to bet him is another bad sign. Be careful with these guys. I will usually chop my bet in half on these types of horses. I am sure as time goes on you will or have recognized other situations where you have to be careful or adjust your line. It is all part of the learning curve. Little tweaks like these will help your bottom line.

Another thing you should pay attention to is horse you do not like that the public likes. For instance you make a horse 30-1 and the public has him at 6-1. That is pretty significant. Further tweaks with these types of horses are likely needed to make your line more accurate.

IMO...these types of discussions stagnate not because the argument is made -- by the "non-believers" -- that "an odds line cannot be accurate"...but because the practitioners of odds-line play do not do a satisfactory job of explaining how they handle some of the more troubling aspects of today's game.

Horse racing is a unique game...not because it's a game of "incomplete information"...but because it's a game of inside information. In this game... some very relevant information, which is unknown to the masses, is known to some. And these "some" are then allowed to bet against the ignorant masses. This is how situations like the dreaded "dead-horse-on-the-board syndrome" are created.

You yourself have acknowledged that the dead-on-the-board horse is a bad sign...and you've stated that you chop your bet in half in these cases. But WHY are you cutting your bets in half in these cases? If you are still showing a profit on these types of bets...then cutting these bets in half would also cut the PROFIT in half -- so it could hardly be advisable. And if you are NOT making a profit with these bets...then cutting them off ALTOGETHER makes a lot more sense than cutting them in half.

My point has always been that these "dead-on-the-board" horses seemingly represent some of our biggest "overlays"...and the fact that we don't have confidence in these "big overlays" proves that we don't have confidence in our method of line-making itself.

And if we don't trust our betting line when it comes to our "biggest overlays"...then what good is our betting line in the first place?

[raybo]

you obviously don't pay to much attention to Mountaineer. there are plenty of horses there that should be legit 3/5 shots that are in the gate at 1-9 and return $4.80 or more, i have seen them pay as much as $7.00. and Mountaineer does open its meet up tomorrow, Saturday evening March 1.

and please correct me if i have this wrong!

You are correct, I don't pay much attention to Mnr, and hundreds of posts on this forum point out the reasons for that.

But, stating something the way you did in your previous post , and then stating as proof, a single track (especially Mnr), hardly changes my feeling on your original post.

To each his own. But, remember, if you keep doing (and thinking) the same things, you will keep getting what you have been getting. So, if you are doing very well with your thinking, keep doing it, but if you aren't, then maybe you ought to make some changes. ;)

[Sapio]

you obviously don't pay to much attention to Mountaineer. there are plenty of horses there that should be legit 3/5 shots that are in the gate at 1-9 and return $4.80 or more, i have seen them pay as much as $7.00. and Mountaineer does open its meet up tomorrow, Saturday evening March 1.

and please correct me if i have this wrong!

Hi lamboguy

Isn't that phenomonen track specific and normally as a result of a cancelled large wager?

Thomas Sapio

[raybo]

Hi lamboguy

Isn't that phenomonen track specific and normally as a result of a cancelled large wager?

Thomas Sapio

Didn't want to say that, but that's only one of the reasons I don't play Mnr. :ThmbUp:

[raybo]

Here's a link to a thread started quite a while ago, that deals with odds line creation. Unfortunately, the point at which the OP stopped posting is prior to the stuff most of us who have been around a while are really interested in knowing. But still, there is some good stuff there, if you can wrap your head around the math.

Probability and Odds (http://alldataexcel.freeforums.org/creeating-money-lines-part-1-probabilty-and-odds-t18.html)

[lamboguy]

another example of odds moving up after the break. 2nd race at Charlestown, 5 horse enters gate at 1/9, in the gate 1-5, goes to the lead no less, wins race and pays $3.80

i will show you plenty of tracks like this including big ones.

[traynor]

another example of odds moving up after the break. 2nd race at Charlestown, 5 horse enters gate at 1/9, in the gate 1-5, goes to the lead no less, wins race and pays $3.80

i will show you plenty of tracks like this including big ones.

One might conclude that someone somewhere knows something that goes completely over the heads of the "value seekers" and is profiting handsomely from that knowledge. Profit is good. Chasing rainbows is bad.

Unless, of course, one is only talking about it, and writing about it, rather than actually doing it.

[lamboguy]

another example this evening from the 5th at Turfway Park. the 5-1 combination leaves gate as a $11 exacta, wins race @ 2-5, 3rd choice runs second @6-1 in 7 horse field, exacta returns $20.

[lamboguy]

moving over to the 6th at Meadowlands race 6. #2 AMERICAN IN PARIS leaves gate at 3-5 returns home a winner and pays $5.60

[traynor]

moving over to the 6th at Meadowlands race 6. #2 AMERICAN IN PARIS leaves gate at 3-5 returns home a winner and pays $5.60

Incremental wagering enables an entry to look like an underlay, avoid the last minute money dump, trick the "value seekers" into pumping money on another entry--and provide a nice return. That is the way the game is played in the big city. Hong Kong, too.

[Hoofless_Wonder]

... So if you do go through the process of making a line, you do need to make modifications to your line. I think as a rule, you should not expect that the public will give you over double your fair odds line. So in the above example re-evaluate and figure out what is going on.

Bottom line is that unless you have a strong reason to stick to your guns, you will at the very least have to change the line on this horse to 7-1 (half of the public line). This will save you from betting too much. If there are some partial reasons why you think the horse is fooling the public maybe 6-1 or 5-1……

This is exactly the wrong approach IMHO, at least for some races. I gain no comfort in having my odds line, as it were, be closer to those reflected on the tote. I agree one must be cautious of "dead on the board" horses, but sometimes you've got to step up and be bold.

Let me be clear, I don't often make an effort at a detailed odds line, simply because there are not many races which offer enough info to do it.

I make hay on the races which offer less data to work with, because those are the ones which offer more overlays. 3YO stakes races in February are great value - the animals are undergoing rapid changes, with some moving forward in great leaps, while other 2YO sensations are not living up to the hype.

What is being left out of the discussion is weighting the race, or a confidence factor. Surely not odds lines are the same? When wagering, does not one have to compare the "fair odds" with the relative comfort level of those odds?

Most bettors feel more comfortable with wagering $100 to win on a 6/5 shot, and can't hardly bet more than $5 on 20-1 shots. Does the 20-1 shot and a $5 wager really mean I'm only 1/20th as confident?

In the FOY, that mule ALMOST FAMOUS was a great price, but not a great bet. The amount of my wager reflected that - a medium sized bet for me, only made slightly larger by his odds.

[raybo]

moving over to the 6th at Meadowlands race 6. #2 AMERICAN IN PARIS leaves gate at 3-5 returns home a winner and pays $5.60

Keep going because you're going to run out of "cases" soon, much sooner than anyone else would by posting races where the opposite happened.

[raybo]

Incremental wagering enables an entry to look like an underlay, avoid the last minute money dump, trick the "value seekers" into pumping money on another entry--and provide a nice return. That is the way the game is played in the big city. Hong Kong, too.

Wow! No kidding? We had no idea that kind of stuff was going on. :rolleyes:

[Poindexter]

IMO...these types of discussions stagnate not because the argument is made -- by the "non-believers" -- that "an odds line cannot be accurate"...but because the practitioners of odds-line play do not do a satisfactory job of explaining how they handle some of the more troubling aspects of today's game.

Horse racing is a unique game...not because it's a game of "incomplete information"...but because it's a game of inside information. In this game... some very relevant information, which is unknown to the masses, is known to some. And these "some" are then allowed to bet against the ignorant masses. This is how situations like the dreaded "dead-horse-on-the-board syndrome" are created.

You yourself have acknowledged that the dead-on-the-board horse is a bad sign...and you've stated that you chop your bet in half in these cases. But WHY are you cutting your bets in half in these cases? If you are still showing a profit on these types of bets...then cutting these bets in half would also cut the PROFIT in half -- so it could hardly be advisable. And if you are NOT making a profit with these bets...then cutting them off ALTOGETHER makes a lot more sense than cutting them in half.

My point has always been that these "dead-on-the-board" horses seemingly represent some of our biggest "overlays"...and the fact that we don't have confidence in these "big overlays" proves that we don't have confidence in our method of line-making itself.

And if we don't trust our betting line when it comes to our "biggest overlays"...then what good is our betting line in the first place?


Thaskalos, I think the point you are trying to make goes something like this. If we had a magical computer that was able to analyze every race and give the exact probability of each entrant that was 100% accurate(sans the other stuff that comes up, 3 went lame since last start, 7 is juiced up today, 9 had a breathing operation since last start, 11 was sick a while back and was purposely held back last two so the connections could unload on him today, 12 out of the blue after being very sore the last 4 races suddenly feels great today), it probably would not do us that much good to have our perfect line, because our line is going to way underestimate the chances of the 3, 7,9, 11, 12. Moreover, the connection and insiders are all going to be sending in cash on these horses driving up the prices and giving us artificial value on the other 7 horses. We are going to be looking at our computers odds line and see the 5) 2-1 paying 7/2 and the 8)4-1 paying 7-1. So we are going to make our bets on these two horses and even though we think we are getting whopping value, chances are we are not getting any value at all. In this instance you are absolutely right. That being said, because we demand a premium, even with all that is against us in this race, we might not be at that much of a disadvantage.

This is the game we play and if you are correct and there is that much working against us in every race than no matter what we do, we probably will never show a long term profit in this game no matter what we do(unless we are privy to the inside information-connections, expert tote watching…..)The only way anyone here finds out if you are right is to go through the process and see if they can make a profit long run. I still feel that you are better off making a line and betting value. The premium you demand helps protect you against the very thing you are warning against. Rules such as cutting your bet down when your horse looks dead board or passing the race when 3 horses are offering value and whatever else you may come up (when you develop a large sample of bets) with helps you find the “danger races” and hopefully helps get you in the black.


Regarding your point about passing on the dead board horse. Theoretically if long term data shows that these plays are losers, you are probably right, BUT, we are human and it is not easy to see a horse that we make 2-1, be dead on the board, pay 5-1 and run off by 5. So by betting half the amount we limit the exposure but still make a nice profit when we are right. This is a much better alternative to me. For those with the discipline to pass because there long term records show that a certain situation is not profitable, more power to them.

By the way I disagree that dead on the board horses are your biggest value situations. That has not been my experience at all. A lot of times horses are dead on the board for fairly obvious reasons or because someone else is getting hammered so much (not because everyone in the barn knows the horse is not sound).

One final note about line making is this. The reason you make a line is because sometimes for some reason when you line a race you are just fare away from 100% probability. Sometime you underestimate the group and you can be a low as all the horses adding up to .83 and other times(a more common experience) you are somewhere between 120 and 130 %. You really have to work to get your numbers ironed out. By the time you do the horse you originally made 2-1 is now 3-1. If you did not go through this process you would be thinking the horse was worth 2-1 when in actuality the fair value is 3-1. Making a line is simply the art of quantifying your opinion.

[Poindexter]

This is exactly the wrong approach IMHO, at least for some races. I gain no comfort in having my odds line, as it were, be closer to those reflected on the tote. I agree one must be cautious of "dead on the board" horses, but sometimes you've got to step up and be bold.

Let me be clear, I don't often make an effort at a detailed odds line, simply because there are not many races which offer enough info to do it.

I make hay on the races which offer less data to work with, because those are the ones which offer more overlays. 3YO stakes races in February are great value - the animals are undergoing rapid changes, with some moving forward in great leaps, while other 2YO sensations are not living up to the hype.

What is being left out of the discussion is weighting the race, or a confidence factor. Surely not odds lines are the same? When wagering, does not one have to compare the "fair odds" with the relative comfort level of those odds?

Most bettors feel more comfortable with wagering $100 to win on a 6/5 shot, and can't hardly bet more than $5 on 20-1 shots. Does the 20-1 shot and a $5 wager really mean I'm only 1/20th as confident?

In the FOY, that mule ALMOST FAMOUS was a great price, but not a great bet. The amount of my wager reflected that - a medium sized bet for me, only made slightly larger by his odds.

I agree with you that playing a line you have a stronger confidence in for a larger bankroll is wise and lines that are shaky should be played on a smaller bankroll(maybe 2/3rds of the former). The ration is a personal choice that long term data will provide the answer to.

Not sure where you derived that I want my odds line close to the public's line. However I know that if I make the 1) 3-1 and he is paying 8/5 and I make the 4) 3-1 and he is 14-1, there is not way the they both have a 25% chance of winning as I projected. By bringing the 4's odds up to 7-1 I am reducing his chance of winning to 12.5%. I am still betting him, just not betting him for as much money as I would if he was 4-1. It is a little illogical, but I am fairly certain that virtually everyone's long term data will show this is the right thing to do. Also as I mentioned in the other post, if I know something I feel most do not, I will "step up and be bold". You mentioned that you are stronger at interpreting development of young horses than the public which gives you a big edge over the public. I would consider this to be grounds to leave the horse 3-1 and not bring him up to 7-1. Just make sure your huge overlay 3-1's in these types of races are indeed popping at a 25% win rate.

[sjk]

The dead on the board thinking never enters into my mind. The better the price I am getting on an overlay the better I like it and the more I bet.

For my situation the horse going off at 14-1 does not win as often as the one going off at lower odds but the extra payoff more than compensates for the reduction in frequency of cashing.

I owned horses for 20 years and never bet enough on one (I generally passed races I had a horse running in) to change the odds. The trainers who worked with me never gave me the impression they expected to win at the windows and generally did not bet. I think the inside money business is overrated.

While it is true that an odds line could be inaccurate because it does not have access to inside information I think it is likely the answer is that the odds line does not properly account for and weight the public information.

Everyone knows that horses who can outsprint their competition and get clear have a greatly enhanced chance of winning so if your line does not correctly identify and reward horses who can do this your results will be impaired. There are a multitude of others items that must be treated properly.

It is easier to make a bad odds line (they are all 6-1) than a good one and it is prudent to evaluate and challenge your line to make it better. In the end you want to get to where you think you have properly accounted for everything you can think of and you have blind confidence in your bets.

[Poindexter]

If one has a fairly consistent contender selection method, and one knows what the hit rate of his top, 2nd, 3rd, 4th contenders are, at different field sizes, and race types, couldn't one calculate the win probabilities of each of those selections, convert it to odds and then fudge it a point or 2 to get a fair odds line?

Yeah, I know, most people don't do a good job of record keeping, and therefore couldn't create a very good line, but suppose someone actually did perform that very important task, record keeping, just for grins.

Rabyo, I think the answer is yes with some form of accuracy. As I think was mentioned earlier, you do miss out on the occasion horse you might make out at 15-1 who might be at 40-1(and no I do not consider a 20-1 morning line at 40-1 as dead on the board-longshots tend to be ignored and drift up, there is no attraction for people to bet on them)

I think you are very correct in using field size as a variable(I would group the field sizes as something like A) 6 or less B) 7-8 C) 9-10 D)11+
I would also classify your A, B, C, D choices something like
A(0) Complete Standout
A (2) Strong A who looks clearly best
A(4) Good A who best in open field
A(5) Someone must get nod

I left the A(1) and A(3) open for horses that fit in between.

then do something similar with the B's C's......

It just seems like more work than actually coming up with a line for each race.. In a few days(when I get the time) I will post the model I use for races that I feel have too many unknowns to line. I just do not have it in my computer so I can't copy and paste it.

[lamboguy]

saturday Parx race 2 #8 HERE COMES DRZ leaves gate 6/5 crosses wire 2-1 and runs out.

no value to that horse though, i had him 3-1

[Hoofless_Wonder]

I agree with you that playing a line you have a stronger confidence in for a larger bankroll is wise and lines that are shaky should be played on a smaller bankroll(maybe 2/3rds of the former). The ration is a personal choice that long term data will provide the answer to.

Not sure where you derived that I want my odds line close to the public's line. However I know that if I make the 1) 3-1 and he is paying 8/5 and I make the 4) 3-1 and he is 14-1, there is not way the they both have a 25% chance of winning as I projected. By bringing the 4's odds up to 7-1 I am reducing his chance of winning to 12.5%. I am still betting him, just not betting him for as much money as I would if he was 4-1. It is a little illogical, but I am fairly certain that virtually everyone's long term data will show this is the right thing to do. Also as I mentioned in the other post, if I know something I feel most do not, I will "step up and be bold". You mentioned that you are stronger at interpreting development of young horses than the public which gives you a big edge over the public. I would consider this to be grounds to leave the horse 3-1 and not bring him up to 7-1. Just make sure your huge overlay 3-1's in these types of races are indeed popping at a 25% win rate.

Your posts seem to indicate that if there's a large difference in your odds line and the tote, you're assuming you've made a mistake - and your odds should be more aligned with the public. While I agree that a review of the odds to glean if the horse is "dead" on the board, or if other horses are being wagered where it "doesn't make sense", there are many races where the tote is just wrong. Yes, over time the tote will outperform most of us, but it smoothes the prices where one ends up wagering on only the smaller overlays, and not the big hay races.

Now I will concede that if I have two 4-1 shots on my odds line, and in one race the horse is 8-5 and in another the horse is 13-1, then yes - the lower odds horse will win more often. However, the 8-5 horse is not bettable, being a huge underlay. And I think we agree that factoring in our long-term strike rates with a particular type of race is useful in determining bet size, but don't we still need a "confidence factor" along with the odds line?

Back to my ill-fated FOY example with ALMOST FAMOUS, my 20% chance to win odds may have been wrong, but I don't believe that long-term results need to validate that - it's very difficult. I'm happy with my 20% odds line indicating a potential play. I do know that when I'm getting 13-1 in a difficult race type where I'll pick 15% winners long-term, I'm getting twice the odds, and that does provide for a margin of safety. If I lowered the odds line to 7-1, then again I'm on the cusp of passing on the race, even though being below his 10-1 ML indicates he might have some wise guy support. It was a tough race for me, even though the result was pretty chalky.

For those of you that do put in the effort to make an odds line for every race, and can use it with success, I tip my hat to you for being a better numbers guy than me. But I agree wholeheartedly with Thaskolas that sometimes the dead money ain't so dead, and that's the time to wager with both hands.

[thaskalos]

Thaskalos, I think the point you are trying to make goes something like this. If we had a magical computer that was able to analyze every race and give the exact probability of each entrant that was 100% accurate(sans the other stuff that comes up, 3 went lame since last start, 7 is juiced up today, 9 had a breathing operation since last start, 11 was sick a while back and was purposely held back last two so the connections could unload on him today, 12 out of the blue after being very sore the last 4 races suddenly feels great today), it probably would not do us that much good to have our perfect line, because our line is going to way underestimate the chances of the 3, 7,9, 11, 12. Moreover, the connection and insiders are all going to be sending in cash on these horses driving up the prices and giving us artificial value on the other 7 horses. We are going to be looking at our computers odds line and see the 5) 2-1 paying 7/2 and the 8)4-1 paying 7-1. So we are going to make our bets on these two horses and even though we think we are getting whopping value, chances are we are not getting any value at all. In this instance you are absolutely right. That being said, because we demand a premium, even with all that is against us in this race, we might not be at that much of a disadvantage.

This is the game we play and if you are correct and there is that much working against us in every race than no matter what we do, we probably will never show a long term profit in this game no matter what we do(unless we are privy to the inside information-connections, expert tote watching…..)The only way anyone here finds out if you are right is to go through the process and see if they can make a profit long run. I still feel that you are better off making a line and betting value. The premium you demand helps protect you against the very thing you are warning against. Rules such as cutting your bet down when your horse looks dead board or passing the race when 3 horses are offering value and whatever else you may come up (when you develop a large sample of bets) with helps you find the “danger races” and hopefully helps get you in the black.


Regarding your point about passing on the dead board horse. Theoretically if long term data shows that these plays are losers, you are probably right, BUT, we are human and it is not easy to see a horse that we make 2-1, be dead on the board, pay 5-1 and run off by 5. So by betting half the amount we limit the exposure but still make a nice profit when we are right. This is a much better alternative to me. For those with the discipline to pass because there long term records show that a certain situation is not profitable, more power to them.

By the way I disagree that dead on the board horses are your biggest value situations. That has not been my experience at all. A lot of times horses are dead on the board for fairly obvious reasons or because someone else is getting hammered so much (not because everyone in the barn knows the horse is not sound).

One final note about line making is this. The reason you make a line is because sometimes for some reason when you line a race you are just fare away from 100% probability. Sometime you underestimate the group and you can be a low as all the horses adding up to .83 and other times(a more common experience) you are somewhere between 120 and 130 %. You really have to work to get your numbers ironed out. By the time you do the horse you originally made 2-1 is now 3-1. If you did not go through this process you would be thinking the horse was worth 2-1 when in actuality the fair value is 3-1. Making a line is simply the art of quantifying your opinion.

Making a line sounds great in theory. I myself have tried line-making many times in the past...with varying degrees of success. I bought line-making methods from seemingly well-meaning and "respected" individuals...who endeavored to transform me from a "gambler" to an "investor". No dice! The methods proved to be much worse than advertised. I also attempted line-making on my own...by using Barry Meadow's highly intuitive method of "if this race were run one hundred times...how many times would this horse emerge victorious". I felt quite uneasy making fine-line determinations by purely intuitive means...and I abandoned the project. I also tried the more "mathematical" line-making procedures...but I found those to be overly restrictive and uncompromising.

I now play without a betting line...concentrating mainly on the exotics. I create value not by designating accurate odds to particular horses...but by identifying profitable situations, usually through the exclusion of well-bet horses that I consider vulnerable. I find that I have more negative opinions about horses than I do positive ones...and the exotics seem, to me at least, to be the more appropriate field for my particular skills.

In short...I fully realize the advantages that an accurate betting line offers. But this style of betting hasn't worked out for me. If there are people out there who are using betting lines successfully...then the fault clearly lies with me.

[raybo]

Rabyo, I think the answer is yes with some form of accuracy. As I think was mentioned earlier, you do miss out on the occasion horse you might make out at 15-1 who might be at 40-1(and no I do not consider a 20-1 morning line at 40-1 as dead on the board-longshots tend to be ignored and drift up, there is no attraction for people to bet on them)

I think you are very correct in using field size as a variable(I would group the field sizes as something like A) 6 or less B) 7-8 C) 9-10 D)11+
I would also classify your A, B, C, D choices something like
A(0) Complete Standout
A (2) Strong A who looks clearly best
A(4) Good A who best in open field
A(5) Someone must get nod

I left the A(1) and A(3) open for horses that fit in between.

then do something similar with the B's C's......

It just seems like more work than actually coming up with a line for each race.. In a few days(when I get the time) I will post the model I use for races that I feel have too many unknowns to line. I just do not have it in my computer so I can't copy and paste it.

Poindexter,

I appreciate your response! And, I see what you're saying (which would require even more record keeping data than I am currently doing, ie; further annotating the "A"s by perceived strength within that group, then the same for the "B"s, and "C"s and "D"s), but your suggestions seem to be saying that you are placing all your emphasis on individual races, while I am only concerned with the long term. My thought is that my "A"s hit at a certain percentage of the time long term, my "B"s hit at a certain percentage of the time long term, etc.. Since we never know for sure which horse will win an individual race, those perceived strength horses might not produce any better than the others, especially when you consider the prices they pay. In other words, the hit rates are not the only factor that is important, but must also include the prices they pay, long term.

As we know, the supposed best predictor, the combined public, still loses money, by a good percentage, long term. So, if we hope to beat the public enough to make a net profit, we must focus our efforts in the long term, not on individual races.

That said, and you probably won't agree, but, imagine this scenario, using a flat bet wagering method (say a variable percentage of bankroll based on preset levels of net profit, or similar system) for all odds playable contenders, and that our win contender selection method is mechanical and automated (exactly the same process all the time):

Track specific scenario:
Aqu, dirt, sprint, 4yo and up, field size 7-8:

Contender "A" long term hit rate = .30 win probability
Contender "B" long term hit rate = .20 win probability
Contender "C" long term hit rate = .18 win probability
Contender "D" long term hit rate = .12 win probability

Odds to 1 formula: 1 / (WP / (1-WP)) (where "WP" is the long term "win probability")

Contender "A" projected odds = 2.33 to 1
Contender "B" projected odds = 4.00 to 1
Contender "C" projected odds = 4.56 to 1
Contender "D" projected odds = 7.33 to 1

Assume our "premium" for fair odds is 1.25, then our minimum betting odds would be:

Contender "A" = 2.33 x 1.25 = 2.92 (rounded to nearest whole number) = 3/1 odds
Contender "B" = 4.00 x 1.25 = 5.00 (rounded to nearest whole number) = 5/1 odds
Contender "C" = 4.56 x 1.25 = 5.69 (rounded to nearest whole number) = 6/1 odds
Contender "D" = 11.50 x 1.25 = 9.17 (rounded to nearest whole number) = 9/1 odds

Does this seem feasible to you, without regard to number of bets made, long term? Basically, we're demanding good prices on our investments, based on the perceived risk involved (obtained from our automated, daily updated, record keeping data). At least that's the way I see it anyway. Quite possible that I'm missing some things?

[traynor]

Accurate odds lines need more than the individual win rates. That is the reason so many attempts to calculate the percentages involved go astray--they only consider the individual entry's win rates while ignoring the fact that the win rates used fail to factor the effect of varying scenarios.

Simply stated, a win probability of 35% (based on past records) may seem reasonable. However, a substantial number of the races used to create that figure may have been races in which (for example) no other entry had a win probability greater than 20% (or some other similar qualifier). The actual win probability of the entry tagged as 35% may be 20% or less in the exact mix of horses in today's race.

Accurate odds lines are more complex than calculations based on individual win probabilities, because the individual win probabilities are based on widely scattered "confounding variables" (most notably the various win probabilities of the other entries in the races used to make the calculations).

It is not simple. If it were, anyone with a computer and a chunk of race data could put together a highly predictive odds line that would be an accurate representation of future events. Crunching numbers and putting a probability tag on a horse is easy. Putting an accurate probability tag on a horse is much less so.

[raybo]

Accurate odds lines need more than the individual win rates. That is the reason so many attempts to calculate the percentages involved go astray--they only consider the individual entry's win rates while ignoring the fact that the win rates used fail to factor the effect of varying scenarios.

Simply stated, a win probability of 35% (based on past records) may seem reasonable. However, a substantial number of the races used to create that figure may have been races in which (for example) no other entry had a win probability greater than 20% (or some other similar qualifier). The actual win probability of the entry tagged as 35% may be 20% or less in the exact mix of horses in today's race.

Accurate odds lines are more complex than calculations based on individual win probabilities, because the individual win probabilities are based on widely scattered "confounding variables" (most notably the various win probabilities of the other entries in the races used to make the calculations).

It is not simple. If it were, anyone with a computer and a chunk of race data could put together a highly predictive odds line that would be an accurate representation of future events. Crunching numbers and putting a probability tag on a horse is easy. Putting an accurate probability tag on a horse is much less so.

Here again, the emphasis is being placed on individual races, rather than on the long term. Trying to set an "accurate" odds line, for a specific race, is not my goal, and IMO, should not be anyone else's goal, unless one is only interested in the results he/she gets for that individual race. I think that is probably the downfall of the vast majority of bettors, excepting those that are skilled enough, and have all the other needed resources, to vary their bet sizes on a per race basis, ie; the whales and syndicates, and others who think they know the "true" value of a specific horse, in a specific race. They fail to admit to themselves that they DO NOT KNOW the "true" value of individual horses, in individual races, because anything can happen in a race, and any horse can have a poor performance in any race, regardless of how "dominant" and "ready" they appear to be in that race.

I'm not addressing the more extreme exotics, of course, because they are a different animal completely, and our results are necessarily more scattered throughout the spectrum. And individual bets become much more important, as any individual race could yield enough profit to make our whole meet, or year, or career (if we hit "the big one" or a few of the really good ones).

Intuitively, focusing on the short term will probably negatively affect the long term. Whereas, if we focus on the long term, using our long term record keeping data, we do a better job of positively affecting that long term, due to those dang "confounding variables", in individual races, some keep harping on. A horse race is just that, "A" horse race, and means very little in the scheme of things, unless of course, you're betting your mortgage payment on that one race. The combination of all the races you play is what makes you profitable or not.

[Robert Fischer]

different types of odds lines





Morning Line Odds - The Track Handicapper's morning line


Expected Odds Line - What You expect the actual odds to be. Your morning line.


Fair Value Odds Line - What You think the odds SHOULD be if they were accurate.


Actual Odds - the Public's actual money in the live pools.

[Poindexter]

Your posts seem to indicate that if there's a large difference in your odds line and the tote, you're assuming you've made a mistake - and your odds should be more aligned with the public. While I agree that a review of the odds to glean if the horse is "dead" on the board, or if other horses are being wagered where it "doesn't make sense", there are many races where the tote is just wrong. Yes, over time the tote will outperform most of us, but it smoothes the prices where one ends up wagering on only the smaller overlays, and not the big hay races.

Now I will concede that if I have two 4-1 shots on my odds line, and in one race the horse is 8-5 and in another the horse is 13-1, then yes - the lower odds horse will win more often. However, the 8-5 horse is not bettable, being a huge underlay. And I think we agree that factoring in our long-term strike rates with a particular type of race is useful in determining bet size, but don't we still need a "confidence factor" along with the odds line?



Back to my ill-fated FOY example with ALMOST FAMOUS, my 20% chance to win odds may have been wrong, but I don't believe that long-term results need to validate that - it's very difficult. I'm happy with my 20% odds line indicating a potential play. I do know that when I'm getting 13-1 in a difficult race type where I'll pick 15% winners long-term, I'm getting twice the odds, and that does provide for a margin of safety. If I lowered the odds line to 7-1, then again I'm on the cusp of passing on the race, even though being below his 10-1 ML indicates he might have some wise guy support. It was a tough race for me, even though the result was pretty chalky.

For those of you that do put in the effort to make an odds line for every race, and can use it with success, I tip my hat to you for being a better numbers guy than me. But I agree wholeheartedly with Thaskolas that sometimes the dead money ain't so dead, and that's the time to wager with both hands.

First off yes, I think confidence in your line should factor in to your bet sizing. (See my last post) Once again that is personal choice.

Second, I think you have misunderstood what I posted. If I make a horse 4-1 and he is 13-1, I am betting the horse, period. No ifs, ands or buts about it. All I am saying is that from a bet sizing standpoint, unless I think I am privy to something most of the public cannot see, I will treat him like a 6.5-1 shot. Also if he is morning line 6-1 opened 15-1 and never came anywhere near the 6-1 prior to closing at 13-1, then that I would consider a dead board horse in which case I would halve the reduced bet size. If his morning line was 10-1 and he opened at 15-1 before settling at 13-1, I would not consider that a dead board horse. It is sort of a subjective(feel) thing. Also, I personally take on value at all levels assuming I haven't lined the horse out(lining out means I have to give him some kind of chance to win, so I may come up with 30-1, but I wouldn't bet him at any price). If i think a 30-1 shot on my line is worth 30-1 and he is 80-1, I will bet him. Why not?

[Robert Fischer]

different types of odds lines




Morning Line Odds - The Track Handicapper's morning line


Expected Odds Line - What You expect the actual odds to be. Your morning line.


Fair Value Odds Line - What You think the odds SHOULD be if they were accurate.


Actual Odds - the Public's actual money in the live pools.



If I were going to go through the process:

1 - I'd take the ML and correct for obvious mistakes (polite odds for longshots, favorites that are clearly too high, etc..), and I'd now have the Expected Odds Line.

2 - I'd then apply any rare or proprietary insight and form the Fair Value Odds Line.

3 - I'd observe the live wagering to see the Actual Odds. I would then compare the Actual Odds to the Expected Odds Line. I'm looking for discrepancies in expected behavior only. (as opposed to any comparisons between Actual Odds and my Fair Value Odds Line).

4 - I would then adjust/smooth my own Fair Value Odds Line as necessary to reflect any unexpected Public behavior.

5 - Pass or Play + Wager Sizing

[Poindexter]

Making a line sounds great in theory. I myself have tried line-making many times in the past...with varying degrees of success. I bought line-making methods from seemingly well-meaning and "respected" individuals...who endeavored to transform me from a "gambler" to an "investor". No dice! The methods proved to be much worse than advertised. I also attempted line-making on my own...by using Barry Meadow's highly intuitive method of "if this race were run one hundred times...how many times would this horse emerge victorious". I felt quite uneasy making fine-line determinations by purely intuitive means...and I abandoned the project. I also tried the more "mathematical" line-making procedures...but I found those to be overly restrictive and uncompromising.

I now play without a betting line...concentrating mainly on the exotics. I create value not by designating accurate odds to particular horses...but by identifying profitable situations, usually through the exclusion of well-bet horses that I consider vulnerable. I find that I have more negative opinions about horses than I do positive ones...and the exotics seem, to me at least, to be the more appropriate field for my particular skills.

In short...I fully realize the advantages that an accurate betting line offers. But this style of betting hasn't worked out for me. If there are people out there who are using betting lines successfully...then the fault clearly lies with me.

It doesn't work for you and you have moved on that is what you should do. I would never try to persuade someone to continue on with something if it doesn't work for them. That would be silly. The fact you have found a better way is a good thing. The whole trying and failing probably gave you a better understanding of value then you perhaps know, and helped you get to where you are today or maybe not.

Regarding the "if this horse was run 100 times how many times would this horse win", I do not really subscribe to that theory or I should say look at it that way. To many confounding variables as you guys call them. I just sort of look at this horse is off a 6 month layoff, working in this manner for this trainer with this jock up.....he should be about 15-1. If he is not ready to win they could run this race 1000 times he is not going to win. The 15-1 just says that the chance of his being cranked up * the chance of him winning if he is, would give him about a 1 in 16 chance of winning. These are the types of horses you also have to watch the tote carefully on. If he is morning line 12, opens at 4-1 and drifts to 8-1, I would say he has a lot better chance than the initial 6.7% chance I gave him. If he is is morning line 12 and opens 25-1 never to see 15-1, he probably has less of a chance then the 6.7% I initially assessed.

[Poindexter]

Poindexter,

I appreciate your response! And, I see what you're saying (which would require even more record keeping data than I am currently doing, ie; further annotating the "A"s by perceived strength within that group, then the same for the "B"s, and "C"s and "D"s), but your suggestions seem to be saying that you are placing all your emphasis on individual races, while I am only concerned with the long term. My thought is that my "A"s hit at a certain percentage of the time long term, my "B"s hit at a certain percentage of the time long term, etc.. Since we never know for sure which horse will win an individual race, those perceived strength horses might not produce any better than the others, especially when you consider the prices they pay. In other words, the hit rates are not the only factor that is important, but must also include the prices they pay, long term.

As we know, the supposed best predictor, the combined public, still loses money, by a good percentage, long term. So, if we hope to beat the public enough to make a net profit, we must focus our efforts in the long term, not on individual races.

That said, and you probably won't agree, but, imagine this scenario, using a flat bet wagering method (say a variable percentage of bankroll based on preset levels of net profit, or similar system) for all odds playable contenders, and that our win contender selection method is mechanical and automated (exactly the same process all the time):

Track specific scenario:
Aqu, dirt, sprint, 4yo and up, field size 7-8:

Contender "A" long term hit rate = .30 win probability
Contender "B" long term hit rate = .20 win probability
Contender "C" long term hit rate = .18 win probability
Contender "D" long term hit rate = .12 win probability

Odds to 1 formula: 1 / (WP / (1-WP)) (where "WP" is the long term "win probability")

Contender "A" projected odds = 2.33 to 1
Contender "B" projected odds = 4.00 to 1
Contender "C" projected odds = 4.56 to 1
Contender "D" projected odds = 7.33 to 1

Assume our "premium" for fair odds is 1.25, then our minimum betting odds would be:

Contender "A" = 2.33 x 1.25 = 2.92 (rounded to nearest whole number) = 3/1 odds
Contender "B" = 4.00 x 1.25 = 5.00 (rounded to nearest whole number) = 5/1 odds
Contender "C" = 4.56 x 1.25 = 5.69 (rounded to nearest whole number) = 6/1 odds
Contender "D" = 11.50 x 1.25 = 9.17 (rounded to nearest whole number) = 9/1 odds

Does this seem feasible to you, without regard to number of bets made, long term? Basically, we're demanding good prices on our investments, based on the perceived risk involved (obtained from our automated, daily updated, record keeping data). At least that's the way I see it anyway. Quite possible that I'm missing some things?

You cannot lump all your races together because every race is different. One race I may have my top choice at 6/5 and my second choice at 2-1 and another race I might have my top choice at 7/2 and my 2nd and 3rd choices at 9/2 each. Obviously the former is a far less contentious race. That 6/5 "A" choice might win at 4/5. Obviously you are not going to play him at 4/5. But more importantly his win % was much higher than your typical win percentage on other A choices. Now if you have him down as an A1 then you have him lumped with your other A1s. Over time your A0's might win at a 50% rate and your a1's t a 40% rate and your a2's at a 30% rate and you A3's at a 25% rate and you a4s at a 20% rate....whatever the percentages might be. Wouldn't it make sense to track over time what % each variation of A choices are hitting them lumping them altogther which gives you no accuracy whatsoever. You might think you are betting value every time one of your A choices is 3-1, but maybe in actuality this "A" is an A4 which over time you have found only hit at a 20% rate. You would actually be betting an undervalued horse disguised as a value horse.

[traynor]

You cannot lump all your races together because every race is different. One race I may have my top choice at 6/5 and my second choice at 2-1 and another race I might have my top choice at 7/2 and my 2nd and 3rd choices at 9/2 each. Obviously the former is a far less contentious race. That 6/5 "A" choice might win at 4/5. Obviously you are not going to play him at 4/5. But more importantly his win % was much higher than your typical win percentage on other A choices. Now if you have him down as an A1 then you have him lumped with your other A1s. Over time your A0's might win at a 50% rate and your a1's t a 40% rate and your a2's at a 30% rate and you A3's at a 25% rate and you a4s at a 20% rate....whatever the percentages might be. Wouldn't it make sense to track over time what % each variation of A choices are hitting them lumping them altogther which gives you no accuracy whatsoever. You might think you are betting value every time one of your A choices is 3-1, but maybe in actuality this "A" is an A4 which over time you have found only hit at a 20% rate. You would actually be betting an undervalued horse disguised as a value horse.

Exactly. It is not that constructing an odds line cannot be accurate. It is that simplistic muddling around with a computer screen full of numbers is not a very useful way to go about constructing an odds line.

It is, however, a step up from: "If this race were run 100 times, how many times would this horse win?" followed by pulling a number out of one's (hat?) and affixing it as a value as if it means something other than a wild and highly subjective guess.

[raybo]

You cannot lump all your races together because every race is different. One race I may have my top choice at 6/5 and my second choice at 2-1 and another race I might have my top choice at 7/2 and my 2nd and 3rd choices at 9/2 each. Obviously the former is a far less contentious race. That 6/5 "A" choice might win at 4/5. Obviously you are not going to play him at 4/5. But more importantly his win % was much higher than your typical win percentage on other A choices. Now if you have him down as an A1 then you have him lumped with your other A1s. Over time your A0's might win at a 50% rate and your a1's t a 40% rate and your a2's at a 30% rate and you A3's at a 25% rate and you a4s at a 20% rate....whatever the percentages might be. Wouldn't it make sense to track over time what % each variation of A choices are hitting them lumping them altogether which gives you no accuracy whatsoever. You might think you are betting value every time one of your A choices is 3-1, but maybe in actuality this "A" is an A4 which over time you have found only hit at a 20% rate. You would actually be betting an undervalued horse disguised as a value horse.

I see what you're saying, but once again, you're trying to create fair odds on individual horses, in individual races. I'm not attempting to do that at all, but rather setting odds for ranked positions with a known, constantly varying, long term hit rate. My experience is that any of my top contenders could win any specific race, regardless of the field's makeup, just at varying hit rates during the long term. But, over that long term, those ranked positions have produced the long term hit rate from my record keeping data. So, if the long term average hit rate is a certain percentage, then my odds must jive with that hit rate or I cannot make net long term profit.

Sure the hit rate varies constantly, but in an automated system like this, so would your fair odds on those ranked positions. As the hit rate declines the odds required would increase, and vice versa. You might be absolutely right, I don't know, but I also don't know if anyone has ever used a system like this, and actually tracked it to determine if it has merit or not. If you have tried this and know it will not work, that's one thing, but if you haven't, maybe it's not so far fetched after all?

Sometimes my top pick is 20/1 (or higher), some times it's less than 1/1, wouldn't those times when the 30/1 actually wins even out the times that the odds are too low to bet and the horse wins anyway? Likewise for the other ranked positions. Just wondering out loud.

[Dark Horse]

I am trying to get a handle on how to attach value to contending horses.

For example, is there a formula to follow on whether a favorite is 6-5 or 2-1?

If it's a 10 horse field and I throw out 5 horses as non-contenders how do I know where to start with the horse I like best? Is it what I am willing to accept as "fair" odds? If I make my top pick 2-1, where do I go from there?

You need excel to do this instantly.

1) translate the line into the corresponding percentage for the winning expectation. 3/1 = 25%, 4/1 = 20%, etc.
2) transpose that percentage into what it would be if the total odds added up to 100% (instead of the 120% or so that the tote board adds up to).
3) If you toss out five horses, as in your example, you may be able to arb the rest of the field. For this you need an arb calculator, which you can also create in excel. It will show you how much to bet on each horse to end up with the same profit.
4) For greater precision, as for your top pick, you can then calculate an overlay to see if the bet has value or not.

You can't control it completely, because the odds will change after the race goes off, but resulting ups and downs should be 50/50 in the long run.

I attached a sample race from earlier today, part of my excel interface. The information you asked for is in the white section on the left (ML, tote, tote %, change in %, and 100%-based percentage). On the right, in the other white-ish section is the arb calculator, where any horse can be entered for changing numbers. (The other, colored stuff is not related to your question).

If you can't program in excel, here's an online arb calculator. Use the right odds. 6/5 = 120, 8/5 = 160, 3/1=300, etc. http://www.sbrforum.com/betting-tools/arbitrage-calculator/

[Sapio]

Unless you can determine that your odds line has a higher information content than the public odds, you will eventually and continually lose.

One should, at minimum, evaluate their favorite against the public favorite.

Thomas Sapio

[Poindexter]

I see what you're saying, but once again, you're trying to create fair odds on individual horses, in individual races. I'm not attempting to do that at all, but rather setting odds for ranked positions with a known, constantly varying, long term hit rate. My experience is that any of my top contenders could win any specific race, regardless of the field's makeup, just at varying hit rates during the long term. But, over that long term, those ranked positions have produced the long term hit rate from my record keeping data. So, if the long term average hit rate is a certain percentage, then my odds must jive with that hit rate or I cannot make net long term profit.

Sure the hit rate varies constantly, but in an automated system like this, so would your fair odds on those ranked positions. As the hit rate declines the odds required would increase, and vice versa. You might be absolutely right, I don't know, but I also don't know if anyone has ever used a system like this, and actually tracked it to determine if it has merit or not. If you have tried this and know it will not work, that's one thing, but if you haven't, maybe it's not so far fetched after all?

Sometimes my top pick is 20/1 (or higher), some times it's less than 1/1, wouldn't those times when the 30/1 actually wins even out the times that the odds are too low to bet and the horse wins anyway? Likewise for the other ranked positions. Just wondering out loud.

I would imagine if you keep good records(which you surely seem to) that you can easily answer your own question with a little bit of research. I assume you have 100's if not 1000's of races you have gone through. Obviously the higher you go up the odds ladder the more likely it is that your A, or b or c or whatever choice is offering value. I do not believe the win percent of all A horses in aggregate has anything to do with the exact price you would need on an A choice for all your A choices in that subset to make a long term profit. I think it is simply a matter of accumulating data and seeing how much profit you make and your roi when your A horse is 3-1 or higher, or 7-2 or higher all the way down the list....or horse d is 8-1 or higher or 9-1 or higher....Basically input data and analyze the data with a what if program and you will find the ideal strike point for an A, B, C, D or maybe there is no strike point. The answer should be at your fingertips.

[classhandicapper]

Unless you can determine that your odds line has a higher information content than the public odds, you will eventually and continually lose.

One should, at minimum, evaluate their favorite against the public favorite.

Thomas Sapio

I agree. This is one test I've done many times.

It has always been my contention that if I can't pick the winner close to as often as the public, I am probably making too many mistakes to win.

In almost all those tests I picked the winner as often, but since I didn't always agree with the favorite, I had higher average prices. But even that result indicates I am making a lot of mistakes. I just proved that so is the public.

[Hoofless_Wonder]

First off yes, I think confidence in your line should factor in to your bet sizing. (See my last post) Once again that is personal choice.

Second, I think you have misunderstood what I posted. If I make a horse 4-1 and he is 13-1, I am betting the horse, period. No ifs, ands or buts about it. All I am saying is that from a bet sizing standpoint, unless I think I am privy to something most of the public cannot see, I will treat him like a 6.5-1 shot. Also if he is morning line 6-1 opened 15-1 and never came anywhere near the 6-1 prior to closing at 13-1, then that I would consider a dead board horse in which case I would halve the reduced bet size. If his morning line was 10-1 and he opened at 15-1 before settling at 13-1, I would not consider that a dead board horse. It is sort of a subjective(feel) thing. Also, I personally take on value at all levels assuming I haven't lined the horse out(lining out means I have to give him some kind of chance to win, so I may come up with 30-1, but I wouldn't bet him at any price). If i think a 30-1 shot on my line is worth 30-1 and he is 80-1, I will bet him. Why not?

Rereading your posts, if I understand your approach correctly, you'll adjust your odds line based somewhat on the public odds, and whether or not those odds reflect your own inside knowledge or other wagering patterns in the race - something like a confidence factor. In the example above you're lowering your odds to 6.5 to 1, which cuts the amount wagered. Yet there are times when you would leave the odds alone, and make a larger wager, based on increased confidence that your number is accurate. I'm looking forward to your future posts on the models you use.

My approach is to set a more static odds line, then determine the amount wagered based on the overall confidence in those odds for the race. So, my 13-1 shot remains at a 20% to win, but my confidence is lower - knowing I'll lose at least 80% of the time. So maybe this really isn't the correct definition of an odds line.

As for betting a 80-1 shot you give a 3% chance at winning - yes, that seems mathematically sound, but you've got to be very confident in your line.

It's when my confidence is higher on a particular horse, say 35-50%, that I usually try to increase my wager. Today's Swale is a great example of what not to do. I initially liked :3: No Nay Never, and singled him on a couple of cheap multi-race tickets. But when 9th race opened for wagering, the odds on him were ridiculously low (I think he opened at 1-5, and went off at 2-5), so I took another look at the race. With his long layoff, and today's race probably not being the major goal, could today's race be just a workout?

At the risk of being accused of redboarding, I ended up with the following odds line:

:3: 1-2 (tote odds 2-5)
:5: 2-1 (tote odds 9.5 to 1)
rest of field: 7-1

With :5: Spot going off at 9-1, I should have been making a large wager - but I couldn't get past thinking the :3: dog would still win, the exacta would be too small, and being a bit scared that the odds on the :5: were just too generous. Usually I pass a race when the odds on top horse are so low, but the tote was tempting me. So, I wagered on a couple of cheap pick threes with my key pony in the last race finishing mid-pack, and placed a small bet on the :5: to win at the last second. I still can't believe I didn't play what turned out to be a great exacta payoff.

Grrgh. Having whiffed in the Gotham and Palm Beach, I'm even over the last couple of weeks in 3YO stakes, but still looking forward to them over the next few months.

[Fox]

One might conclude that someone somewhere knows something that goes completely over the heads of the "value seekers" and is profiting handsomely from that knowledge. Profit is good. Chasing rainbows is bad.

Unless, of course, one is only talking about it, and writing about it, rather than actually doing it.

Are you one of the enlightened few?

[lamboguy]

Unless you can determine that your odds line has a higher information content than the public odds, you will eventually and continually lose.

One should, at minimum, evaluate their favorite against the public favorite.

Thomas Sapio
very sound advice

[classhandicapper]

Making a line sounds great in theory. I myself have tried line-making many times in the past...with varying degrees of success. I bought line-making methods from seemingly well-meaning and "respected" individuals...who endeavored to transform me from a "gambler" to an "investor". No dice! The methods proved to be much worse than advertised. I also attempted line-making on my own...by using Barry Meadow's highly intuitive method of "if this race were run one hundred times...how many times would this horse emerge victorious". I felt quite uneasy making fine-line determinations by purely intuitive means...and I abandoned the project. I also tried the more "mathematical" line-making procedures...but I found those to be overly restrictive and uncompromising.

I now play without a betting line...concentrating mainly on the exotics. I create value not by designating accurate odds to particular horses...but by identifying profitable situations, usually through the exclusion of well-bet horses that I consider vulnerable. I find that I have more negative opinions about horses than I do positive ones...and the exotics seem, to me at least, to be the more appropriate field for my particular skills.

In short...I fully realize the advantages that an accurate betting line offers. But this style of betting hasn't worked out for me. If there are people out there who are using betting lines successfully...then the fault clearly lies with me.

I think all these methods can probably work, but at the end of the day, when I made lines, almost all my plays wound up being in races where I had a specific piece of information that made me think a particular horses was either better or worse than public perceptions. So I just stopped making the lines or ranking the horses. Now I immediately focus my attention on the horse I know something special about and more or less intuit the fair odds on him without worrying about the appropriate odds on everyone else.

[raybo]

Are you one of the enlightened few?

According to many of his posts, yes he is, at least in his own mind anyway. :lol:

[Robert Fischer]

I think all these methods can probably work, but at the end of the day, when I made lines, almost all my plays wound up being in races where I had a specific piece of information that made me think a particular horses was either better or worse than public perceptions. So I just stopped making the lines or ranking the horses. Now I immediately focus my attention on the horse I know something special about and more or less intuit the fair odds on him without worrying about the appropriate odds on everyone else.
We have adapted somewhat similar approaches.
I demonstrated the basics of my odds-line process in post #70 of this thread.
No one is currently paying me to do this process, so I do it in my head.
Step 2 knocks out most of the races:

2 - I'd then apply any rare or proprietary insight and form the Fair Value Odds Line.
The special information is what allows you to form the Fair Value Odds Line. With the absence of any special insight, I have to assume that for all intents and purposes, the Actual Odds are "fair", which rules out focusing on that race as a key wagering race. Obviously, that race is still "live" for use in a multi-race sequence if the value in another leg is substantial.

[PaceAdvantage]

According to many of his posts, yes he is, at least in his own mind anyway. :lol:This has been a nice thread so far...plenty of stuff to think about and consider. Let's hope it doesn't degrade into a war of words between you and traynor. In fact, let's advise all moderators to not let this happen...by force if necessary.... :lol: :ThmbUp:

[raybo]

This has been a nice thread so far...plenty of stuff to think about and consider. Let's hope it doesn't degrade into a war of words between you and traynor. In fact, let's advise all moderators to not let this happen...by force if necessary.... :lol: :ThmbUp:

I got ya! I'll mind my manners.

[Overlay]

I think all these methods can probably work, but at the end of the day, when I made lines, almost all my plays wound up being in races where I had a specific piece of information that made me think a particular horses was either better or worse than public perceptions. So I just stopped making the lines or ranking the horses. Now I immediately focus my attention on the horse I know something special about and more or less intuit the fair odds on him without worrying about the appropriate odds on everyone else.
When you use the term "something special", I don't think that that necessarily has to convey the idea of "inside information", or of something that is not accessible to the general public. The "specific piece of information" can be a positive performance attribute that is in plain sight, but that the public has ignored or undervalued because it has become oversold on an aspect of another horse's record which, even if it may be a positive attribute, has less weight or influence than the public is giving it credit for.

Nor have I found it necessary to rely on intuition in such cases. The use of statistical probabilities gives me greater confidence and precision in detecting and exploiting such opportunities, while not decreasing their wagering value.

[raybo]

When you use the term "something special", I don't think that that necessarily has to convey the idea of "inside information", or of something that is not accessible to the general public. The "specific piece of information" can be a positive performance attribute that is in plain sight, but that the public has ignored or undervalued because it has become oversold on an aspect of another horse's record which, even if it may be a positive attribute, has less weight or influence than the public is giving it credit for.

It could also mean that one has a set of combined publicly available data that it's very likely the public does not use, at least not enough of them to mean anything in the odds. This is why being able to manipulate the available data, to make unique data from it, is so important.

[Exotic1]

This has been a nice thread so far...plenty of stuff to think about and consider. Let's hope it doesn't degrade into a war of words between you and traynor. In fact, let's advise all moderators to not let this happen...by force if necessary.... :lol: :ThmbUp:

Very nice thread. Some smart and practical folks out there that I need to learn from.

[formula_2002]

I am trying to get a handle on how to attach value to contending horses.

For example, is there a formula to follow on whether a favorite is 6-5 or 2-1?

If it's a 10 horse field and I throw out 5 horses as non-contenders how do I know where to start with the horse I like best? Is it what I am willing to accept as "fair" odds? If I make my top pick 2-1, where do I go from there?

as Cole Porter wrote, EXPERIMENT..with about 500 races, Don't change one thing in your selection process.
if you do change any one thing..start all over, always with new races.
don't use previous races, always use races that have not been run.

you should find you will save your self a lot of money if you dont bet on those 500 races. if you do make any money,.. call me :)

[classhandicapper]

When you use the term "something special", I don't think that that necessarily has to convey the idea of "inside information", or of something that is not accessible to the general public. The "specific piece of information" can be a positive performance attribute that is in plain sight, but that the public has ignored or undervalued because it has become oversold on an aspect of another horse's record which, even if it may be a positive attribute, has less weight or influence than the public is giving it credit for.

Nor have I found it necessary to rely on intuition in such cases. The use of statistical probabilities gives me greater confidence and precision in detecting and exploiting such opportunities, while not decreasing their wagering value.

By "special" I also meant a specific insight that is not likely to be reflected on the board.

I'm sure it can work fine using statistical probabilities.

I elaborated earlier on why I stopped doing that. My estimated probabilities would often be different depending on whose speed figures and other information I was using. So I could never be sure in a very specific case whether the horse I was considering was actually an overlay or I was just using an inaccurate figure.

Now, I still look at figures, but they are almost never a source of a bet. If anything it's the other way around. If I feel confident a high profile figure maker has a specific figure wrong, I will be interested in horses coming out of that race (to bet on or against depending on the direction of the error). That would be a "special" insight.