Betting lines from will pays

[imofe]

I was wondering if it is possible to make an odds line on horses from the P3 will pays. There is one track I play where the P3 will pays seem to be much sharper than the win board.

Here is an actual example. $1 P3 will pays
1) $97
2) $794
3) $346
4) $329
5) $72
6) $150
7) $794
8) $3010
9) $646
10) $210

Now the #5, who is the low P3, is 3/2 on the board. The #6 is 10-1 with less than a minute to post. It seems to me that if #5 is paying $5.0, the #6 should be paying around $ 10.4. #6 won the race and eventually paid $ 20.

What I would like to do is be able to make an odds line from the will pays and use it as a tool to be combined with my before the card handicapping. Is it possible to do this with the will pays?

[Dave Schwartz]

Sure. At least I think so.

First, check out this post (http://www.paceadvantage.com/forum/showpost.php?p=1493789&postcount=177).

In theory, you could convert these will-pays to $2 odds and plug them into that spreadsheet. The result would be the odds on the horses.

I have done this race for you in the graphic below.

http://www.horsestreet.com/BBSImages/AdjML-02-WillPays.jpg

[imofe]

Sorry Dave. Still a little lost. I understand the M/L section. But for the #6 how does the M/L of 74 turn into a Pcts. of 1.33 and then a Norm Pct. of 18.42 ?

[davew]

I will change headers to make more understandable for you
(Dave used a spreadsheet from another application I suspect)

M/L becomes PK3 will pays
Pcts becomes percentage of entire pool
Norm Pct becomes live percentage on this horse/entry
Adj ML becomes odds


This works alright unless a longer shot won one of the first legs

[imofe]

Okay. But if I do not have this chart handy, how do I figure out that the #6 has 1.33 % of the pool based on the M/L section?

[Dave Schwartz]

DaveW,

It still adjusts to the payoffs in the last leg. However, the higher the payoffs, the greater the distortion in the projected lines.

This was just something I threw together in about 20 minutes. Please do make whatever improvements you deem necessary and distribute as you see fit.

[Ocala Mike]

For the mathematically challenged, a relatively unsophisticated approach to this is to simply divide the will pays by the M/L. A prime example of the simplicity of this is to check the D/D will pays on a horse called ETHNIC DANCE a while back at Del Mar. Horse was 12/1 on the M/L, and was the SHORTEST double; won and paid $12.80, I believe.

[Dave Schwartz]

For the mathematically challenged, a relatively unsophisticated approach to this is to simply divide the will pays by the M/L. A prime example of the simplicity of this is to check the D/D will pays on a horse called ETHNIC DANCE a while back at Del Mar. Horse was 12/1 on the M/L, and was the SHORTEST double; won and paid $12.80, I believe.

Mike,

Oh, that's good


Dave

[thaskalos]

Mike,

Oh, that's good


Dave

Hey, Dave...

For a guy who claims to read virtually everything handicapping-related that comes out...you've sure missed a few things.

This daily-double theory goes all the way back to Al Illich...who thought of it in the '30s... :)

[imofe]

For the mathematically challenged, a relatively unsophisticated approach to this is to simply divide the will pays by the M/L. A prime example of the simplicity of this is to check the D/D will pays on a horse called ETHNIC DANCE a while back at Del Mar. Horse was 12/1 on the M/L, and was the SHORTEST double; won and paid $12.80, I believe.

I am not really a fan of this approach because of the morning line. Many times I see a 10/1 ML that I think deserves 5-1. When the P3 and board both show 5-1 there is no value here. But by using the 10-1 line it appears there is. I think it might be better to divide the will pays by the actual odds. But I was looking to make an odds line as soon as the P3 showed to avoid confusion near post. Of course in your example the horse was still 5-1 and was the lowest double, so value was there is you believe in that double pool over the win pool.

[Dave Schwartz]

Oh, I get that. I have that book and read it many years ago.

Perhaps I did not make it clear enough. I was thinking out loud I guess.

It was the idea that you could EASILY take ANY payoff and divide by ML to produce a metric. Then you take ALL of those metrics and average them together (or do something a little fancier) and produce an overall metric.

It just struck me in that moment.

[Ocala Mike]

Right, thas; I never meant to imply that it was my revelation.

The OP is correct, in that you need a pretty tight M/L; some are jokes.

Incidentally, I have found that lots of insights into the game originated in the 30's. Must have been a lot of guys with time on their hands during the depression looking to escape the bread lines. I sometimes use a dropdown "system" that my dad passed on to me that still (sometimes) works, and he learned it in the 30's.

[thaskalos]

Right, thas; I never meant to imply that it was my revelation.

The OP is correct, in that you need a pretty tight M/L; some are jokes.

Incidentally, I have found that lots of insights into the game originated in the 30's. Must have been a lot of guys with time on their hands during the depression looking to escape the bread lines. I sometimes use a dropdown "system" that my dad passed on to me that still (sometimes) works, and he learned it in the 30's.
I know that you didn't present the idea as your own.

I was just kidding around... :)

[Robert Fischer]

I was wondering if it is possible to make an odds line on horses from the P3 will pays. There is one track I play where the P3 will pays seem to be much sharper than the win board.

Here is an actual example. $1 P3 will pays
1) $97
2) $794
3) $346
4) $329
5) $72
6) $150
7) $794
8) $3010
9) $646
10) $210

Now the #5, who is the low P3, is 3/2 on the board. The #6 is 10-1 with less than a minute to post. It seems to me that if #5 is paying $5.0, the #6 should be paying around $ 10.4. #6 won the race and eventually paid $ 20.

What I would like to do is be able to make an odds line from the will pays and use it as a tool to be combined with my before the card handicapping. Is it possible to do this with the will pays?
How will reality work in favor of this theory ?

or When will the 'pick will-pays point to betting value in the current race?





There would have to be discrepancies between the pick will-pays, and the current race odds,
AND those discrepancies would have to be caused by superior wagering insight from the multi-race bettors.

I would guess, that in general, the current race odds are superior to the will-pays. Multi-race will-pays require coverage, and can depend in part on how the chalk does in the earlier legs.
There may be subsets in which the will pays do reflect superior wagering insight from the multi-race bettors.

[imofe]

Every track is different. But after watching this track for quite a while I believe that the P3 will pays are a better indicator than the win pools. Do not get me wrong. In most of the races the win pool is very similar to the P3 pool. But discrepancies do pop up like the one I mentioned in the initial post. There is another positive to looking for these kind of horses. It almost forces the bettor to look for value on the win board. Someone may think they like a horse that they feel will go off at a good price. When the horse goes off at something like 3-1 when the horse was a 10-1 ML some bettors are still inclined to make the wager. The P3 may indicate 4-1 and the horse is not worth a value bet anymore.

Anyway, I am not trying to convince anyone that this works or does not work. I just wanted to know if anyone had a way to turn the will pays into an odds line to use as another handicapping tool.

[Ocala Mike]

There may be subsets in which the will pays do reflect superior wagering insight from the multi-race bettors.



You're not looking for "superior wagering insight." The theory is that the "smart money, insider betting, stable money" (call it what you will) is going to have their charge covered in the multi-race pools rather than the single race pools in an attempt to:

1. Maximize their score.
2. Shield it from the betting public.

[TrifectaMike]

Every track is different. But after watching this track for quite a while I believe that the P3 will pays are a better indicator than the win pools. Do not get me wrong. In most of the races the win pool is very similar to the P3 pool. But discrepancies do pop up like the one I mentioned in the initial post. There is another positive to looking for these kind of horses. It almost forces the bettor to look for value on the win board. Someone may think they like a horse that they feel will go off at a good price. When the horse goes off at something like 3-1 when the horse was a 10-1 ML some bettors are still inclined to make the wager. The P3 may indicate 4-1 and the horse is not worth a value bet anymore.

Anyway, I am not trying to convince anyone that this works or does not work. I just wanted to know if anyone had a way to turn the will pays into an odds line to use as another handicapping tool.

Of all the exotic pools, the exacta pool is the most efficient estimater of win odds.

Mike

[Robert Fischer]

You're not looking for "superior wagering insight." The theory is that the "smart money, insider betting, stable money" (call it what you will) is going to have their charge covered in the multi-race pools rather than the single race pools in an attempt to:

1. Maximize their score.
2. Shield it from the betting public.

I think your "smart money" means the same thing as my "superior wagering insight". :ThmbUp:

[imofe]

Of all the exotic pools, the exacta pool is the most efficient estimater of win odds.

Mike

I completely agree with you Mike. Problem is that the exacta pool keeps changing. It can be very hard to keep up with last minute changes in the exacta payouts.

[imofe]

You're not looking for "superior wagering insight." The theory is that the "smart money, insider betting, stable money" (call it what you will) is going to have their charge covered in the multi-race pools rather than the single race pools in an attempt to:

1. Maximize their score.
2. Shield it from the betting public.

Yes and no. I suggest to use this to help your initial opinion of the race. Not as smart money on a horse I would never have bet, and now I will.

Here are a few examples in the last week that show how it helped me.

I was interested in playing a P4. In two of the races I thought the probable post time favorite was vulnerable. In the first leg I think it is a two horse race. Problem is that the multiples are getting too high to produce a acceptable ticket for my return. In the first leg, one of the two horses is 3/5 and the other 2-1. However, the P3 will pays have these two horses as just the opposite. The 2-1 should be 3/5. This allowed me to 1) play the ticket 2) actually get the better number in the first part of the P4 because people see the win board and are drawn to the 3/5.

The next example is a race where I have 3 contenders who can range in price from 5-1 to 15-1. Another race where I feel the favorite is vulnerable. These contenders have question marks. One is coming off a layoff. Another is a horse that I want a bet on only if the horse is leaving. The third is a dropper with questionable form but has raced against better. When the P3 will pays came up I noted that one of the horses was about 5-1 in the will pays and the other two were right at the odds line. The favorite was lukewarm in the P3 but bet on the win end. I got a $22.0 return on my contender who was 5-1 in the will pays.

Do all races work out so nicely? Of course not. Sometimes the horse you thought was vulnerable jogs. But you do not have to catch every race with these returns.

[Capper Al]

I was wondering if it is possible to make an odds line on horses from the P3 will pays. There is one track I play where the P3 will pays seem to be much sharper than the win board.

Here is an actual example. $1 P3 will pays
1) $97
2) $794
3) $346
4) $329
5) $72
6) $150
7) $794
8) $3010
9) $646
10) $210

Now the #5, who is the low P3, is 3/2 on the board. The #6 is 10-1 with less than a minute to post. It seems to me that if #5 is paying $5.0, the #6 should be paying around $ 10.4. #6 won the race and eventually paid $ 20.

What I would like to do is be able to make an odds line from the will pays and use it as a tool to be combined with my before the card handicapping. Is it possible to do this with the will pays?

The answer is yes. I have written a spreadsheet just for that. Instead of feeding pool amounts, inputting payout amounts will return odds also. I have been considering using it because watching the tote-board isn't working against late money flashes.

[imofe]

In my head I can roughly figure out what a horse should be in comparison to the low one. In the example I gave, the #6 is about twice as much as #5 and #10 is about 3 times as much as #5. But just how low should the low one be?

[Bennie]

I always look at the dd, pick3 and pick4 willpays to see if there are "live" horses. My friend likes to call these "hidden money" People who bet the horse in the multi race wagers leading into the race. If they are live in their wagers they pass on any win wagers. I myself use this in reverse. If I am live in my multi race wager, I look to see if someone was taking action but not really showing in the win pool. I might play an exacta box with the "live" horse and my top 2 selections. Just my thoughts.

[Ocala Mike]

I always look at the dd, pick3 and pick4 willpays to see if there are "live" horses. My friend likes to call these "hidden money" People who bet the horse in the multi race wagers leading into the race. If they are live in their wagers they pass on any win wagers.



:ThmbUp: :ThmbUp: :ThmbUp: :ThmbUp:

Bennie and I must have studied at the same handicapping school. It doesn't happen that often, but it is amazing what prices you can get on absolutely "live" horses. These are the situations where you hear a lot of "that double (or P/3 or P/4) came back less than a parlay" from those that weren't attentive.

[jdhanover]

With all this, you need to be careful of smallish P3 or P4 pools when there has been a longshot in the early race(s)....can see skewed payoffs (and be careful that some are not paying for 3/4 for example) that mean nothing. In bigger pools I suspect this may have some value but would need to be tracked and studied.

[Ocala Mike]

Agree with above post. The analysis of the will-pays is less effective when a ridiculous longshot or other "wrong horse" figures in the first half of a double or an early race of any multi.

Like everything else, however, there is one caveat, namely that you might pick up on a live horse that was "back-wheeled," obviously with the inclusion of the "wrong horse."

[Robert Fischer]

Like everything else, however, there is one caveat, namely that you might pick up on a live horse that was "back-wheeled," obviously with the inclusion of the "wrong horse."

when you say "back-wheeled", do you mean in current race exactas?

[Bennie]

I believe he means if your "key" horse is in the 6th race, you "backwheel" from that horse into the 5th and 4th race instead of starting your multi race with the 6th.

[Robert Fischer]

I believe he means if your "key" horse is in the 6th race, you "backwheel" from that horse into the 5th and 4th race instead of starting your multi race with the 6th.

that makes sense.

the wheeled 2nd in exactas thing may be useful to look at too, who knows..




look for differences in pk3 will-pay, and win pools
try to see if the horse in question would be of the "Smart Money" subset (first time starter, or trainer change, or any change that signals intent, or ?)
or you could skip the subsets idea and just bet every time you see a difference...

[Longshot6977]

In case someone reading this thread is wondering how the DD angle works or exacta angle works, here is a little portion of my toteboard spreadsheet. I do the same for exacta payoffs. Take a look at the payoff on the :5:. He has the lowest DD ratio of 7.0 while having a 15-1 M/L. He's 'live' in the DD, so a win/place bet could also be used on him.

[dkithore]

I completely agree with you Mike. Problem is that the exacta pool keeps changing. It can be very hard to keep up with last minute changes in the exacta payouts.


Do you know of any utility that could monitor the volatile changes in the tote-exacta will pays? any thoughts?

[imofe]

No, I do not. At many tracks, even the ones that have a good amount in the pools, the changes between just before and just after post can be significant. At tracks will smaller pools, the task becomes impossible. If you can find a track where the exactas do not change much after the race, I would make looking at those payoffs part of your handicapping. The reason I prefer the P3 is that those will pays are set in stone ( except for a late scratch) and you have plenty of time to use them in your decision making.

[dkithore]

No, I do not. At many tracks, even the ones that have a good amount in the pools, the changes between just before and just after post can be significant. At tracks will smaller pools, the task becomes impossible. If you can find a track where the Exactas do not change much after the race, I would make looking at those payoffs part of your handicapping. The reason I prefer the P3 is that those will pays are set in stone ( except for a late scratch) and you have plenty of time to use them in your decision making.
Thanks for the feedback. This sudden changes seem to be almost universal phenomena and eessp. for a US player-with no Betfair advantage of fixed prices, one is stuck. Last night at UK two of my win contenders dropped from 12 to 5-2 and one from 8 to 5/2. Again this phenomena is track specific. Less gyrations at Ascot and Epsom on big events days.

ADW TWINSPIRES OR THE HOST uk TRACKES DO NOT OFFER P3-DD OR P4. So options are limited. MY ONLY OPTION IS TO FIND A WAY TO PRODUCTS ALLOW FIXED PRICES.

Any thoughts or ideas by anyone??

Any ideas anybody.

[Segwin]

At the risk of being noobish what is a "will pay"?

[Dave Schwartz]

...This sudden changes seem to be almost universal phenomena and eessp. for a US player-with no Betfair advantage of fixed prices, one is stuck. Last night at UK two of my win contenders dropped from 12 to 5-2 and one from 8 to 5/2. Again this phenomena is track specific. Less gyrations at Ascot and Epsom on big events days.

ADW TWINSPIRES OR THE HOST uk TRACKES DO NOT OFFER P3-DD OR P4. So options are limited. MY ONLY OPTION IS TO FIND A WAY TO PRODUCTS ALLOW FIXED PRICES.

Any thoughts or ideas by anyone??

Any ideas anybody.

I have no clue how to do a better job with multi-race wagers, but for single race bets I have stopped using the tote board altogether beyond capturing scratches and prices.

This is working for me.

[Robert Fischer]

At the risk of being noobish what is a "will pay"?
Will Pays indicate the amount that you will win, if you are alive in a multi-race bet such as the pick-3.

This graphic shows the $2 Pick-3 will pays for yesterday's 8th race at Belmont.

http://www.fotoshack.us/fotos/70427willpays.jpg

5 ended up winning the the 8th, so if you had played 7-4-5, you had a big pick-3.

[Capper Al]

Will Pays indicate the amount that you will win, if you are alive in a multi-race bet such as the pick-3.

This graphic shows the $2 Pick-3 will pays for yesterday's 8th race at Belmont.

http://www.fotoshack.us/fotos/70427willpays.jpg

5 ended up winning the the 8th, so if you had played 7-4-5, you had a big pick-3.

You know what happened. Here's the odds as my program figures them from the payoffs:

1: 4.97
2: 7.84
3: 4.51

4: 2.82*
5: 18.65
6:

7: 3.45

[Segwin]

Will Pays indicate the amount that you will win, if you are alive in a multi-race bet such as the pick-3.

This graphic shows the $2 Pick-3 will pays for yesterday's 8th race at Belmont.

http://www.fotoshack.us/fotos/70427willpays.jpg

5 ended up winning the the 8th, so if you had played 7-4-5, you had a big pick-3.

Thank you Robert.

[DRIVEWAY]

You know what happened. Here's the odds as my program figures them from the payoffs:

1: 4.97
2: 7.84
3: 4.51

4: 2.82*
5: 18.65
6:

7: 3.45

If I multiply the mutuals in the 6th and 7th race, I come up with app. 153.

If I divide the pick 3 payoffs by 153 the results are.

1. 3.70
2. 5.49
3. 3,42
4. 2.37
5. 12.18
7. 2.76

Actual Odds were
1. 3.40
2. 8.60
3. 3.45
4. 2.35
5. 6.20
7. 3.95

Been doing it this way for a long time. Surprised by you oddsline. How are you doing it differently or should I say How am I doing it differently.

Thanks for your input

[tanner12oz]

I have always had the theory that the multirace wagers contain much more smart or at least informed money then the single race wagers

[AndyC]

Every track is different. But after watching this track for quite a while I believe that the P3 will pays are a better indicator than the win pools. Do not get me wrong. In most of the races the win pool is very similar to the P3 pool. But discrepancies do pop up like the one I mentioned in the initial post. There is another positive to looking for these kind of horses. It almost forces the bettor to look for value on the win board. Someone may think they like a horse that they feel will go off at a good price. When the horse goes off at something like 3-1 when the horse was a 10-1 ML some bettors are still inclined to make the wager. The P3 may indicate 4-1 and the horse is not worth a value bet anymore.

Anyway, I am not trying to convince anyone that this works or does not work. I just wanted to know if anyone had a way to turn the will pays into an odds line to use as another handicapping tool.

I charted pools and payoffs for years and would certainly not recommend trying to use P3 will-pays as an indication of value. Looking at Belmont today the biggest P-3 pool was $101K. Considering that will-pays are base on possible winning tickets you are basing a value on a very small amount of money. Usually 10% or less of the pool. Do you really want to hang your hat on what $10,000 or less of action indicates? Depending on the first 2 legs the amount of live money could be under $1,000. Keep in mind that the usual P-3 pool is far less than $101K.

[Capper Al]

If I multiply the mutuals in the 6th and 7th race, I come up with app. 153.

If I divide the pick 3 payoffs by 153 the results are.

1. 3.70
2. 5.49
3. 3,42
4. 2.37
5. 12.18
7. 2.76

Actual Odds were
1. 3.40
2. 8.60
3. 3.45
4. 2.35
5. 6.20
7. 3.95

Been doing it this way for a long time. Surprised by you oddsline. How are you doing it differently or should I say How am I doing it differently.

Thanks for your input

Let's go over this offline and see what we have.

[imofe]

I charted pools and payoffs for years and would certainly not recommend trying to use P3 will-pays as an indication of value. Looking at Belmont today the biggest P-3 pool was $101K. Considering that will-pays are base on possible winning tickets you are basing a value on a very small amount of money. Usually 10% or less of the pool. Do you really want to hang your hat on what $10,000 or less of action indicates? Depending on the first 2 legs the amount of live money could be under $1,000. Keep in mind that the usual P-3 pool is far less than $101K.

I can not speak for every track. Maybe at Belmont the P3 money is not better than the win board. But I did chart one track for 3 months and the P3 will pays were a much better predictor of who was going to win than the win pool. I have since gone back and used Dave's chart and the formula that was mentioned by Driveway. Regardless of which one I used, the P3 pool was the better pool. The reason I charted it is because I kept seeing hidden winners. So in all probability it has been going on longer than I have tracked it.

Again, I am not saying to use this as a complete handicapping formula. It should be used to compliment your existing handicapping.

[dkithore]

I can not speak for every track. Maybe at Belmont the P3 money is not better than the win board. But I did chart one track for 3 months and the P3 will pays were a much better predictor of who was going to win than the win pool. I have since gone back and used Dave's chart and the formula that was mentioned by Driveway. Regardless of which one I used, the P3 pool was the better pool. The reason I charted it is because I kept seeing hidden winners. So in all probability it has been going on longer than I have tracked it.

Again, I am not saying to use this as a complete handicapping formula. It should be used to compliment your existing handicapping.
A good share. Thx.

[AndyC]

I can not speak for every track. Maybe at Belmont the P3 money is not better than the win board. But I did chart one track for 3 months and the P3 will pays were a much better predictor of who was going to win than the win pool. I have since gone back and used Dave's chart and the formula that was mentioned by Driveway. Regardless of which one I used, the P3 pool was the better pool. The reason I charted it is because I kept seeing hidden winners. So in all probability it has been going on longer than I have tracked it.

Again, I am not saying to use this as a complete handicapping formula. It should be used to compliment your existing handicapping.

Shouldn't it work at every track? Which tracks did you use and what was the win percentage of the willpay favorite versus the win board? What was the average pool size of the P-3s you tested?

Yesterday at Churchill there were 8 P-3s with an average pool size of $14,884. Looking at the last P-3 with the largest pool of $25,207 I estimate that the amount of live money going into the final leg was about $325. I doubt that such a small amount of money would have more predictive accuracy than a win pool.

I would also would like to know how such information, if accurate, could compliment your handicapping outside of the instances where there are big discrepancies.

[imofe]

You have to go to my original post. I was using a calculation that I felt could be improved. So I asked what was the best way to turn the P3 into an accurate odds line. I am not trying to convince anyone that it works. If you feel what works at one track should work at all of them, then we have a difference of opinion. If you think what determines if this works or not is the money on live tickets, we also have a difference of opinion. It is the whole P3 pool from leg 1 to leg 3.

You will have to find out through research if this or other pools ( P4, exacta) outperform the win pool and if you can use it. I gave a few examples in an earlier post as to how it helped compliment my handicapping.

When I say it has outperformed the win pool it is based on one thing. Are the odds lower or higher in the P3 odds line compared to the win board. I generally use a 1.75 times odds and up to see if the horse is any value. So a horse that is 4-1 in the P3 should be about 7-1 at post to be a value horse.

[davew]

some ADWs keep probable daily doubles from previous races

I like looking at the favorite couple of horses probables, especially if a longer shot won the race


so instead of just looking at willpays from the previous race winner, I can also see the probables if other horses would have won (especially the well bet also rans)

[Robert Fischer]

Some situations seem to be more applicable to "smart money" than others.

I believe (or like to believe) that there are logical reasons for why a horse could have a better performance in these situations(or in any unique angle).

If you can build upon that angle with that logic, I believe that it can at times strengthen the quality of the play.

Let's try 2nd time starter+ FTL and bet heavy in the pick-3. So maybe the idea is that the trainer didn't have him fully cranked first time out. In other words he gave him kind of an in-race workout to get his hooves wet. So then it may be possible to build on that angle by looking for that same scenario. Well... his first race he was entered in 5 furlongs. He broke 4th by 4-lengths into a fast pace and he finished 5th by 5-lengths. Today he's entered for a mile.

OK.... well, on the replay of the 5f race, he ran kind of evenly and he had enough pace, that at a mile he could be near the pace today. Building. He finished 5th of 6 horses... but you thought it was a smooth ride. He wasn't being whipped like crazy before the turn and he looked like he changed leads ok and ran fairly evenly. Building some more. Maybe you have enough for a bet here already. Now you look into the trainer. Well he's Gary Capuano. Maybe you remember another 2nd time starter he had win the stretchout. Maybe you look at his stats and his 2nd time numbers are better than his first time numbers. Well.. he hits a higher% and ROI 2nd time out... maybe this is a trainer who likes to give his horses a race?

So now you've built on that angle, and I believe you get a stronger play.

[AndyC]

You have to go to my original post. I was using a calculation that I felt could be improved. So I asked what was the best way to turn the P3 into an accurate odds line. I am not trying to convince anyone that it works. If you feel what works at one track should work at all of them, then we have a difference of opinion. If you think what determines if this works or not is the money on live tickets, we also have a difference of opinion. It is the whole P3 pool from leg 1 to leg 3.

You will have to find out through research if this or other pools ( P4, exacta) outperform the win pool and if you can use it. I gave a few examples in an earlier post as to how it helped compliment my handicapping.

When I say it has outperformed the win pool it is based on one thing. Are the odds lower or higher in the P3 odds line compared to the win board. I generally use a 1.75 times odds and up to see if the horse is any value. So a horse that is 4-1 in the P3 should be about 7-1 at post to be a value horse.

Your premise, I believe, is that smart money in the P-3 will produce a better odds line than the win pool. If that is true, are there tracks with no smart money so looking at P-3 willpays isn't applicable? I find that unlikely, so yes we have a difference of opinion.

Let's look at this live ticket/whole pool situation. What happens if a horse is bet extremely heavily in the 3rd leg by insiders who have the dope on a sure winner but the bettors don't include the winner of the first leg on any ticket? The willpays will not reflect the true action of the horse because only live tickets are used. Any odds line derived from will pays ONLY reflect the live tickets on each horse in the last leg.

[FocusWiz]

some ADWs keep probable daily doubles from previous races

I like looking at the favorite couple of horses probables, especially if a longer shot won the race

so instead of just looking at willpays from the previous race winner, I can also see the probables if other horses would have won (especially the well bet also rans)I like these because it is my opinion that the smart money (if there is any) may not pick the winner of the race and we may then be using less informed betting as a predictor of the next race. Using the will-pays assumes that the winning horse had backing from the smart money, which may often be the case, but I was not seeing a strong pattern there.

I have been "watching" the daily double pools at irregular intervals at a few tracks and have used a rather contrived formula I use to compute how much a horse is "bet down" in these pools. Basically it compares the double probables against a combination of actual odds of the first leg and the morning line of the second. The morning line of the second leg is first adjusted to (somewhat unscientifically) account for known scratches. I generally exclude higher odds horses in the first leg from the calculation except if they are bet down significantly.

Since Mr. Fischer posted Race 8 from Belmont on 9/11, here are the ratios I obtained from the 7th/8th race that day (from a spreasheet I use):
Double Betting: 1 / 1.42

Double Betting: 5 / 2.50

Double Betting: 3 / 3.43


In interpreting the above, it would indicate that :1: was overbet the most in the double pool followed by :5: and :3: . The :5: won paying $14.40 with :1: coming in second for an exacta of $59.00. The :3: finished 4th. About 60% of winners are in these top three results. I use this information in a way similar to how the OP wants to use the Pick 3 will pays; not to handicap per se, but to be aware of where the money seems to be going. I have not tweaked my formulas for this (since I do not have the skills nor the data to evaluate all the possible permutations), but it seems that there is a lot of "unsmart" money bet on certain horses. For example, I often see the Morning Line favorite bet down in the double wagering but then do not see comparable support in the actual race. As Dave Schwartz has often noted, longshots get significant support at least partly because they are longshots. Number pairs (like 7/7) are also overly popular. My guess is that some bettors will select a horse they like and then look to pair it with something in the Double, doing less handicapping or having less "informed" information for the second leg.

AndyC asked if the Pick 3 should work at all tracks. The Double pools certainly don't. I was watching one race (I think at Mountaineer) where the entire Double pool was less than what my uncle used to bet on a race. The Double wagering gave me no insight into what would happen in the next race at all. However, even at a track like Belmont, I have noticed a pattern where Saturdays are less reliable than other days of the week. My guess is that there may be informed wagers and uninformed wagers and that this ratio will change from day to day.

[Capper Al]

How are you playing smart money horses? I'm looking at taking the top two that I agree with and throw in one or two of my long shots for a tri. Any other suggestions on how to interpret the board?

[AndyC]

AndyC asked if the Pick 3 should work at all tracks. The Double pools certainly don't. I was watching one race (I think at Mountaineer) where the entire Double pool was less than what my uncle used to bet on a race. The Double wagering gave me no insight into what would happen in the next race at all. However, even at a track like Belmont, I have noticed a pattern where Saturdays are less reliable than other days of the week. My guess is that there may be informed wagers and uninformed wagers and that this ratio will change from day to day.

I have serious doubts that using P-3 willpays for an odds line has much value at any track. Certainly not at small pool tracks.

[AndyC]

I have long been an advocate of having P-3,P-4, P-5, and P-6 toteboards that can be easily shown on a monitor. The board would have as many columns as races in the bet. A P-3 would have 3, a P-4 would have 4, etc. There would be as many rows as needed to show the number of entrants in each race. What the board would show is how the total pool money is spread over each race. So if there was a P-3 with $100K in the pool column 1 would show how the $100K was distributed on each entrant in the first race. Column 2 would show how the $100K was distributed on each entrant in the second race and the same for column 3. Each column would add up to $100K. Quite easily done and it would add much daylight to many blind bets.

Being a P-6 player, I have seen many times where horses in down legs with no form (especially maidens) were being bet way down in the win pools and other pools long after the P-6 betting was closed. Really an unfair situation for most bettors.

[imofe]

I know there have been posts that concentrate on whales and computer groups and how they do or do not change our chances as individual bettors. But think about the average guy that goes to the track. He picks up a program or form as he walks into the door. He bets win, dd, some exactas, tris, or 10 cent supers. He will look to play a P3, P4 , P5 or P6 if he has a few chalk singles ( especially at the beginning) or if there is a carryover or maybe a guaranteed pool. Now, if he is making any of the one race bets ( win, ex, tri, sup) he uses the win board as his guide. He does not use exacta probables or P3 will pays as his guide. If a horse was the low P3 and went off at 5-1 how many people would even notice? If a horse was 8/5 ML and went off at 5-1 how many people would notice? The current win odds are used by just about everyone. They can and will be distorted at times.

When I first started noticing this I wondered if it was worth the effort to track the P3 for a longer period of time. After all, we have a tendency to noticed the winners but not the losers. Then when it is looked at long term, we see all the losers. I asked a guy who I believe is a real good horseplayer if he thought it was worth it. His answer shocked me. "I have known that the win pool is the weakest pool at that track for years", he said.

A lot of posters are talking of smart money. Now I am not saying horses never get hidden in multi- race bets. But I am talking about the pool itself being a better indicator of who will win. As an experiment anyone can try this out at their favorite track. Identify horses you think will be the favorite and you think are vulnerable before you get to the track. Use the multi-race will pays, or exacta board if it does not fluctuate a lot after the race goes off to determine a separate odds line. Look for races where the will pays agree with you that the odds on the favorite are lower than they should be. Try to play your contenders that show value from the wins odds to the multi-race line. Give it a try for a week or so. You might be surprised at your results.

[imofe]

Here is an example from a little while back

Actual odds P3 odds

Chalk 3/5
P3 odds 2-1

Contender #1 4-1
P3 odds 4-1

Contender #2 7-1
P3 odds 4-1

Contender #3 5-1
P3 odds 8-1


I believe the chalk is vulnerable in this race. It is a really nice horse but is slowly losing form. Contender #2 shows my only value. The #3 contender is the reason I do not use ML. The horse is 12-1 ML and some would say is the hardest hit P3. But the horse is also really bet on the win board.

The race came contender #2 followed by contender #3. I am mostly a win and exacta player so I caught a nice win price and an exacta. Some people who bet tri's and supers may have been able to exploit this race to a greater degree than I did. I am not strong at putting together the bottom part of tri's and supers, so I rarely play them.

[AndyC]

The current win odds are used by just about everyone. They can and will be distorted at times.

Despite what you might believe the will pays reflect live tickets. How else would someone compute the will pays? Because the amount of live tickets is so small big distortions can be the norm and not the exception.

[imofe]

Yes. The will pays reflect live tickets. Maybe I am doing a bad job of explaining my opinion. The will pays also reflect dead tickets because someone has to pay the live tickets. So when I say that it counts from leg 1 to leg 3 this is what I mean. You can not just use the winning ticket if you believe the whole P3 pool is significant.

Let's say that a 20-1 wins a race. Does that make all the money bet on every other horse insignificant? No. That is how the horse got to be 20-1. It is the same thing in the P3 pools. The payouts are based on all the bets, not just the part at the end where you see the will pays.

[FocusWiz]

imofe,

I tend to use the double wagering in a similar manner, but with different rationale. I like to watch the doubles because the bets were committed during the last race and were likely bet down in the closing minutes of the first leg. After that, the information is static.

When I see a favorite bet down in the current race which had little or no support in the double wagering, I am likely to become suspicious of it, but primarily because I expect the favorites and longshots to generally have a little more support in the double due to the way many bettors wager. As I indicated earlier, the other reason is that the pool information is available and I can see how the wagering was done with other horses besides the winner. In fact, I have seen races where each of the top four horses was paired with a different doubles "favorite" in the next race. I use this as an indicator that the race is quite open.

This is not an exact science, since the best horse will not necessarily have the most backing and some bettors who are acting on a tip have no idea what a Daily Double or a Pick 3 actually is. When I refer to "smart" money or "informed" bettors, I am not just talking about inside information, but also individual handicappers who have analyzed the race. The odds on the current race and in the various pools reflect a mixture of information, misinformation, solid tips, and wild guesses. Developing an odds line as you are attempting is similar to what I am doing when I look for my ratios. We are trying to pick out the useful information out of the background noise.

I am still trying to integrate this into an overall wagering strategy. I do not want to blindly bet the doubles favorite but I don't want to ignore it either. For now, I look at it and if it supports my existing contenders, I may consider them higher and if not, I may back off. That is not quite a strategy, though.

[imofe]

Good info FocusWiz. Some tracks just have an early/ late double but some have more in the middle of the card. I like your logic concerning the favorite that is cold in the double payouts, especially if I was already looking to play against this chalk in my handicapping. Many times the exactas will reveal the same thing. A horse is 6/5 but the exactas are too high with him on top and there may even be exactas that are lower with him running second.

[AndyC]

Let's say that a 20-1 wins a race. Does that make all the money bet on every other horse insignificant? No. That is how the horse got to be 20-1. It is the same thing in the P3 pools. The payouts are based on all the bets, not just the part at the end where you see the will pays.

You are mixing apples and oranges. In the win pool ALL tickets are live until the race is over. In the P-3 pool if there are 100 live tickets and the #1 horse has 10 of them then the odds determined by the P-3 pool would be 9-1. A horse with 20 live tickets would be 4-1, etc, etc. Dead money doesn't even come into the computation.

[Capper Al]

You are mixing apples and oranges. In the win pool ALL tickets are live until the race is over. In the P-3 pool if there are 100 live tickets and the #1 horse has 10 of them then the odds determined by the P-3 pool would be 9-1. A horse with 20 live tickets would be 4-1, etc, etc. Dead money doesn't even come into the computation.

I have been trying to make this point to my friend. On my app I don't even want to see dead money horses while he on the other hand wants every horse ranked.

[Dave Schwartz]

Dorcus & DaSilva, in Science and Betting (1961 or so), spoke about the bet UP horses being positive as well as the bet downs.

I have found this to be true.

One is driven by hit rate, the other by payoff.

[AndyC]

Dorcus & DaSilva, in Science and Betting (1961 or so), spoke about the bet UP horses being positive as well as the bet downs.

I have found this to be true.

One is driven by hit rate, the other by payoff.

From back in the days when all betting was done on-track and when exactas, trifectas, P-3s, etc didn't exist. Not to mention computers for doing last second calculations and making automatic bets. Probably not research you can hang your hat on for today's players.

[Ocala Mike]

At NYRA, I still prefer to use the DD will pays rather than the P/3 will pays, especially where it is the significantly larger pool. In today's 9th race, I didn't come up with CRACKERJACK JONES until I looked at the DD will pays.

1 66.50 3/1
2 405.00 30/1
3 41.40 5/2
4 25.80 4/1
5 68.50 8/1
6 39.60 5/1
7 SCRATCHED
8 81.00 15/1
9 72.00 10/1

A simple division of the M/L (the right-hand figure) into the will pay reveals the 8 as the best. He didn't get bet in the P/3, though.

The DD pool was 59150, P/3 pool 34953.

[imofe]

MIke,

If you find that DD work better at a track, stick with it. The bottom line is that another pool ( ex, DD, P3 ,P4) can be a better indicator of who will win than the win pool. But you then in turn bet into that win pool.

[AndyC]

MIke,

If you find that DD work better at a track, stick with it. The bottom line is that another pool ( ex, DD, P3 ,P4) can be a better indicator of who will win than the win pool. But you then in turn bet into that win pool.


The win pool is a very efficient pool for determining probabilities of winning. See http://faculty.chicagobooth.edu/Richard.Thaler/research/pdf/parimutual%20betting%20markets.pdf

To say that looking at will-pays from a P-3, which by definition is filtered by the results of 2 other races, will produce a more efficient betting line than the win pool is wishful thinking at best.

[imofe]

You are mixing apples and oranges. In the win pool ALL tickets are live until the race is over. In the P-3 pool if there are 100 live tickets and the #1 horse has 10 of them then the odds determined by the P-3 pool would be 9-1. A horse with 20 live tickets would be 4-1, etc, etc. Dead money doesn't even come into the computation.

What you are missing here is that the tickets to the #1 have to be good through the first two legs. Not just any ticket to the 1. Lets say this horse goes off at 19-1 and half of those 10 tickets are made by a group in the know. The win board reflects odds as if those 5 tickets did not exist. But the 5 tickets must also contain the winner in the first two legs or it is useless. It is the overall success of the P3 bettors which makes it valuable. Not just that they used #1 in the last leg. That being said, you were correct in that the odds are determined by live tickets. I stand corrected.

I appreciate your comments and feedback. I also respect you asking questions and not taking what I am saying as fact. But I will share this with you. At this track over the last 3 months if you had bet any horse that was the low P3 and was not the favorite, you would have a nice profit. This is about as basic as one can get. But I also wanted to be able to take advantage of horses that were not the low will pay, but were lower than there win odds. That is why I asked about an odds line.

[Dave Schwartz]

From back in the days when all betting was done on-track and when exactas, trifectas, P-3s, etc didn't exist. Not to mention computers for doing last second calculations and making automatic bets. Probably not research you can hang your hat on for today's players.

I was not suggesting that this would be a profitable strategy.

Merely pointing out that the other side of the coin also has merit as a metric.

[AndyC]

What you are missing here is that the tickets to the #1 have to be good through the first two legs. Not just any ticket to the 1.

Therein lies the problem. A horse could have 20% pool bet on it in the 3rd leg but due to bad results in the first 2 legs it may have only 10% of the live tickets bet to it. Having 2 races as a filter before the 3rd leg doesn't make the live tickets any smarter.

Having relatively small pools a smart bettor is not likely going to buy the first 2 legs just to be alive to his smart horse. So it could be that smart money may disappear in the P-3 due to key horses losing in the 1st 2 legs.

I would have to believe that any indication of smart money would be more apt to show when the first 2 legs are won by logical contenders.

[AndyC]

I was not suggesting that this would be a profitable strategy.

Merely pointing out that the other side of the coin also has merit as a metric.

I remember the early bet drift up winners very well from my tote watching days in the 70s.

[imofe]

[QUOTE=AndyC]Therein lies the problem. A horse could have 20% pool bet on it in the 3rd leg but due to bad results in the first 2 legs it may have only 10% of the live tickets bet to it.

If this is the case most of the time then we have nothing. But what if the opposite was happening more often than what you mentioned? What if the chalk in leg 3 actually had more tickets made to him, but after 2 legs, there are more live tickets to someone else? Now we have a situation where the public will still bet the horse that had more actual tickets made to them, but the horse is not the low will pay.

[AndyC]

If this is the case most of the time then we have nothing. But what if the opposite was happening more often than what you mentioned? What if the chalk in leg 3 actually had more tickets made to him, but after 2 legs, there are more live tickets to someone else? Now we have a situation where the public will still bet the horse that had more actual tickets made to them, but the horse is not the low will pay.

If after 2 legs there are more live tickets to a horse other than the horse that had the most original tickets bet on him I would say that it pretty much proves my point that will-pays are very unreliable. The filtering out process of the first 2 legs is not related to the chances of any horse in the 3rd leg. All it does is skew reality.

[FocusWiz]

I think I see points on both sides of this discussion. Using the will-pays will pare down the information to a specific set of horses in the prior races. It will definitely skew the results and reduce the sample set. Whether it skews it in the right direction or the wrong direction we will know about 35 minutes later when the results come in, which is probably not the goal.

I was going to post this yesterday, because Ocala Mike posted something that really peaked my interest because my figures had not alerted me to CRACKERJACK JONES since I focus on the prior race pool rather than the will-pays. I did have indications in my spreadsheet, but I missed them.

From my rather unscientific analysis of those races, there were four 9th race horses that were "too close to call" in the Doubles wagering in the 8th race (i.e., for the 8th/9th Double), the :4: :9: :6: and :8: . This was at least partly due to heavy betting on the 6/4 Double combination, neither of which ultimately figured in the results of either race.

When the dust cleared from the 8th race, though, all three of the horses which finished "in the money" were paired with CRACKERJACK JONES in the double wagering. In this case, the will-pays would have been skewed towards the more desired result if any of these had won. I can just as well cite dozens of examples (I won't) where the winner of a race was paired with an untimate noncontender but the overall Doubles pool clearly picked the winner.

As I indicated earlier, the goal is to separate the information from the noise. Hindsight is an excellent tool for that, but it does not help in the actual wagering.

[imofe]

If after 2 legs there are more live tickets to a horse other than the horse that had the most original tickets bet on him I would say that it pretty much proves my point that will-pays are very unreliable. The filtering out process of the first 2 legs is not related to the chances of any horse in the 3rd leg. All it does is skew reality.

Not that unreliable if the low one keeps winning most of the time. I guess time will tell.