Question for Mathcapper

[BitPlayer]

On Saturday, I was using the DD probables from the prior race to project odds for the BC dirt races. I noticed that the discrepancy between my projected odds and the off odds seemed larger than I\'m used to.  When I was playing around with projecting odds during the Saratoga meet, the median discrepancy was around 10%.  On Saturday, it was almost 20%.

Is that something you have noticed on big days?  I\'m wondering if it might be a function of pool size, increased casual money, or field size.

[Boscar Obarra]

Pools were wacky BC day.

 You\'re correct, there were more divergences than usual.

 I saw some huge bets on longshots in the doubles, that never followed through into the win pool .

[trwhis2]

This is likely due to the BCBC.

Players are betting higher than normal amounts into double pools because of contest rules.  Obviously, they’re doing it inefficiently.

[Boscar Obarra]

What rules are you referring to?

 There was a 75-1 shot bet down to to under 20-1 in the doubles . Not talking with only 1 horse either

[Mathcapper]

Bit --

Other than what Boscar noted about some longshots getting bet down in the doubles   much more than in the win pool (like that 16 horse in the Mile that was 18-1 in the Will Pays and went off at 84-1), I didn\'t really see all that much that was out of the ordinary.

Perhaps those longshots that were bet down in the doubles were skewing the average discrepancy you calculated. I don\'t know the specifics of that BCBC contest format so I can\'t really speak to it, but trwhis2 may have a point about longshot doubles getting bet down by contest players looking to leap to the top of the leaderboard.

It\'s also possible the bigger fields have something to do with the discrepancy you observed. Whenever there\'s a really big field, there\'s often a few horses whose implied odds vary significantly from the final odds, although I would think the fact that you\'re looking at the median instead of the average discrepancy would account for that.

In any event, I\'ve posted the ML\'s, implied Will Pays odds (based on top 3 favorites from prior race) and Final Odds for each of Saturday\'s BC races below in case you want to check your data.

Generally, the only thing I\'ve observed over the years on big days in regards to Will Pays is that I sometimes see the smart money acting in reverse. On a day-to-day basis, the win pool is the most efficient pool, but on really big days, I\'ve seen the opposite to be true on occasion (the Palace Malice Belmont comes to mind -- 7-1 in Will Pays, 13.8-1 final odds).

This could just be selection-bias on my part, I\'ve never really done a study on it. But it makes sense in that on big days the sharp money may be getting overwhelmed in the win pool by the all the uninformed public money that only comes in on such days.

Didn\'t really see that effect on Saturday though -- six of the nine BC winners went off below their implied Will Pays odds, including the winners of the last three biggest races on the card -- Accelerate, Enable and Monomoy Girl.

F&M Sprint   CD   3
   ML   DD   FL
1   4.4   6.39   5.39
2   16.3   39.3   39.4
3   21.7   52.9   65
4   32.5   71   59.7
5   10.9   7.1   6.45
6   16.3   21.9   15
7   10.9   31.7   33.1
8   16.3   23.8   21.4
9   21.7   38.1   46.9
10   21.7   52.7   43.7
11   10.9   10.4   7.51
12   16.3   15.5   33.1
13   1.81   0.72   0.92
14   21.7   37   27.7
         
Turf Sprint   CD   4
   ML   DD   FL
1   21.5   19   19.3
2   10.8   22.5   27.4
3   21.5   37.7   36.6
4   16.1   18.3   18.3
5   3.81   4.62   4.08
6   12.9   13.4   13.2
7   21.5   14.8   12.2
8   16.1   14.7   12.2
9   4.35   6.43   7.12
10   21.5   18.4   19.3
11   6.48   1.88   2.05
12   12.9   28.4   33.5
13   21.5   21.3   15.2
14   6.48   12.5   16.3
         
Dirt Mile   CD   5
   ML   DD   FL
1   2.55   2.07   2.7
2   20.3   53.4   65.6
3   20.3   35.1   30.7
4   15.2   22.3   23.3
5           
6   5.08   10.1   10.6
7   6.1   5.75   5.35
8   20.3   23.8   14.9
9   15.2   22.1   23.3
10   1.64   1.02   0.9
         
F&M Turf   CD   6
   ML   DD   FL
1   5.45   3.92   4.76
2   21.6   39.5   69.1
3   3.84   2.92   2.66
4   21.6   28.2   38.8
5   32.3   22.5   38.8
6   3.3   4.6   3.19
7   32.3   23.9   55.5
8   32.3   93   104
9   10.8   11.5   11.6
10   5.45   7.12   5.28
11   21.6   32.5   34.6
12   16.2   18.4   31.5
13   16.2   19.7   50.3
14   16.2   8.15   6.33
         
Sprint   CD   7
   ML   DD   FL
1   6.43   6.51   6.21
2   6.43   5.02   4.66
3   21.3   59   58.7
4   16   56.2   35
5   1.97   1.75   1.68
6   21.3   65.1   53.6
7   12.8   51.1   39.2
8   6.43   4.56   4.66
9   2.71   2.13   2.6
         
Mile   CD   8
   ML   DD   FL
1   11.7   14.8   21.1
2   9.76   7.44   7.41
3   14.6   14.3   13.7
4           
5   5.84   3.72   2.68
6   11.7   21.7   20
7   7.8   5.89   5.31
8   9.76   11.3   11.6
9           
10   5.84   9.41   7.41
11   19.5   15.8   21.1
12   5.84   6.53   6.36
13   9.76   13   14.8
14   14.6   15   21.1
15   29.3   35.5   45.3
16   29.3   18.3   84.2
         
Distaff   CD   9
   ML   DD   FL
1   12.5   29.4   22.1
2   3.68   1.65   2.15
3   20.9   43.3   35.7
4   31.3   80.5   89.3
5   15.7   22.9   20
6   31.3   83.1   90.3
7   6.28   6.88   5.3
8   12.5   35.2   30.5
9   8.37   7.35   6.35
10   6.28   6.48   8.44
11   2.12   2.08   1.94
         
Turf   CD   10
   ML   DD   FL
1   12.8   10.4   22.2
2   1.12   0.94   0.8
3   12.8   7.85   7.48
4   10.6   10.2   9.68
5   10.6   12.1   10.1
6   31.8   49.3   53.2
7   12.8   16.7   12.7
8   31.8   90.4   75.7
9   31.8   40.9   136
10   21.2   31.1   41.9
11   31.8   26.6   37
12   4.82   6.06   5.29
13   31.8   33.2   114
         
Classic   CD   11
   ML   DD   FL
1   13.2   9.54   14.6
2   22   14.4   19.8
3   8.86   7.68   5.23
4   22   25.5   31.2
5   33   39.6   93.5
6   6.67   4.49   3.67
7   5.58   6.7   6.27
8   22   26   40.5
9   13.2   12.1   13.5
10   11.1   7.26   14.6
11   6.67   6.95   9.39
12   33   35.6   53
13   22   38.6   55.1
14   2.84   4.93   2.63

[analizethis]

In your computation of \"will pay odds\" do you apply any adjustment for takeout differences?

Thanks.

Bob

[Mathcapper]

Hey Bob --

Yes, I take into account the takeout rate for both the daily double and the win pool for the particular track as well as the difference between the two in the computations.


I also make a small adjustment for breakage and put all three lines (Morning Line, Will Pays implied odds, Final Line) on the same scale based on the win takeout and breakage estimate for the particular track.

[BitPlayer]

I use the DD probables from the top two favorites rather than the top three finishers, so there is some discrepancy between our numbers.  Your approach did better than mine.  You are also right that the projected odds for long-priced-DD horses are off a bit more (as a percentage of the projected odds) than those for the shorter prices. Thanks very much for your help.

[Mathcapper]

Bit -- I mistyped in previous post (edited), I use the top 3 favorites, not the top 3 finishers, so we should be pretty close.