Days Matter

[Dave Schwartz]

I must be on ignore with a lot of people. Or else nobody believes me.

Eliminating based upon 30 days is just a bad idea in just about every category.

Sweet spot is in yellow. 26-42 days.

[thaskalos]

Dave,

Has anybody thought of doing a recency study exclusively on those horses who won or just missed winning in their most recent start? You know, "striking while the iron is hot"...and all that...

[PaceMasterT]

Originally Posted by Dave Schwartz

I must be on ignore with a lot of people. Or else nobody believes me.

Eliminating based upon 30 days is just a bad idea in just about every category.

Sweet spot is in yellow. 26-42 days.

Dave,

Would you say that this phenomenon is due to the fact that a lot more horses run on 30 days or greater rest than they used to? As the win percent looks to be about even across DSLR, would you say that DSLR has no handicapping value? Look at Keeneland today, 3 races with all layoffs and/or first time starters, 4 races with 1 horse that ran in past 28 days, 1 race with 2 horses coming back within 28 days, and 2 races with more than half of the horses coming back within 28 days. Sure seems like the odds are stacked in favor of the layoff horses.


I know the "IV" is slightly skewed to longer layoff, but that is probably from old guard underbetting layoffs.

[Dave Schwartz]

Has anybody thought of doing a recency study exclusively on those horses who won or just missed winning in their most recent start? You know, "striking while the iron is hot"...and all that...

Thaskalos,

I never have but it is a good idea. Also, perhaps looking at horses that finished close and a few other related things, like stretch gain vs loss, etc..

Would you say that this phenomenon is due to the fact that a lot more horses run on 30 days or greater rest than they used to? As the win percent looks to be about even across DSLR, would you say that DSLR has no handicapping value? Look at Keeneland today, 3 races with all layoffs and/or first time starters, 4 races with 1 horse that ran in past 28 days, 1 race with 2 horses coming back within 28 days, and 2 races with more than half of the horses coming back within 28 days. Sure seems like the odds are stacked in favor of the layoff horses.


I know the "IV" is slightly skewed to longer layoff, but that is probably from old guard underbetting layoffs.

PaceMaster,

If you think about it, IV has nothing to do with betting. That would be PIV (Pool Impact Value).

I think the biggest reason is that trainers have read those books, too, but, yes, there are changes in trends.

It is like the 2nd-after-layoff angle in sprints that Quirin wrote about. As I recall, that factor has little, if any value.

Another one that is related comes from Fred Davis: Times on track in last 30 days (i.e. races plus works). It used to be (as I recall) a good thing about 15 years ago to have 4 or 5 activities. Now, 5 is too many and, in some kinds of races, 4 is also too many.

Simply put, racing has changed.

Oh, and one more, to illustrate the point: Recall "Suspicious drop downs." That is, horse runs a big race (win or lose) then drops 2 or 3 levels. We were trained not to bet those, so in the end, they are not bad signs after all.

[Capper Al]

Dave,

You're stat stuff is great in P&P 2012. I must have read it at least 4 times. The question is in the conclusion reached. Most stat and system players look for the happy path. This would be a place like where getting 10/1 odds on something that occurs one out of four times would be a simple mathematical bounty, a sort of fixed roulette wheel. With a take-out and a knowledgeable public, this isn't going to happen. Your whole book supports this premise and so do many other authors. The fact is that when you find a losing proposition, you have found a bucket of losers. Are all those in the losers bucket bad? Of course, not. Some losing buckets will never pay out, but some will. A handicapper is forced to go threw the trash to find value using his know of horses and the racing game. By reapplying himself and handicapping at this point, value may be found. With your large database advantage, you might want to run a secondary query on what traits wins in the losers DSLR elimination bucket. In the end, with all the machines and math, it still comes down to handicapping.

[Johnny V]

My thinking on this has undergone a big change over the last few recent years. DSL was always a factor for me in the claimers and still is somewhat although to a much lesser degree these days. DSL IMO can safely be seriously downgraded in most cases in the current racing enviorment.

[Lemon Drop Husker]

I have no stats to back up this claim, but the older and more experienced the horse the less inversely proportional the layoff effects their next out ability.

I tend to stay away from races in which a horse that scares me, or interests me, that is off a long layoff. I simply have never seen any pattern whatsover to bet for or against.

[Tom]

HTR studies similar to Dave's on layoffs.
This not your grandfather's game anymore.

[JohnGalt1]

Dave, I like the data you provide here.

As you know statistics without context can be misleading.

By that I mean, have you broken down your data into sprints or routes as first race after lay offs? I've read that a route horse wins more frequently than sprinters after long lay offs since longer races are less stressful than sprints.

Of course there is no way to know if lay off horses have had quality work outs before their comebacks. Quality work outs being frequent and at long enough distances for a prime effort.

Statistics can be a useful guide, but we all, (or at least I do) handicap each race as a unique event. If I handicap a race and like a horse with a 4% trainer ridden by a 3% jockey I know I'm fighting the statistical of winning, I will bet it and take my lumps, or reward.

Your example of double drop downs may not be a negative anymore since purses are so high at some tracks that the owner may not care as much if it is claimed.

Another bit of advice I've read in older books is never bet a horse doing two new things--a dirt sprinter trying a turf route for the first time as an example. That's where higher odds lie if I can scope out a likely winner.

[Dave Schwartz]

The question is in the conclusion reached. Most stat and system players look for the happy path. This would be a place like where getting 10/1 odds on something that occurs one out of four times would be a simple mathematical bounty, a sort of fixed roulette wheel.

Yes, the handicapper looks for a happy ending even if none is to be found.

Get this: There are no factors that, by them selves, will show profit. NONE.

With a take-out and a knowledgeable public, this isn't going to happen. Your whole book supports this premise and so do many other authors. The fact is that when you find a losing proposition, you have found a bucket of losers. Are all those in the losers bucket bad? Of course, not. Some losing buckets will never pay out, but some will. A handicapper is forced to go threw the trash to find value using his know of horses and the racing game. By reapplying himself and handicapping at this point, value may be found. With your large database advantage, you might want to run a secondary query on what traits wins in the losers DSLR elimination bucket.

They are all losing propositions to one degree or another.

In the end, with all the machines and math, it still comes down to handicapping.

What I hear you saying is that if I do not look at the statistics, I can just figure it out.

I call that the "head in the sand" approach.

IMHO, people win or lose by understanding (you can call this "handicapping" if you'd like) the horses in a given race and situation. The stats tell you what to measure but the measurement must be race-specific.

As an example, let us say that you still subscribe to that incorrect theory that a layoff is a bad thing (despite the fact that statistics indicate you'd be better off preferring the layoff horse over the horse back in 12-14 days).

If you find a race where all of the horses are layoff horses then the impact of the factor is nullified completely.

I am not suggesting that this is a handicapping approach, I am suggesting that the player sinks or swims based upon his analysis of individual races.

If someone asks me why did I play a horse in Race A the answer could be because he had a class edge that resulted in a positive expectancy, while in Race B it could be because he had an early speed edge that resulted in a positive expectancy.

The commonality is, of course, positive expectancy.

Ultimately, that becomes the question in EVERY race: "Which horses have a positive expectancy?"

Most handicappers simply will not do the work to answer that question.

[Tom]

I think the value of the data is the generally, the layoff is not as important as it once was. Don't use it as an elimination factor, but keep the horse and dig deeper - distance, trainer, class, sex.......you might find that the trainer NEVER wins over 60 days, then eliminate.

[raybo]

Generally (but not exclusively, due to my sometimes having questions or suspicions about a horse's condition (either positive or negative)):

Any horse with a layoff of 45 days or more gets a good second (old school) look, not an automatic elimination from win contention.

Conversely, if a horse is coming back in less than 15 days I will also take a good second look, looking for negatives that allow me to eliminate that horse. If there are none then that factor is ignored.

[Capper Al]

What I hear you saying is that if I do not look at the statistics, I can just figure it out.

I call that the "head in the sand" approach.

Dave,

I am not doing this. Why would I read your book on stats four or more times if I did? The stats are as given and set the stage and scope with which the horse must be evaluated within. Not every bucket of losers has value horses in it. One has to look at the bucket and hope with their racing knowledge that they can find a horse worthy of taking a chance on.

[Dave Schwartz]

Why would you look at "a bucket of losers" for value when you could look at a "bucket of not-so-bads" for value?

Or do you mean look at a "bucket of horses not bet so much?" because that is a different column.

[acorn54]

i'll paraphrase what mike pizzolla said in one of his videos, this horses days away pattern raises doubts of his ability to win at 3-1; at 10-1 he has a good chance.
"let the bet make you"-mike pizzolla