Last Minute Odds Drop Analysis

[barn32]

OK, if this is a dumb idea I'm fine with that. (I'm not Trifecta Mike for Christ sakes.)

Has anyone broken down last minute odds drops according to last minute odds?

In other words, it seems to me from much of the discussion I've read about this topic that most last minute odds drops are on short(er) priced horses. So if you arrange the odds from low to high, how much (and how often) do they drop for each odds segment.

Example:

Odds

1 (even)
2 to one
3 to one
4 etc.
5
6
7
8
9
10

You can break these down even further (like 8/5, 9/5, 5/2, 7/2, etc.) but you get the idea.

So if you have this data, your analysis would read something like this, give or take.

1 50%, 30% (Horses at even money drop in odds 50% of the time, and average a 30% drop.)
2 25%, 17% (Horses at 2 to 1 drop in odds 25% of the time, and average a 17% drop.)
3 10%, 5% (etc)
4 5%, 3%
5 2%, 1%
6 1% .05%
7 .05%, .02%
8 .02%, .01%
9 .01%, .005%
10 .009%, .0045%

Horses with last minute odds of 3 to 1 or less show the largest percentage drops. 4 to 1 and higher, and the drops are negligible.

Conclusion: Assuming the above is accurate (or close) at one minute to post only play horses 4 to 1 or higher. (Or 5/2, 3 to 1 or 7/2,--whatever your data tells you.)

It might also be track dependent, I don't know.

All right, let me have it.

[DeltaLover]

What you describe here is not a simple task.

In the past, I had written a live odds feed retriever trying to detect patterns related to price movements trying to detect the closing. I was only considering odds while ignoring anything else like trainer, owner, or the overall profile of each horse.

I have to admit that this exercise was proven to be a fruitless project, convincing me to abandon the idea that I can come up with this kind of a final price predictor.

[FocusWiz]

I likewise agree with DeltaLover.

I have seen some patterns where "generally" very short horses odds-on entries rise in odds during the last minute with a reduction in the odds for next "tier" of favorites and an overall increase in the odds for the longest wagers, but nothing that I could hang my hat on.

[traynor]

It depends on the track, and the cause of the odds drop/rise. Lumping all tracks/situations/scenarios together blurs any (decipherable) distinctions. As with most other aspects of horse racing, looking at the "big picture" often hides all the useful information that would be readily available to more astute observation.

[Robert Fischer]

(edit- Just read Traynor's post and it appears we are on similar wavelengths)

The way the track collects it's data from the hubs, and the way a few large players opt to play that specific track,scenario,day,etc... make a big difference

Some tracks show 2.60 to show on eventual 2.20 horses all the time

Some do similar with $5 horses that start at 2-1.

Some scenarios, tracks show $3.80-$4 horses opening at 2-5.

Aside from something you can't control, such as some random guy placing $10K on a horse, a lot of it comes down to knowing the play inside and out, and knowing the patterns of how that play often gets played by the public at that track.

[Dave Schwartz]

I did a SMALL study a couple of years ago where I took the toteboard at the first refresh flash of zero minutes to post and compared it to the "winner cirlce" odds (i.e. final).

Be aware that the "first refresh flash" means that when the mins-to-post said "0" for the first time I waited until the pool totals refreshed, which could be as much as 30 seconds later.


I divided the tracks into 3 groups:

1. low rebate tracks (i.e. less than 1.5% or even 0%)

2. Mid rebate tracks (i.e. >1.5% to 5.99%) - note these would be lower now.

3. High rebate Tracks (6% and up)


Could only be called anecdotal because the sample size was small - only around 200 races total.

Percent of final pool that was on the board at 0 minutes:
Group 1. 43%
Group 2. 35%
Group 3. 27%


The Winning Horse: Odds changed - i.e. went up or down a full tick:
Note: Definition of "full tick" means that a horse that was 4.3:1 went off at 3.3:1 or below. Horses that dropped below 3/1 were required to drop 0.5. Thus, 2.3 would have to go to 1.8:1 or below.

Group 1.
38% went up
20% stayed the same
42% went down

Groups 2&3.
10% went up
17% stayed the same
73% went down



Please note that I am sharing this not because I think the deal is done, but that perhaps this will give someone who wants to do a more complete study a starting point.

[MJC922]

Dave thanks for sharing.

[traynor]

I think the size of the mutuel pools and the relative predictability of races at a given track are more useful ways to categorize than rebates. Lost (or apparently lost) in the analyses of pool trends is that in order to profit from rebates, one still has to pick a LOT of winners. Maintaining a high enough strike rate to get into a range at which rebates produce an overall profit is not easy. Especially at the tracks with the best rebates.

Additionally, the amount one can dump into the mutuel pools at the last minute (without diminishing the return into the red, with or without a rebate) will vary widely--even at the same track on the same day. Other factors are involved in such situations than the amounts of the pools at various points, but that is a great starting point.

[Dave Schwartz]

I think the size of the mutuel pools and the relative predictability of races at a given track are more useful ways to categorize than rebates.

Traynor,

I must disagree.

IMHO, the higher the rebate, the more attractive the pool is to whales.

The greater the percentage of whale money in the pool, the more difficult it is for anyone to make a profit.

I liken it to walking into a poker room where there are two identical poker games (i.e. same rules, buy-ins, etc.). Table #1 has a slightly lower rake, thereby attracting the best players in the room. The lesser players gravitate towards the table #2.

My belief would be that table #2 would be easier to beat.

[OverlayHunter]

Percent of final pool that was on the board at 0 minutes:
Group 1. 43%
Group 2. 35%
Group 3. 27%

Dave, do I understand you to being saying that, on average, 73% of the dollars wagered to win on Group 3 races enters the pool after the first refresh flash of zero minutes to post?

[Dave Schwartz]

Yes.

[cj]

Originally Posted by Dave Schwartz

Yes.

Makes sense, everyone has learned the odds early mean nothing at these places. Hell, the odds late mean very little too.

[Cratos]

Originally Posted by Dave Schwartz

Traynor,

I must disagree.

IMHO, the higher the rebate, the more attractive the pool is to whales.

The greater the percentage of whale money in the pool, the more difficult it is for anyone to make a profit.

I liken it to walking into a poker room where there are two identical poker games (i.e. same rules, buy-ins, etc.). Table #1 has a slightly lower rake, thereby attracting the best players in the room. The lesser players gravitate towards the table #2.

My belief would be that table #2 would be easier to beat.

I agree with your conclusion.

[Cratos]

Originally Posted by Dave Schwartz

Yes.

Again, I agree with you.

[traynor]

Originally Posted by Dave Schwartz

Yes.

It seems that might be a very productive area to study. Interesting.