Player's expectation in races where the post time favorite loses.

[Jeff P]

Below are JCapper Data Window results for horses other than the post time favorite in races where the post time favorite failed to finish first. The time period covered is ytd 2011 current through end of day Thur Nov 17, 2011.

From the query results, it can be seen that profitable play is possible - with no contender selection and/or no "handicapping" whatsoever - provided the player is able to zero in on races where the post time favorite fails to finish first.

This sample should come as no surprise to anybody. (I consider it to be similar in nature to samples showing that a positive roi is possible provided the player is able to to pick the horse that has the lead at the second call.)

That said, you might want to think twice before discarding these sample results as trivial.

There are many strategies for identifying races where the post time favorite is unlikely to win. It is possible to assess post time favorites using standard fare such as speed, pace, class, and form. Successful strategies may also involve human connections, trips, suitability to surface and distance, track bias, breeding, and even overreactions by the crowd to "black type" horses. More advanced strategies involve creating custom factors to assess the likelihood of a loss by a post time favorite. (Believe it or not this can be done with a fair degree of accuracy.)

The type of strategy involved is not important and the strategy need not be perfect. The goal is merely to create a reliable predictor for races where the post time favorite is statistically "up against it." Of course, the better the strategy - the better the player's result.

First, all starters in the database that aren't post time favorites - in races where the post time favorite failed to finish first:
Data Window Settings:
Connected to: C:\JCapper\exe\JCapper2.mdb
999 Divisor Odds Cap: None

SQL: SELECT * FROM STARTERHISTORY
WHERE (RANKODDS > 1
AND FINPOSFAV1 > 1)


Data Summary Win Place Show
------------- ---------- ---------- ----------
Mutuel Totals 462411.70 360236.30 329508.10
Bet -403950.00 -403950.00 -403950.00
------------- ---------- ---------- ----------
Gain/Loss 58461.70 -43713.70 -74441.90

Wins 28332 47331 68707
Plays 201975 201975 201975
PCT .1403 .2343 .3402

ROI 1.1447 0.8918 0.8157
Avg Mut 16.32 7.61 4.80




Next, the above data broken out by odds rank:

By: Odds Rank

Rank Gain Bet Roi Wins Plays Pct Impact
---- -------- -------- ------ ---- ----- ----- ------
1 0.00 0.00 0.0000 0 0 .0000 0.0000
2 17848.30 56086.00 1.3182 9613 28043 .3428 2.4437
3 13943.40 56694.00 1.2459 6544 28347 .2309 1.6457
4 12874.40 56618.00 1.2274 4619 28309 .1632 1.1632
5 9527.00 55984.00 1.1702 3137 27992 .1121 0.7989
6 6495.70 53100.00 1.1223 2055 26550 .0774 0.5518
7 3314.90 44928.00 1.0738 1194 22464 .0532 0.3789

8 -317.80 33612.00 0.9905 646 16806 .0384 0.2740
9 -1457.90 22782.00 0.9360 315 11391 .0277 0.1971
10 -2169.00 13922.00 0.8442 133 6961 .0191 0.1362
11 -630.70 6212.00 0.8985 55 3106 .0177 0.1262
12 -929.60 3326.00 0.7205 18 1663 .0108 0.0772
13 197.00 452.00 1.4358 3 226 .0133 0.0946
14 -214.00 214.00 0.0000 0 107 .0000 0.0000
15 -6.00 6.00 0.0000 0 3 .0000 0.0000
16 -6.00 6.00 0.0000 0 3 .0000 0.0000
17 -4.00 4.00 0.0000 0 2 .0000 0.0000
18 -2.00 2.00 0.0000 0 1 .0000 0.0000
19 -2.00 2.00 0.0000 0 1 .0000 0.0000





-jp

.

[Dave Schwartz]

What point would you make from this, Jeff?

Do you find these results surprising?

I mean, you have (effectively) removed 37% of all your wagers - and they were all losers.

[classhandicapper]

The one problem with this approach is that most vulnerable favorites are not total throw outs with no chance to win. They are horses with some flaws and question marks. So even though they may be over bet, some still win. IMO you rarely find a favorite that's so bad you can profit betting the rest of the field. You usually still have to come up with good horse at good value.

I'm not saying this is a bad way to play. It's a great way to play, but producing profits is not as easy as this makes it look.

[Capper Al]

The one problem with this approach is that most vulnerable favorites are not total throw outs with no chance to win. They are horses with some flaws and question marks. So even though they may be over bet, some still win. IMO you rarely find a favorite that's so bad you can profit betting the rest of the field. You usually still have to come up with good horse at good value.

I'm not saying this is a bad way to play. It's a great way to play, but producing profits is not as easy as this makes it look.

I agree. Vulnerable favorites -- What's old is new again. How will one know if it is a vulnerable favorite without knowing how and when to eliminate?

[bob60566]

I agree. Vulnerable favorites -- What's old is new again. How will one know if it is a vulnerable favorite without knowing how and when to eliminate?

They use there database.

[bob60566]

On all given races there is not one peice of data that is unknown to the wise guys or whales not problem to me.

As they say when in Rome do as the Romans do

[Dave Schwartz]

I "eliminate" about 14% of all horses below 7/2.

My goal with said horses is not to expect that they never win. That would be unrealistic.

Instead, the goal is that at the end of the day, a $2 flat bet on the ones I have eliminated will return a $net of around $1.40; that is, lose about 30% per wagered dollar.

In other words, I want to determine that there is "dead money" in the pool, because that will (logically) make my horses more playable.


Recently, I have been working on creating my own little "BetFair" approach in American racing. That is, deciding which of the low-odds horses can be bet against and then bet them along with whatever other contenders I have.

The more pool I have eliminated, the more advantage I show. The greater my advantage, the more I wager - spread across my 2,3,4 or even 5 horses. (Note that I only bet horses that are flat-bet profitable according to my numbers.)

I realized how well this was working when, after several hundred races, I found that the more money I bet in a race, the higher my $Net.

The "Full Kelly" on these wagers comes back with some pretty hefty amounts. Often as much as 40% of bank! No, I do not wager 40% of TOTAL BANKROLL. However, I will wager 70% of SESSION bankroll.

Imagine that I might have a race where I project 90% of the time one of my horses will win and I am going to get (say) 2/5 in the race. That is a 26% advantage, divided by odds of .4:1 = 65%!

(I actually max out at 50% of Session BR.)

The question is, "In those races, do I really win 90% of the time?"

The answer is, "No, but so close as to be a negligible difference." My 90% (and up) races right now combine for 88% winners, down only slightly from the projected sum of 92%.


Regards,
Dave Schwartz

[Dave Schwartz]

On all given races there is not one peice of data that is unknown to the wise guys or whales not problem to me.

I would disagree with that.

Oh, they certainly know what all the data is, but they don't use it all. Nobody can.

They have a very good model - better than mine. Their strengths are in making a strong "global" model. Their weaknesses can be seen in races which are a little different than the "average" race.

That is where we (lesser players - like you and me) can do very well. It has caused me to pass about 40% of all races I handicap. Of the 60% left, my average bet size is slightly over 25% of bank!


Regard,
Dave Schwartz

PS: If there is enough interest, I'd consider doing a show about this.

[Robert Fischer]

I like the Data Windows. Neat.

I recommend that everyone browsing this thread go over the data at least once. For some of us this may be pretty straightforward data interpretation, and for others it may or may not be a new concept.

Don't even worry about the strategy itself the first time, and whether you play that way - just read the stats for the data (remember baseball cards?) and see if it matches what you would expect.

Important to note that we are looking at the set of results data for when the Post Time Favorite diid NOT win.

[Capper Al]

They use there database.

They use their database to eliminate or bubble up a horse to the top of their contender list. This is exactly my point to the punters. A beginning player can save a lot of time and money if they focus their efforts according to the nature of the game. I respect the commercial products sold here, yet the advice given should be to the punters best interest. Matter of fact, if the punter learns form, trainer intent, and conjures up some eliminations of their own, they could use the commercial products to their best advantage. But first they have to learn the game.

[Capper Al]

On all given races there is not one peice of data that is unknown to the wise guys or whales not problem to me.

As they say when in Rome do as the Romans do

To me "Wise Guys" implies mob and/or inside cheating. "Whales" refers to syndicates. Are you using these terms differently?

[Capper Al]

Dave's replies are on the mark, no non-sense.

Dave,

I'm considering raising my minimum from odds of 4/1 to 5/1. Once a player goes for value, they are betting against the favorite which on average wins 1 out of 3. If the average punter hits 1 out 4 then the math follows:

Favorite loses 2/3 * 1/4 = 1/6 or odds of 5 to 1.

I do better than the average punter, but feel grounded by sticking with the basics. Of course, what is reality? I do play action bets and suck up the loss from my winnings.

[TrifectaMike]

Once a player goes for value, they are betting against the favorite...

Bullshit meter: Red Alert!

Value Betting

A person who wants to make money betting on horses needs to look for odds that contradict with his/her own probability estimates for a race. This is strictly a mathematical approach to betting. You do not necessarily need to believe in the horse you put your money on. As long as the odds given are better than the purely mathematical chance of winning the race, it is a value bet.

Mike (Dr Beav)

[TexasDolly]

"Recently, I have been working on creating my own little "BetFair" approach in American racing. That is, deciding which of the low-odds horses can be bet against and then bet them along with whatever other contenders I have."

Dave, when you get a minute please elaborate on the comment about finding which horse can be bet against and then betting on them.
Thank you,
TD

[Capper Al]

Bullshit meter: Red Alert!

Value Betting

A person who wants to make money betting on horses needs to look for odds that contradict with his/her own probability estimates for a race. This is strictly a mathematical approach to betting. You do not necessarily need to believe in the horse you put your money on. As long as the odds given are better than the purely mathematical chance of winning the race, it is a value bet.

Mike (Dr Beav)

You should know about BS. You dish a lot of it out here. But, in theory, you are right in this case. Some favorites are under valued because players avoid low odds horses. In practice, your weakness, this is almost impossible to implement especially for the average punter. Splitting hairs between a horse going off at odds on money that really is worth 3 to 5 does happen. Being able to predict these before the race, forget it.

[Capper Al]

"Recently, I have been working on creating my own little "BetFair" approach in American racing. That is, deciding which of the low-odds horses can be bet against and then bet them along with whatever other contenders I have."

Dave, when you get a minute please elaborate on the comment about finding which horse can be bet against and then betting on them.
Thank you,
TD

I second that. I think I'm doing it, but would like to see what you have.

[Dave Schwartz]

I may have confused things a little with my wording.

What I meant was to determine which low-priced horse is the "target" that can be wagered against.

Then I take my other contenders and combine them in a dutch.

Some races do not contain a low-priced horse that can be wagered against. I have worked hard to find an alternative solution in races such as those. Sometimes I can bet JUST those low-priced priced horses.


Hope I said it better this time.

[PICSIX]

I may have confused things a little with my wording.

What I meant was to determine which low-priced horse is the "target" that can be wagered against.

Then I take my other contenders and combine them in a dutch.

Some races do not contain a low-priced horse that can be wagered against. I have worked hard to find an alternative solution in races such as those. Sometimes I can bet JUST those low-priced priced horses.


Hope I said it better this time.

Dave, I thought you flat bet your contenders?

[Dave Schwartz]

To me "Wise Guys" implies mob and/or inside cheating. "Whales" refers to syndicates. Are you using these terms differently?

I have never heard of a single "syndicate" anywhere that bets $100m per year.

[Dave Schwartz]

I used to - until I discovered that my COMBINED hit rate was accurate enough to use Kelly. Suddenly, I am a huge proponent of Kelly!

That would be another change in my thinking!

What always stood in the way oft hat in the past is my inability to build an accurate line. Now that I seem to have accomplished that, Kelly works.

Don't misunderstand me - were I not working at becoming a "high-limit" player, I would still use HorseMarket Investing (http://store.thehorsehandicappingauthority.com/products/Horse-Market-Investing.html) or a modified Opponent Method (http://store.thehorsehandicappingauthority.com/products/The-Opponent-Method.html) approach.

Things change when you wager at maximum in every race.

Dave

[Jeff P]

What point would you make from this, Jeff?Dave, my point is the same one you already articulated… that by identifying dead money in the pools, the player has the ability to improve his or her expectation (considerably) even with no “handicapping” or contender selection whatsoever.


My goal with said horses is not to expect that they never win. That would be unrealistic…Just so you know, I’m perfectly aware that 100% accuracy when it comes to elimination of dead money in the pools can never be achieved. However, I purposely structured the query driving the sample results posted above so as to show the player’s expectation in the event 100% accuracy were somehow achieved.


Do you find these results surprising?No, the data doesn’t surprise me. However, I do find it interesting that profitable play can be had (again with no handicapping and no contender selection) given strong enough accuracy in screening out dead money. To that end, I’ve been doing some R&D recently and have developed what appears to be a very good dead money screen – one that covers a lot of situations and is 100% mechanical.

I started this thread to see if others out there had been down this same road.


-jp

.

[Jeff P]

Recently, I have been working on creating my own little "BetFair" approach…I’m glad you mentioned Betfair.

Exchange wagering is often attacked by those intent on keeping to the status quo.

One argument that I hear repeated over and over again is that betting against horses leads to cheating.

What people making that argument fail to understand - or are afraid to admit - is the extent to which betting against horses successfully is possible in the pari-mutuel pools. (I think the data posted at the top of this thread confirms that.)



-jp

.

[Dave Schwartz]

Jeff,

IMHO, BetFair will not succeed in the U.S.

The only way it can is if there is a way to create a betting exchange that makes it possible and practical to bet exotics. How the heck would anybody be able to figure that out (except hi-tech guys like you and me, and of course, whales)?


Dave

[cj]

It is usually easier in my opinion to find a less heavily bet horse for elimination than the favorite. Get rid of a 5 to 1 shot and you are playing straight up.

[Jeff P]

Exotics should be easy enough to figure out...

My opinion is that fair pay for a 4-7 straight exacta in MTH R3 is $21.00. My opinion for a 4-5 straight exacta in that same race is $24.00... 4-2 $63.00... 4-6 $78.00.

Therefore I'd offer less than that to others.

Someone with a different opinion about that race snaps up my offers - or makes offers of their own on the same or other combos.

True, there are lots of combos in an exacta matrix. But programming an interface to handle that shouldn't be a game stopper.

There, we just figured it out.


-jp

.

[Dave Schwartz]

CJ,

I would say that you and I agree, except (maybe) with the frequency-thing.

IMHO, a single toss out at (say) 5/1 is not enough to warrant playing the race. Of course, I might be able to combine 2 or 3 such toss outs to make a playable race.


The majority of my plays are fueled by finding at least one horse under 7/2 that qualifies as a "bad bet" from my perspective.

The remainder of my plays are when there is no such horse and the two or three most-bet horses are (in my estimation) the only horses that are likely to win the race.

It is this second kind of play that is fueled by two or more toss outs at above 3/1.


Dave

[Dave Schwartz]

There, we just figured it out.

:lol:

The problem is the betting interface.

How does someone wade through the trifecta combinations in a 10-horse field to make a bet?

[PICSIX]

Dave, my point is the same one you already articulated… that by identifying dead money in the pools, the player has the ability to improve his or her expectation (considerably) even with no “handicapping” or contender selection whatsoever.


Just so you know, I’m perfectly aware that 100% accuracy when it comes to elimination of dead money in the pools can never be achieved. However, I purposely structured the query driving the sample results posted above so as to show the player’s expectation in the event 100% accuracy were somehow achieved.


No, the data doesn’t surprise me. However, I do find it interesting that profitable play can be had (again with no handicapping and no contender selection) given strong enough accuracy in screening out dead money. To that end, I’ve been doing some R&D recently and have developed what appears to be a very good dead money screen – one that covers a lot of situations and is 100% mechanical.

I started this thread to see if others out there had been down this same road.


-jp

.

Jeff, Have you looked at field size as a stand alone to see where the "Public" makes the biggest mistake (landing on the losing favorite).

Thanks,
Mike

[grant miller]

how close to post time can this work? we know that odds change once there off. great thread!

[Elliott Sidewater]

Jeff:

Interesting that when fav loses ROIs are positive for 2nd through 7th choice, only with a huge database can you uncover a nugget like that. On sheer intuition, I would have never guessed you can go that deep and still have a positive ROI. Is there something you could do with the trainer's win percentage to identify favorites that lose more often than the national average? Is less than 12% low enough, or is it something lower, like less than 10%? Or would this one factor not be predictive enough to matter?

Thanks,
Elliott

[lansdale]

Below are JCapper Data Window results for horses other than the post time favorite in races where the post time favorite failed to finish first. The time period covered is ytd 2011 current through end of day Thur Nov 17, 2011.

From the query results, it can be seen that profitable play is possible - with no contender selection and/or no "handicapping" whatsoever - provided the player is able to zero in on races where the post time favorite fails to finish first.

This sample should come as no surprise to anybody. (I consider it to be similar in nature to samples showing that a positive roi is possible provided the player is able to to pick the horse that has the lead at the second call.)

That said, you might want to think twice before discarding these sample results as trivial.

There are many strategies for identifying races where the post time favorite is unlikely to win. It is possible to assess post time favorites using standard fare such as speed, pace, class, and form. Successful strategies may also involve human connections, trips, suitability to surface and distance, track bias, breeding, and even overreactions by the crowd to "black type" horses. More advanced strategies involve creating custom factors to assess the likelihood of a loss by a post time favorite. (Believe it or not this can be done with a fair degree of accuracy.)

The type of strategy involved is not important and the strategy need not be perfect. The goal is merely to create a reliable predictor for races where the post time favorite is statistically "up against it." Of course, the better the strategy - the better the player's result.

First, all starters in the database that aren't post time favorites - in races where the post time favorite failed to finish first:
Data Window Settings:
Connected to: C:\JCapper\exe\JCapper2.mdb
999 Divisor Odds Cap: None

SQL: SELECT * FROM STARTERHISTORY
WHERE (RANKODDS > 1
AND FINPOSFAV1 > 1)


Data Summary Win Place Show
------------- ---------- ---------- ----------
Mutuel Totals 462411.70 360236.30 329508.10
Bet -403950.00 -403950.00 -403950.00
------------- ---------- ---------- ----------
Gain/Loss 58461.70 -43713.70 -74441.90

Wins 28332 47331 68707
Plays 201975 201975 201975
PCT .1403 .2343 .3402

ROI 1.1447 0.8918 0.8157
Avg Mut 16.32 7.61 4.80




Next, the above data broken out by odds rank:

By: Odds Rank

Rank Gain Bet Roi Wins Plays Pct Impact
---- -------- -------- ------ ---- ----- ----- ------
1 0.00 0.00 0.0000 0 0 .0000 0.0000
2 17848.30 56086.00 1.3182 9613 28043 .3428 2.4437
3 13943.40 56694.00 1.2459 6544 28347 .2309 1.6457
4 12874.40 56618.00 1.2274 4619 28309 .1632 1.1632
5 9527.00 55984.00 1.1702 3137 27992 .1121 0.7989
6 6495.70 53100.00 1.1223 2055 26550 .0774 0.5518
7 3314.90 44928.00 1.0738 1194 22464 .0532 0.3789

8 -317.80 33612.00 0.9905 646 16806 .0384 0.2740
9 -1457.90 22782.00 0.9360 315 11391 .0277 0.1971
10 -2169.00 13922.00 0.8442 133 6961 .0191 0.1362
11 -630.70 6212.00 0.8985 55 3106 .0177 0.1262
12 -929.60 3326.00 0.7205 18 1663 .0108 0.0772
13 197.00 452.00 1.4358 3 226 .0133 0.0946
14 -214.00 214.00 0.0000 0 107 .0000 0.0000
15 -6.00 6.00 0.0000 0 3 .0000 0.0000
16 -6.00 6.00 0.0000 0 3 .0000 0.0000
17 -4.00 4.00 0.0000 0 2 .0000 0.0000
18 -2.00 2.00 0.0000 0 1 .0000 0.0000
19 -2.00 2.00 0.0000 0 1 .0000 0.0000





-jp

.

Hi Jeff,

Much thanks for this valuable insight into the game. The reason I mention TM is that the path you're pursuing here seems similar to that mentioned by TM, about certain 'race types' as being inherently more profitable than others. Although he didn't refer to it directly, he seemed to be suggesting the use of the Bayesian technique of Harsanyi modelling. I hope he'll correct my error if I'm wrong. You may already be aware of this approach, but if not, it involves transforming a game of incomplete information (i.e. poker, horseracing,) into a game of imperfect but useable (profitable) information.

In this case, the 'player' whose game you've modelled is the 'losing favorite', to determine its 'payoff function'. And as you've pointed out, you don't need for this 'player' to lose every race, for your model to still be profitable.

The question I'd like to ask, or the query I'd like to see the results of, would involve the question that was being asked (and possibly has already been answered) by TM, about whether a very high percentage of 'chaos' or, I would add, 'contentious' races, are exactly the kind of race in which the favorite can profitably be bet against.

Thanks again for this and your many other contributions to the game.

Cheers,

lansdale

[CincyHorseplayer]

I'm not reading anybody else's responses Jeff.

The dominant part of my game is betting against suspect favorites.It's the only thing worth betting on.Period.

[Robert Fischer]

Jeff P,

Please check your pms.

[classhandicapper]

It is usually easier in my opinion to find a less heavily bet horse for elimination than the favorite. Get rid of a 5 to 1 shot and you are playing straight up.

I agree.

But you face the same problem here that to you do with bad favorites. These bad 5-1 shots are almost never bad enough to totally eliminate. Some will win. So you can't cover the entire take with a single insight like this.

It's also fairly easy to find horses that are 20-1 or 30-1 that should be 100-1 or 200-1. Eliminating horses like that doesn't cover much of the take, but sometimes in large fields there are 3 or 4 of them. Cumulatively, that helps a lot too

[Robert Fischer]

These bad 5-1 shots are almost never bad enough to totally eliminate.


this isn't true.

there are favorites all over the price range that you can eliminate (and when you start getting up to >5-1 they are more numerous.)

the odds of the horse has NOTHING to do with you eliminating it. You would just like to eliminate some of the chalk so that there is a reason to bet, so you don't start a race searching for 50-1 shots that you can toss with confidence...

If you want to complain about the system, say that the bad 1-5 shots are almost never bad enough to truly eliminate - not that the bad 5-1 shots are too infrequent. You act like all we know is the odds, and we have to win most of our races, and we have a real raw deal here...

[classhandicapper]

this isn't true.

there are favorites all over the price range that you can eliminate (and when you start getting up to >5-1 they are more numerous.)

the odds of the horse has NOTHING to do with you eliminating it. You would just like to eliminate some of the chalk so that there is a reason to bet, so you don't start a race searching for 50-1 shots that you can toss with confidence...

If you want to complain about the system, say that the bad 1-5 shots are almost never bad enough to truly eliminate - not that the bad 5-1 shots are too infrequent. You act like all we know is the odds, and we have to win most of our races, and we have a real raw deal here...

I don't think you are understanding my point at all. I'll try to be clearer.

I find horses I think are very overbet as favorites, 5-1 shots, and in every other odds category all the time, but I rarely find favorites or 5-1 shots I think are close to 0% to win.

The typical 5-1 shot might win between 13%-14% of the time. The ones I hate win way less than that. But they still win some of the time. So that's not nearly enough of an insight to create a positive expectation game on the rest of the horses. You still need to find a second inefficiency besides hating that 5-1 shot to be profitable.

The whole premise of this thread is how eliminating horses creates value on the other horses (which I agree with). I am simply arguing it's damn near impossible to find short priced horses with a 0% chance of winning. So it's harder to create enough value to win than simply finding a single over bet horse.

[Capper Al]

I predetermine if the horse is capable or not before the tote-board lights up. If I like him and the odds happen to be good, I'll play him. If the odds are low, I'll use him in my gimmicks with another higher odds horse that I like. If I have eliminated him, he's out no matter what the odds.

[Turkoman]

I predetermine if the horse is capable or not before the tote-board lights up. If I like him and the odds happen to be good, I'll play him. If the odds are low, I'll use him in my gimmicks with another higher odds horse that I like. If I have eliminated him, he's out no matter what the odds.

I agree. What you're saying makes a lot of sense to me, and takes us back to the basics. In other words, the handicapping should be done way before the race is about to begin. That way you'll already know who you like or not. Then, before they're off, you take it a step further, deciding who you really like and at what price.

Turkoman

[bob60566]

I would disagree with that.

Oh, they certainly know what all the data is, but they don't use it all. Nobody can.

They have a very good model - better than mine. Their strengths are in making a strong "global" model. Their weaknesses can be seen in races which are a little different than the "average" race.

That is where we (lesser players - like you and me) can do very well. It has caused me to pass about 40% of all races I handicap. Of the 60% left, my average bet size is slightly over 25% of bank!


Regard,
Dave Schwartz

PS: If there is enough interest, I'd consider doing a show about this.

My point exactly

Mac:)

[Jeff P]

Jeff, Have you looked at field size as a stand alone to see where the "Public" makes the biggest mistake (landing on the losing favorite).

Thanks,
MikeI have.

Below are results from my calendar year 2011 database - current as of a few days ago - post time favorites only - broken out by field size.

All post time favorites:
Data Summary Win Place Show
Mutuel Totals 78907.40 82748.80 83449.90
Bet -93532.00 -93532.00 -93532.00
Gain -14624.60 -10783.20 -10082.10

Wins 17234 26994 32743
Plays 46766 46766 46766
PCT .3685 .5772 .7001

ROI 0.8436 0.8847 0.8922
Avg Mut 4.58 3.07 2.55

Post time favorites broken out by field size:
Field Size Gain Bet Roi Wins Plays Pct Impact
1 0.00 0.00 0.0000 0 0 .0000 0.0000
2 0.20 6.00 1.0333 2 3 .6667 1.8091
3 1.10 126.00 1.0087 40 63 .6349 1.7229
4 -146.10 1102.00 0.8674 272 551 .4936 1.3396
5 -637.70 5548.00 0.8851 1259 2774 .4539 1.2316
6 -2230.90 14540.00 0.8466 2972 7270 .4088 1.1093
7 -2830.40 19102.00 0.8518 3673 9551 .3846 1.0436
8 -2705.40 17774.00 0.8478 3247 8887 .3654 0.9915
9 -2413.30 14106.00 0.8289 2413 7053 .3421 0.9284
10 -2071.10 11940.00 0.8265 1939 5970 .3248 0.8813
11 -705.70 4390.00 0.8392 693 2195 .3157 0.8567
12 -757.40 4232.00 0.8210 626 2116 .2958 0.8028
13 -29.60 380.00 0.9221 63 190 .3316 0.8998
14 -92.30 280.00 0.6704 35 140 .2500 0.6784
15 0.00 0.00 0.0000 0 0 .0000 0.0000
16 -2.00 2.00 0.0000 0 1 .0000 0.0000
17 -2.00 2.00 0.0000 0 1 .0000 0.0000
18 0.00 0.00 0.0000 0 0 .0000 0.0000
19+ -2.00 2.00 0.0000 0 1 .0000 0.0000

As you can see, field size has an impact on win rate for post time favorites. (The bigger the field, the lower the win rate.)



-jp

.

[Dave Schwartz]

Jeff,

I am really confused about the excitement in this thread. Please help me to understand.

Unless I have misunderstood, what you are (essentially) saying is that things go really well if the favorite doesn't win.

There has to be more... please fill in the blanks for me.


Dave

[Jeff P]

Jeff:

Interesting that when fav loses ROIs are positive for 2nd through 7th choice, only with a huge database can you uncover a nugget like that. On sheer intuition, I would have never guessed you can go that deep and still have a positive ROI. Is there something you could do with the trainer's win percentage to identify favorites that lose more often than the national average? Is less than 12% low enough, or is it something lower, like less than 10%? Or would this one factor not be predictive enough to matter?

Thanks,
Elliott

Trainer win pct does have an impact. But that impact isn't so strong as to shape the outcome so that it mimics the trainer's historical win percent.

From my Q4 2011 database...

First, all post time favorites:
Data Window Settings:
999 Divisor Odds Cap: None
Filters Applied: FAV=

Surface: (ALL*) Distance: (All*)
(From Index File: C:\2011\Q4_2011\pL_profile.txt)

Data Summary Win Place Show
Mutuel Totals 11763.20 12368.90 12630.90
Bet -14272.00 -14272.00 -14272.00
Gain -2508.80 -1903.10 -1641.10

Wins 2596 4050 4953
Plays 7136 7136 7136
PCT .3638 .5675 .6941

ROI 0.8242 0.8667 0.8850
Avg Mut 4.53 3.05 2.55

Next, the above data broken out by trainer win pct:
By: Trainer Win Pct

------- ------ ------- ------- ------- ---- ----- ----- ------
>=Min < Max Gain Bet Roi Wins Plays Pct Impact
------- ------ ------- ------- ------- ---- ----- ----- ------
-999.00 0.00 0.00 0.00 0.0000 0 0 .0000 0.0000
0.00 5.00 -350.10 1368.00 0.7441 214 684 .3129 0.8600
5.00 10.00 -288.80 1384.00 0.7913 223 692 .3223 0.8858

10.00 15.00 -498.80 2532.00 0.8030 436 1266 .3444 0.9467
15.00 20.00 -450.50 2978.00 0.8487 539 1489 .3620 0.9950
20.00 25.00 -287.50 2492.00 0.8846 504 1246 .4045 1.1119
25.00 30.00 -248.10 1782.00 0.8608 354 891 .3973 1.0921
30.00 35.00 -193.40 952.00 0.7968 183 476 .3845 1.0568
35.00 40.00 -59.70 326.00 0.8169 64 163 .3926 1.0793
40.00 45.00 -31.80 128.00 0.7516 24 64 .3750 1.0308
45.00 50.00 2.40 52.00 1.0462 14 26 .5385 1.4801
50.00 55.00 -58.90 140.00 0.5793 20 70 .2857 0.7854
55.00 60.00 -0.40 8.00 0.9500 2 4 .5000 1.3744
60.00 65.00 2.80 12.00 1.2333 4 6 .6667 1.8326
65.00 70.00 -5.20 24.00 0.7833 5 12 .4167 1.1454
70.00 75.00 0.00 0.00 0.0000 0 0 .0000 0.0000
75.00 80.00 1.20 6.00 1.2000 1 3 .3333 0.9163
80.00 85.00 2.00 4.00 1.5000 1 2 .5000 1.3744
85.00 90.00 0.00 0.00 0.0000 0 0 .0000 0.0000
90.00 999999.00 -44.00 84.00 0.4762 8 42 .1905 0.5236

-jp

.

[andymays]

And what happens if you're in P.B. with a bunch of wasted chicks screaming at your window?

Feel free to delete. LOL

[andymays]

And what happens if you're in P.B. with a bunch of wasted chicks screaming at your window?

Feel free to delete. LOL

Let me answer for you.

You let it ride!

[Jeff P]

Hi Jeff,

Much thanks for this valuable insight into the game. The reason I mention TM is that the path you're pursuing here seems similar to that mentioned by TM, about certain 'race types' as being inherently more profitable than others. Although he didn't refer to it directly, he seemed to be suggesting the use of the Bayesian technique of Harsanyi modelling. I hope he'll correct my error if I'm wrong. You may already be aware of this approach, but if not, it involves transforming a game of incomplete information (i.e. poker, horseracing,) into a game of imperfect but useable (profitable) information.

In this case, the 'player' whose game you've modelled is the 'losing favorite', to determine its 'payoff function'. And as you've pointed out, you don't need for this 'player' to lose every race, for your model to still be profitable.

The question I'd like to ask, or the query I'd like to see the results of, would involve the question that was being asked (and possibly has already been answered) by TM, about whether a very high percentage of 'chaos' or, I would add, 'contentious' races, are exactly the kind of race in which the favorite can profitably be bet against.

Thanks again for this and your many other contributions to the game.

Cheers,

lansdale

Great question!

I've been using RV (aka race volatility, chaos, degree of difficulty or any other name you want to call it) in my models since the early 1990's.

In most models, increased RV translates to increased player expectation.

The first time I looked at a data set with losing post time favorites removed broken out by field size I was a little surprised by what I saw.

The following sample will serve to illustrate.
Data Window Settings:
Connected to: C:\JCapper\exe\JCapper2.mdb
999 Divisor Odds Cap: None

SQL: SELECT * FROM STARTERHISTORY
WHERE (RANKODDS <> 1
AND FINPOSFAV1 > 1)


------------- ---------- ---------- ----------
Data Summary Win Place Show
------------- ---------- ---------- ----------
Mutuel Totals 465892.20 363099.00 332192.40
Bet -407098.00 -407098.00 -407098.00
Gain 58794.20 -43999.00 -74905.60

Wins 28546 47699 69236
Plays 203549 203549 203549
PCT .1402 .2343 .3401

ROI 1.1444 0.8919 0.8160
Avg Mut 16.32 7.61 4.80

The above sample broken out by field size:
---------- -------- -------- ------ ---- ----- ------ ------
Field Size Gain Bet Roi Wins Plays Pct Impact
---------- -------- -------- ------ ---- ----- ------ ------
1 0.00 0.00 0.0000 0 0 .0000 0.0000
2 1.40 2.00 1.7000 1 1 1.0000 7.1306
3 31.00 88.00 1.3523 22 44 .5000 3.5653
4 571.70 1630.00 1.3507 267 815 .3276 2.3360
5 3508.60 11786.00 1.2977 1469 5893 .2493 1.7775
6 9094.00 41726.00 1.2179 4163 20863 .1995 1.4228
7 13698.50 67790.00 1.2021 5656 33895 .1669 1.1899
8 10496.00 75940.00 1.1382 5435 37970 .1431 1.0207
9 8445.90 71288.00 1.1185 4468 35644 .1254 0.8938
10 7599.30 70232.00 1.1082 3930 35116 .1119 0.7980
11 2230.60 29006.00 1.0769 1459 14503 .1006 0.7173
12 2367.30 31824.00 1.0744 1442 15912 .0906 0.6462
13 516.90 2960.00 1.1746 123 1480 .0831 0.5926
14 229.30 2728.00 1.0841 108 1364 .0792 0.5646
15 0.00 0.00 0.0000 0 0 .0000 0.0000
16 16.90 30.00 1.5633 1 15 .0667 0.4754
17 -21.00 32.00 0.3438 1 16 .0625 0.4457
18 0.00 0.00 0.0000 0 0 .0000 0.0000
19+ 7.80 36.00 1.2167 1 18 .0556 0.3961

As you can see, successful elimination of a post time favorite in a small field creates a higher player expectation than successful elimination of a favorite in a big field.

Why is this so?

I believe it has to do with the amount of "dead money" (as a percentage of the total pool) identified by the player.

In a small field, the favorite typically has a higher percentage of the total pool than a favorite in a big field. Thus, successful elimination of the favorite in a small field results in more bang for the buck than it does in a bigger field.

Edit: However, that said, I find that large (confusing) fields offer the best wagering opportunities. (At least that's what my records tell me.)


-jp

.

[Jeff P]

Jeff,

I am really confused about the excitement in this thread. Please help me to understand.

Unless I have misunderstood, what you are (essentially) saying is that things go really well if the favorite doesn't win.

There has to be more... please fill in the blanks for me.


Dave

I'm crafting a strategy based on an accurate dead money eliminator.

Imho, such a thing has a lot of potential.

Nothing more. Nothing less.


-jp

.

[Aner]

Jeff,

I am really confused about the excitement in this thread. Please help me to understand.

Unless I have misunderstood, what you are (essentially) saying is that things go really well if the favorite doesn't win.

There has to be more... please fill in the blanks for me.


Dave

I think there is merit in identifying false favorites. I used one consistency factor to sort morning line favorites into strong and weak groups. Looking at races I have handicapped in November, I find the following:

Strong morning line favorites.

Races = 96
Wins = 36
frequency = 38%
betting loss = 7%

Weak morning line favorites.

Races = 122
Wins = 25
frequency = 20%
betting loss = 45%


Yes, an extremely small data base, but the trend is so strong it probably is real. I think one could key strong favorites in the first and second hole for vertical exotics. Weak favorites should be excluded in the win hole.

[Jeff P]

The one problem with this approach is that most vulnerable favorites are not total throw outs with no chance to win. They are horses with some flaws and question marks. So even though they may be over bet, some still win. IMO you rarely find a favorite that's so bad you can profit betting the rest of the field. You usually still have to come up with good horse at good value.

I'm not saying this is a bad way to play. It's a great way to play, but producing profits is not as easy as this makes it look.

Just so you know, I fully agree with the bolded part.

I never meant to imply that complete throw outs exist in the realm of post time favorites.

However, I do believe enough accuracy can be achieved so as to make something like this a great way to play.


-jp


.

[bob60566]

Jeff

If you find a false Favourite in your database for given race and are looking for value.
Would you then use your strong angle to get the horse that can get to the second call in the lead using your database to ID that horse for value.

Still handicapping.

[Dave Schwartz]

Have I missed the INDICATORS of a false favorite?

What I see is after the fact.


Again, I ask, "What have I missed?"

[thaskalos]

Have I missed the INDICATORS of a false favorite?

What I see is after the fact.


Again, I ask, "What have I missed?"

Dave is right...this has been common knowledge for many years.

If I am not mistaken, Dick Mitchell was the first to suggest that, if the race favorite could be safely eliminated from winning, long term profits were practically assured...no matter which contender the player chooses to wager on.

And even if we can't eliminate the favorite...safely eliminating ANY two or more horses with a combined win percentage of about 33% would accomplish the same thing.

Talking about it is one thing though...

Actually doing it is an altogether different matter.

I confess that I tried doing this a few years ago...and I failed miserably.

[Robert Fischer]

I don't think you are understanding my point at all. I'll try to be clearer.

I find horses I think are very overbet as favorites, 5-1 shots, and in every other odds category all the time, but I rarely find favorites or 5-1 shots I think are close to 0% to win.

The typical 5-1 shot might win between 13%-14% of the time. The ones I hate win way less than that. But they still win some of the time. So that's not nearly enough of an insight to create a positive expectation game on the rest of the horses. You still need to find a second inefficiency besides hating that 5-1 shot to be profitable.

The whole premise of this thread is how eliminating horses creates value on the other horses (which I agree with). I am simply arguing it's damn near impossible to find short priced horses with a 0% chance of winning. So it's harder to create enough value to win than simply finding a single over bet horse.

right , - a "fair" 5-1 shot will win 162/3%. - before we go and make it more complicated (you are already pulling a Hansen and sprinting out and leaving the class, Class). we can jump back to this convo, but if we start here im going to have to tell a story within a story within a story type of hypotheticals.

lets at least start with POST TIME FAVS since that is what Jeff querried for his little teaser that he shared. (like I said we can jump back ahead to a dif. "range" like 5-1 which will be the exact same premise only now we can use jeff's data window ) - also didn't exactly come across with the "tone" i had wanted to in my response to your post above..

The simple version - is that if POST TIME FAVS were winning @ 35%
(for the sake of this example .35 is just fine for a ballpark est.) Then 65% of the time you would have been in play for similar numbers to Jeff - and 35% of the time you would have LOST ALL. now we have a starting point. BIG PICTURE.
-Now you start to "crop" your big pic and take out the red-eye, and you will have endless online facebook babes.
For Jeffs data window he showed how you would be if you could have been exactly right on those ballpark35%. Maybe you could only get it down to 17.5% or HALF. And maybe you would specialize in ONLY playing some of the "odds-ranges" that Jeffs Data WINDOW show good feedback in - that overlap your FAVORITE ODDS RANGES TO JUDGE (this was kind of a weak re-combo of your 5-1 shots on my part but hey). Like if you love how you are uncanny at choosing quality amongst 4-1to6-1 horses you could simply play that range, you could query that range on jeffs window data, you could just use Jeffs data on page1 of this thread and assume that 5-1 may be around 3rd -4th Post time Odds choice or anything similar...
The more you do these things the "smaller your facebook mugshot" until you are just showing off the "assets"...
unfortunately u finally get a that facebook chick out on a "drink and a date" and you find out her slutty pics must have used quite a bit of photoshop filtering themselves :ThmbDown: wait whut?!?
bad metaphors/inet dating rants aside
I think you see where I am going... it's just a tool where you can get info on how the pools have been performing. For some it could help immensely if they were hampered with pitfalls.

The point isn't that we CAN beat EVERY POST TIME WINNING FAVORITE. The point (at least should be) that such querys are possible in the first place, that Jeffs program does such querys, and then any specific insights that these specific querys happen to share.

[Robert Fischer]

Have I missed the INDICATORS of a false favorite?

What I see is after the fact.


Again, I ask, "What have I missed?"

I think its probably just your experience as a horseplayer and your skill with software.

Initially, I was a little surprised that you began introducing your own product in post 7 of this thread. I read it again and it's clear that you did not do that as a sort of direct competition in Jeffs thread. This is just an interesting topic, and I think all the horseplayers are interested in the topic, we all work with the topic day in and day out in some form or fashion. And If we see numbers like this, the INSTINCT is for the BS meter to go off, and for our own opinions on the best way to approach such an animal - we all feel the need to pitch in and discuss.
One of the reasons Vulnerable Favorites has been a recurring popular thread here at PA.

To be blunt, I think that for some of the initiated horseplayers, these specific "data windows" will be "merely" a straight-forward data interpretation. - What have I missed?
However, probably the majority of players, in terms of numbers (total horseplayer population?) could benefit from looking over data windows like these. Very important to note that these are races where the post time favorite did NOT win. It would probably be shocking that the game isn't EASIER if we could see some of the backwards approaches that players throw their time into.
I think the demo here is an effective example but I'm not sure that the beginners even see it at all, and that a a lot of established horseplayers are comfortable with the premise to begin with. Really surprised that more middle-range players aren't bugging Jeff for further querys. And like you say the raw data isn't awe inspiring without the factors themselves.

[Dave Schwartz]

Robert,

So you are saying that the fact that other odds ranges do well when favorites lose is THE POINT of this thread?


Dave

[PICSIX]

Great question!

I've been using RV (aka race volatility, chaos, degree of difficulty or any other name you want to call it) in my models since the early 1990's.

In most models, increased RV translates to increased player expectation.

The first time I looked at a data set with losing post time favorites removed broken out by field size I was a little surprised by what I saw.

The following sample will serve to illustrate.
Data Window Settings:
Connected to: C:\JCapper\exe\JCapper2.mdb
999 Divisor Odds Cap: None

SQL: SELECT * FROM STARTERHISTORY
WHERE (RANKODDS <> 1
AND FINPOSFAV1 > 1)


------------- ---------- ---------- ----------
Data Summary Win Place Show
------------- ---------- ---------- ----------
Mutuel Totals 465892.20 363099.00 332192.40
Bet -407098.00 -407098.00 -407098.00
Gain 58794.20 -43999.00 -74905.60

Wins 28546 47699 69236
Plays 203549 203549 203549
PCT .1402 .2343 .3401

ROI 1.1444 0.8919 0.8160
Avg Mut 16.32 7.61 4.80

The above sample broken out by field size:
---------- -------- -------- ------ ---- ----- ------ ------
Field Size Gain Bet Roi Wins Plays Pct Impact
---------- -------- -------- ------ ---- ----- ------ ------
1 0.00 0.00 0.0000 0 0 .0000 0.0000
2 1.40 2.00 1.7000 1 1 1.0000 7.1306
3 31.00 88.00 1.3523 22 44 .5000 3.5653
4 571.70 1630.00 1.3507 267 815 .3276 2.3360
5 3508.60 11786.00 1.2977 1469 5893 .2493 1.7775
6 9094.00 41726.00 1.2179 4163 20863 .1995 1.4228
7 13698.50 67790.00 1.2021 5656 33895 .1669 1.1899
8 10496.00 75940.00 1.1382 5435 37970 .1431 1.0207
9 8445.90 71288.00 1.1185 4468 35644 .1254 0.8938
10 7599.30 70232.00 1.1082 3930 35116 .1119 0.7980
11 2230.60 29006.00 1.0769 1459 14503 .1006 0.7173
12 2367.30 31824.00 1.0744 1442 15912 .0906 0.6462
13 516.90 2960.00 1.1746 123 1480 .0831 0.5926
14 229.30 2728.00 1.0841 108 1364 .0792 0.5646
15 0.00 0.00 0.0000 0 0 .0000 0.0000
16 16.90 30.00 1.5633 1 15 .0667 0.4754
17 -21.00 32.00 0.3438 1 16 .0625 0.4457
18 0.00 0.00 0.0000 0 0 .0000 0.0000
19+ 7.80 36.00 1.2167 1 18 .0556 0.3961

As you can see, successful elimination of a post time favorite in a small field creates a higher player expectation than successful elimination of a favorite in a big field.

Why is this so?

I believe it has to do with the amount of "dead money" (as a percentage of the total pool) identified by the player.

In a small field, the favorite typically has a higher percentage of the total pool than a favorite in a big field. Thus, successful elimination of the favorite in a small field results in more bang for the buck than it does in a bigger field.

Edit: However, that said, I find that large (confusing) fields offer the best wagering opportunities. (At least that's what my records tell me.)


-jp

.

Jeff, this is the end of the pond I've been fishing in. We (seasoned horse players) tell new horse players to stay away from short fields. Bad advice in my opinion.

I'm not a database guy, so, thank you for posting this. I know that I can continue down this path.

Thanks,

Mike

[classhandicapper]

Just so you know, I fully agree with the bolded part.

I never meant to imply that complete throw outs exist in the realm of post time favorites.

However, I do believe enough accuracy can be achieved so as to make something like this a great way to play.


-jp


.

Agreed. At least half of the races I bet involve a short priced horse I don't like much.

[classhandicapper]

I have.

Below are results from my calendar year 2011 database - current as of a few days ago - post time favorites only - broken out by field size.

All post time favorites:
Data Summary Win Place Show
Mutuel Totals 78907.40 82748.80 83449.90
Bet -93532.00 -93532.00 -93532.00
Gain -14624.60 -10783.20 -10082.10

Wins 17234 26994 32743
Plays 46766 46766 46766
PCT .3685 .5772 .7001

ROI 0.8436 0.8847 0.8922
Avg Mut 4.58 3.07 2.55

Post time favorites broken out by field size:
Field Size Gain Bet Roi Wins Plays Pct Impact
1 0.00 0.00 0.0000 0 0 .0000 0.0000
2 0.20 6.00 1.0333 2 3 .6667 1.8091
3 1.10 126.00 1.0087 40 63 .6349 1.7229
4 -146.10 1102.00 0.8674 272 551 .4936 1.3396
5 -637.70 5548.00 0.8851 1259 2774 .4539 1.2316
6 -2230.90 14540.00 0.8466 2972 7270 .4088 1.1093
7 -2830.40 19102.00 0.8518 3673 9551 .3846 1.0436
8 -2705.40 17774.00 0.8478 3247 8887 .3654 0.9915
9 -2413.30 14106.00 0.8289 2413 7053 .3421 0.9284
10 -2071.10 11940.00 0.8265 1939 5970 .3248 0.8813
11 -705.70 4390.00 0.8392 693 2195 .3157 0.8567
12 -757.40 4232.00 0.8210 626 2116 .2958 0.8028
13 -29.60 380.00 0.9221 63 190 .3316 0.8998
14 -92.30 280.00 0.6704 35 140 .2500 0.6784
15 0.00 0.00 0.0000 0 0 .0000 0.0000
16 -2.00 2.00 0.0000 0 1 .0000 0.0000
17 -2.00 2.00 0.0000 0 1 .0000 0.0000
18 0.00 0.00 0.0000 0 0 .0000 0.0000
19+ -2.00 2.00 0.0000 0 1 .0000 0.0000

As you can see, field size has an impact on win rate for post time favorites. (The bigger the field, the lower the win rate.)



-jp

.

I have to tell you, that insight shocked the crap out of me.

I have one automatic play in my arsenal. It's partially field size dependent. But I need large fields. I think the reason it works is because even horrible horses often get bet down to 20-1 or 30-1 when they should be 100-1. That adds a lot of dead money to the pool. So this was surprising for me to see. Thanks.

[Red Knave]

So you are saying that the fact that other odds ranges do well when favorites lose is THE POINT of this thread?
Not every thread needs to have a point. Look at Chat if you don't believe me. :)

I'm following this thread to see where it goes. We posters are like witnesses to an accident. We all saw the same thing but we all describe it differently but yet all truthfully. Everyone, yourself included, has a different take on it.

Jeff made a statement backed up with solid data and he's stated that he believes he can determine when the favorite is dead money with some certainty. Not everyone can do that. It's not always obvious.

[Dave Schwartz]

Jeff made a statement backed up with solid data and he's stated that he believes he can determine when the favorite is dead money with some certainty. Not everyone can do that. It's not always obvious.

Knave,

Thank you!

That is what I was looking for and kept asking for! Apparently, I missed that. I will go back and try to understand what I missed.


Dave

[traveler]

That's the key Dave - you understand Canadian, eh! You owe Red a Tim Horton's coffee with a shot of his favourite "sweetener".

[Robert Fischer]

Robert,

So you are saying that the fact that other odds ranges do well when favorites lose is THE POINT of this thread?


Dave

Dave , no, I assume that the point of this thread was to introduce Jeff's JCapper Data Window Query Program.

[Robert Fischer]

surprised more guys aren't asking Jeff to run specific query.

Basically what Jeff has presented us, is the classic vulnerable favorite problem, and a program that can answer if we ask the right questions

[Dave Schwartz]

Basically what Jeff has presented us, is the classic vulnerable favorite problem, and a program that can answer if we ask the right questions

Robert,

I completely missed that. I assumed that this was an open discussion, regarding the Player's Expectations when the favorite loses.

That is why I kept asking questions.

LOL - I suppose that is why I had a hard time getting answers. Perhaps the thread could have been titled better.

Thanks for the heads up.


Dave

[bob60566]

]surprised more guys aren't asking Jeff to run specific query.[/b]

Basically what Jeff has presented us, is the classic vulnerable favorite problem, and a program that can answer if we ask the right questions

Ok is this a right question
How many vunerable favourites lose the first three races at any given track.

Mac:)

[overthehill]

if you can eliminate it from first can you just as easily eliminate it from second? if so how does boxing your contenders in exacatas compare to dutch betting them to win? historically my highest percentage returns have come from betting against a heavy favorite and hitting an exacta of some kind. going from memory the instances where this was successful were in dirt sprint stakes where the heavy chalk was drawn outside of the the other contenders and did not have early speed.

[Robert Fischer]

WHOOPS

Dave ,
re-read the section you quoted from me, and your reply.

In the section where you quoted me, i was talking about the program!, (something like - 'Jeff has presented us... with a program that can answer the classic vulnerable favorites problem if we ask the right questions'...)and there seems to have been an interpretation that I was talking about this discussion thread. Oh no! I would never moderate a line of questioning in this forum, pretty sure this confusion has a very low probability of ever happening again, but if it does, please send some PMs around! I hope you have caught it since then, as I am just checking back in 8am wed morning.

this is a good thread, i would hate for one of the sites more interesting contributors to refrain from asking a good question even for reasons however remotely related to something i may have posted.

Jay



surprised more guys aren't asking Jeff to run specific query.

Basically what Jeff has presented us, is the classic vulnerable favorite problem, and a program that can answer if we ask the right questions

Robert,

I completely missed that. I assumed that this was an open discussion, regarding the Player's Expectations when the favorite loses.

That is why I kept asking questions.

LOL - I suppose that is why I had a hard time getting answers. Perhaps the thread could have been titled better.

Thanks for the heads up.


Dave

[Dave Schwartz]

Robert,

Thank you for your response. I get it now.

Dave

[formula_2002]

I think it may be possible to identify those races where the favorite will win less than expected and ofcourse when it will more often than expected.

It's a bit of a slippery slope, but looking at the different tote board pools, seems to imply when the fav will have a better expectation of winning.

[lansdale]

For anyone who wants to run this, I'd be interested to know the ROI on all second and third favorites in races where the favorite is as described in subject line. Thanks for any help in advance.

[grant miller]

but when do you make your bet? 20 min to post? 10 mtp? 5mtp ? post time?--aint you betting blind at small handle tracks? otb money never shows untill 1/16 of the race has been run!

[Robert Fischer]

Thanks Dave
was kind of thinking aloud early in thread, and did a poor job expressing myself.

:ThmbUp:

Robert,

Thank you for your response. I get it now.

Dave

[Capper Al]

but when do you make your bet? 20 min to post? 10 mtp? 5mtp ? post time?--aint you betting blind at small handle tracks? otb money never shows untill 1/16 of the race has been run!

I do my best at 2 minutes to post.

[pondman]

Matter of fact, if the punter learns form, trainer intent, and conjures up some eliminations of their own, they could use the commercial products to their best advantage. But first they have to learn the game.

I'm a contrarian who goes against the morning line and the crowd in most of my bets. I think it's one of the few ways of beating the races for the average person. You've hit on an important point-- trainer intent. I'd also include the purse.

However, it sounds reasonable with enough resources, technology (direct access to the pool), and data a person could to skim from the pool. But it's not going to be suitable for most people.

[formula_2002]

but when do you make your bet? 20 min to post? 10 mtp? 5mtp ? post time?--aint you betting blind at small handle tracks? otb money never shows untill 1/16 of the race has been run!
the key is to set up the program to answer those questions.
currenty the program down loads the data at 3 to 4 minuets to post time.

[RichieP]

Very interesting stuff Jeff!

Over in the "Matchup" area of Pace and Cap I started posting a specific series of races couple of months ago (all posted at least a full day before they ran.). In the 18 races (run on all surfaces all over North America) that have run since the post time fav has won exactly three of them. Jim "The Hat" Bradshaw hated chalk and gave me clues what to "look" for to show their vulnerability.

I am pretty good at seeing them now but to be able to lay it out and "code it" is another story alltogether man. Congrats and best of luck in your endeavor :)

[RichieP]

Very interesting stuff Jeff!

Over in the "Matchup" area of Pace and Cap I started posting a specific series of races couple of months ago (all posted at least a full day before they ran.). In the 18 races (run on all surfaces all over North America) that have run since the post time fav has won exactly three of them. Jim "The Hat" Bradshaw hated chalk and gave me clues what to "look" for to show their vulnerability.



Make that 3 favs in 19 races after todays Grade 1 Clark Handicap posted up yesterday :). Three of the Matchers had the nice 9/2 winner nailed in a one horse pick, Mr. Hat would be proud :ThmbUp: