Shouldn't a morning line's probability total equal

[Turfday]

100 % plus the takeout? In other words, the percentage total derived from the probability of odds on all entrants should be 100% + the takeout at that track?

Therefore, at Santa Anita, shouldn't the line "total" be between 120-125% when the odds percentages are totaled up on all entrants per race right now?

I've noticed that new Oak Tree / Santa Anita morning line maker Jon White's totals often go more than 130%.

[46zilzal]

You are just figuring this out now? Derby pools have been that way for years above 122-124%

[Overlay]

I would imagine that some of that level of inaccuracy would be the result of a combination of rounding, and of artificial constraints on the odds (particularly on the high-odds side), in order to make a race more competitive from a wagering standpoint, or as a result of the line being based on a projection of the public's betting habits, rather than the linemaker's personal opinion. Also, larger field sizes like the Derby would tend that way, simply because there are so many more horses in the race than a typical field.

[markgoldie]

I knew a linemaker (now deceased) who used a very simple group of tables that were arranged in order of the lowest morning line. This table gave sample odds depending on odds' position and field size. If he deviated from the sample one way or the other, he would simply add or subtract a point or two somewhere else. Naturally, this led to improper balancing on occasion, although to my knowledge, he never received complaints about this issue (although frequently, he did receive ones regarding the accuracy of his handicapping).

At any rate, if they still use this method, it might account for the improper balancing of the totals.

[CBedo]

100 % plus the takeout? In other words, the percentage total derived from the probability of odds on all entrants should be 100% + the takeout at that track?

Therefore, at Santa Anita, shouldn't the line "total" be between 120-125% when the odds percentages are totaled up on all entrants per race right now?

I've noticed that new Oak Tree / Santa Anita morning line maker Jon White's totals often go more than 130%.The morning line maker is trying to get a line that will be what the public thinks, and as you know, the public odds are takeout adjusted, so the total will always add up to well over 100%. It's pretty easy to adjust if you want. Just convert the m/l to a probability and divide it by the total of all the horse's probabilities. This will give you the 100% morning line.

[eqitec]

I compare final morning lines after scratches vs. actual odds at 20 minutes before post as a way of detecting early "action". My problem is how to adjust the morning linemaker's line. Is there any mathmatical formula that should be applied to the lines of all starters after scratches so that the total does end up as 100% + takeout?
I've been doing this manually by adjusting each horse by an amount that makes sense to me. E.g., in a 6 horse race three 20-1 shots might get adjusted down to 15-1 while the top three favorites at 2-1, 3-1, and 4-1 might get adjusted down to 8-5, 5-2 and 7/2.
Can anyone suggest a better way to make these adjustments?

[CBedo]

Here's an example from Hawthorne's 9th:

1) Start with the morning lines.
2) Convert to win probabilities. The total of this is well over 100%
3) Scratch some horses and retotal. This could be over 100%, but could be under depending on scratches.
4) Divide each probability by that total. This is the true morning line probability.
5) Convert the probabilities back to a true morning line.

Hope this helps.


Horse ML Prob Scratch 100% Adusted_Odds
1 6 14.3% 14.3% 12.1% 7.3
2 30 3.2%
3 5 16.7% 16.7% 14.1% 6.1
4 12 7.7% 7.7% 6.5% 14.4
5 6 14.3% 14.3% 12.1% 7.3
6 15 6.3%
7 4 20.0% 20.0% 16.9% 4.9
8 12 7.7% 7.7% 6.5% 14.4
9 8 11.1% 11.1% 9.4% 9.7
10 9/2 18.2%
11 12 7.7%
12 6 14.3% 14.3% 12.1% 7.3
13 15 6.3% 6.3% 5.3% 18.0
14 15 6.3% 6.3% 5.3% 18.0

Total 153.9% 118.5% 100.0%

EDIT: Actually, the 13 & 14 were scratched as well, but hopefully, you get the process.

[ranchwest]

I would suspect that some variance is due to rounding. For instance, morning line odds are almost always even numbers above 5/1. So, you have 6/1, 8/1, 10/1, 12/1, then 15/1, 20/1, 25/1 and 30/1. There are seldom values higher than 30/1. If the arithmetic should have had a horse at 7/1, it will become 8/1 and 18/1 will become 20/1 and 16/1 goes to 15/1. 50/1 goes to 30/1. Or, at least that's what appears to be happening.

[The Hawk]

I knew a linemaker (now deceased) who used a very simple group of tables that were arranged in order of the lowest morning line. This table gave sample odds depending on odds' position and field size. If he deviated from the sample one way or the other, he would simply add or subtract a point or two somewhere else. Naturally, this led to improper balancing on occasion, although to my knowledge, he never received complaints about this issue (although frequently, he did receive ones regarding the accuracy of his handicapping).

At any rate, if they still use this method, it might account for the improper balancing of the totals.

Costy Caras used to do this at Charles Town. His lines added up to what they were supposed to but they were often terrible.

[Turfday]

Using California as an example, where the pari-mutuel handle on win bets is 15.43%, it's no big deal to see morning line odds converted to probability percentages of more than 130% or even higher in some races?

And one of the reasons is that above 5/1, we don't see odd numbers (like 7/1, 9/1, etc.)?

[Overlay]

Using California as an example, where the pari-mutuel handle on win bets is 15.43%, it's no big deal to see morning line odds converted to probability percentages of more than 130% or even higher in some races?

And one of the reasons is that above 5/1, we don't see odd numbers (like 7/1, 9/1, etc.)?

I'm not disputing your observation about Mr. White's morning lines, but could you post one (or more) of them on the board, so that we could dissect and discuss specific data?

[Turfday]

Note: certainly not every race comes up this way, but I've seen quite a few. So wondering if this is de rigeur in making a morning line.

Oak Tree, Friday, Oct. 30 (5th race)

1 ENTRY, 8/1 = 11%

2 REALLY UPTOWN, 5/2 = 27%

3 SWEET TIGGER, 5/1 = 16%

4 COO CACHOO, 8/1 = 11%

5 DEL MAR TRUCKEE, 15/1 = 6%

6 SEEUAT STICKNSTEIN, 3/1 = 24%

7 PURSE LUVA, 4/1 = 19%

8 GILLIGAN, 4/1 = 19%

[The Hawk]

Note: certainly not every race comes up this way, but I've seen quite a few. So wondering if this is de rigeur in making a morning line.

Oak Tree, Friday, Oct. 30 (5th race)

1 ENTRY, 8/1 = 11%

2 REALLY UPTOWN, 5/2 = 27%

3 SWEET TIGGER, 5/1 = 16%

4 COO CACHOO, 8/1 = 11%

5 DEL MAR TRUCKEE, 15/1 = 6%

6 SEEUAT STICKNSTEIN, 3/1 = 24%

7 PURSE LUVA, 4/1 = 19%

8 GILLIGAN, 4/1 = 19%

It's even higher than you think. 4-1 is equal to 20% (or points), 3-1 is 25% and 5/2 is 28.5%. So it comes out to at least 137, which is way too high, especially for an eight-horse field.

[Turfday]

I thought I was imagining things.

[stu]

When I made the morning line for Sunland, Zia Park, Ruidoso, Arapahoe, and Manor, I followed the same rule:

- the morning line is supposed to total 100% plus the takeout percentage and 1 percentage point for each horse in the race.

I was taught by other ML makers that this was the industry standard.

[MakinItHappen]

When I made the morning line for Sunland, Zia Park, Ruidoso, Arapahoe, and Manor, I followed the same rule:

- the morning line is supposed to total 100% plus the takeout percentage and 1 percentage point for each horse in the race.

I was taught by other ML makers that this was the industry standard.

Stu,

Any idea what the logic is for the "and 1 percentage point for each horse in the race." Just curious...

Thanks & Best of Luck,

MakinItHappen

P.S. - If anyone else knows the answer or cares to guess, feel free!

[MakinItHappen]

It's even higher than you think. 4-1 is equal to 20% (or points), 3-1 is 25% and 5/2 is 28.5%. So it comes out to at least 137, which is way too high, especially for an eight-horse field.

Does anyone think there is any possibility that the Morning Line is being intentionally deflated a bit to give the appearance of value and therefore perhaps encourage additional wagering? For example, if a horse should be 4-1 ML on a mathematically balanced ML (even with takeout) but instead the ML is listed as 3-1, perhaps some may glance and the board and see the horse at 4-1 vs the 3-1 ML and conclude that perhaps there is a little value there and place a wager. As simplistic as this is, it still seems to be a bit more clever than I am willing to give "them" credit for... :lol:!

If this were being done intentionally, it would show up in routinely high aggregate race total win probabilities, as is being discussed here... just food for thought.

Best of Luck Everyone!

MakinItHappen

[stu]

Stu,

Any idea what the logic is for the "and 1 percentage point for each horse in the race." Just curious...

Thanks & Best of Luck,

MakinItHappen

P.S. - If anyone else knows the answer or cares to guess, feel free!

Just guessing -- breakage.

[BMeadow]

I randomly checked a few recent days at Santa Anita and found that Bob is right--Jon White has, on several occasions, had morning lines that total closer to 130 than to 120.

With a takeout of 15.43% to win in California, the morning line should roughly total 119-120 (119 x .8457 = 100.63, which would cover breakage). Allowing for rounding might push this up to 124 or so. But 130? Mathematical laziness.

Jon's a good guy who, I'm sure, could beat me in a horse-racing trivia contest of 100 questions by a score of 100 to 0. But his math on the morning line has, from time to time, been less than stellar.

[Turfday]

Hi Barry,

Thank you for stepping in and validating what I initially thought with your expert opinion.

[Java Gold@TFT]

Can someone please breakdown what a normal M/L would be for the BC Classic (with the math). He posted 6 out of 13 horses at 12-1 which of course will get blown away in real life. What SHOULD the M/L look like with obvious bias as to which horsess will take more money than others? I don't see Zenyatta favored. Last I looked Ladbrokes had Rip at 3/2 and her at 4-1. Obviously a British bias and we don't know how his hoof is but i still see him favored over her and Summer Bird being equal odds with Zenyatta. Anyone care to make a guess?

[boomman]

Just guessing -- breakage.

Yep that's my understanding of what it represents-Boomer