Whales: A chink in the armor?
 

I am sure that some people get tired of me going on about the whales. If so, just ignore this post.

I study them because they're our major competition; the single thing that makes this game most difficult to beat.

Well, in truth, I study the IMPACT of whales.

The battle has been very one-sided. That is, every year the game gets tougher, as the efficiency in the factors becomes more even.

This post will illustrate how I came to this conclusion. Before we are finished, I will also tell you why I think this can actually work in our favor.


I have drawn races from Feb. 01 to July 15th from each of the last 2 years. 5f to 9f, fast tracks, 5+ betting interests, no FTS, and races with entries have been removed.


An example:
http://practicalhandicapping.com/des...0Stats-PSR.jpg

Let me tell you what you are looking at.

(This factor is Jim Cramer's Projected Speed Rating. It is very much like BRIS Prime Power.)

#1. We are concentrating on the $Net column. This is the return per $2 wager for the different ranks. (FH is actually half-the-field-plus-1 horse and does not include 1-2-3.)


#2. See how the $Nets flow downward? This is indicative of how aware the public is of this factor.

In the case of a high-level factor such as PSR or BPP, it actually means that the public is dialed into the factors that make up these "factors." (In our software, we call these "Objects.")


#3. What we are really interested in is the ratio between the top rank and the bottom rank. We call that the "$Net Ratio."

The higher the $Net Ratio is, the more difficult it is to make money with longshots because the bigger prices are going to come from the Rear Half (RH) of the field.

In other words, the better the public is at isolating on the "hot horses," the more they will be bet down. This will effectively take some of those RH horses and cause the winners to no longer be long-priced.

#4. What we want to do at #4 is compare the 2017 $Net Ratio with the 2018 ratio.

What we see in this illustration is that the PSR factor has become more efficient. That is, the ratio has "flattened."

BTW, look at just how bad a "rear-half" horse really is.

Also, note how there is a distinct flattening of the top 3. My belief is that this is indicative of (at least philosophically) the whales doing a better job of finding "true value" in those top 3 ranks.

As they get closer to the same $Net, they have been effectively rendered useless because they make no difference in the bottom line.


Let's look at another example.

http://practicalhandicapping.com/des...20Stats-FT.jpg

Here we are looking at a factor in our system known as "Composite Final Time." It is created from 13 different ways of looking at speed ratings, such as Last Race, Best-of-Last-2-Races, Average-of-the-Last-3-Races, Best Ever, etc.

It is safe to assume that to a large degree, these two tables illustrate just how dialed in the public is to the impact of speed ratings.

Look at the gigantic shift in efficiency in the ratio!

Understand that the normal annual shift (in major factors like these) has been something in the vicinity of 2-3%. That's kind of like saying, "The public is getting 2-3% smarter every year."

This particular jump was in the magnitude of 10%!

This is unprecedented. I check these figures almost every year and have never seen a jump like this.

By comparison, Quirin Early Speed Points have been tuned in for so long that they hardly change at all.

(As a side note, a logical question for a handicapper to ask themselves is, "What did I do this year to get 3% smarter?")


Let's do another.

http://practicalhandicapping.com/des...ats-PSROld.jpg

Here's an older version of HDW's Projected Speed Rating. It has been around much longer, does not correlate with the tote board as much as PSR (or BPP), and, like Q ES Pts, has been dialed in for several years.

Still, it moved downward over 3%. (129-125) / 125 =3.1%.

In the next post (which will take me a couple of hours) I will address the impact on the tote board as well as a summary of this.

I contend that there is a silver lining to this. I believe the whales have created a vulnerability and I will share it with you later today.

Of course, I welcome comments.


Dave

Good stuff Dave...looking forward to part II

Next, we look at the impact on the tote board.

http://practicalhandicapping.com/des...tats-PubCh.jpg

The odds change in the opposite way.

What we see here is a widening of the top-to-bottom $Nets.

This is the impact of even more winners being pushed towards the top. Now, some of this is caused by smaller fields.

Look at the 2017 stats. See how the $net was relatively even between the first 3 choices ($1.68, $1.66, $1.64), then dropped off at "FH" ($1.49)?

A summary of 2017 would be:

1st, 2nd, 3rd = best
FH = bad
RH = horrible

Now look at 2018. Thus far this year, we've seen the actual favorite rise a couple of cents, over the next two ranks.

A summary of 2018 would be:

1st = best
2nd, 3rd, FH = good
RH = horrible

In other words, favorites are returning more money at a slight expense to the 2nd and 3rd choices, while the FH has improved to become equal with 2nd/3rd. The RH has gotten even worse.

To say this in fewer words, the public is more dialed in on everything above the rear half. (This is shown clearly in the $Net ratio change.)

Another thing to say would be that some of the winning longshots of the rear half are being bet down so that they pay less.

Finally, there is this...

http://practicalhandicapping.com/des...ts-Summary.jpg

This is the summary table. I ran it against these key factors but usually run it against about 200 factors.

There are 4 metrics:
$Net Ratio
IV Ratio (Impact Value)
PIV Ratio (Pool Impact Value)
PubCh Ratio

Remember that these ratios are always rank1 divided by RH to produce a ratio except for PubCh which is the other way around. That is, it is the average public choice rank of RH divided by average public choice rank of the favorite.

That last one can be confusing. Essentially, the 1st 3 of these ratings are beneficial to the handicapper if the winners are pushed to the top. (i.e. better ranks mean more money returned.)

However, the public choice thing means we'd rather see the RH horses be HIGHER ranks for public choice.

The scoring column is merely all the ratios multiplied together.

The key point here is that the PubCh factor - that is, rank for lowest odds - has moved from rank of 5th to rank of 2nd in one year.

This is absolutely unprecedented.

What does it mean?
It means that the days of just picking longshots with no regards for how strong the low-odds horses are or aren't and being competitive are pretty much over.

It means that if you are going into a race with an idea that sounds even remotely like, "I just pretend that all the horses under 8/1 are not in the race," you are done before you get started, because the low odds horses are going to pound you.

Permit me to give you some sobering stats:


http://practicalhandicapping.com/des...20-1up2018.jpg

Do you want to connect on 20/1 horses? They come up 2.69% of the time. That is once every 37 races.


http://practicalhandicapping.com/des...1up-8h2018.jpg

What's that you say? You only play 8-horse fields and above?

Well, that will add a couple of points to your chances. You're now all the way up to 4.36%.

One of those occurs every 23 races. Hope you have the right horse in that race.

Let's be more realistic. (And back to all races again.)
86% of all winners are under 8/1.
79% of all winners are under 6/1.
70% of all winners are under 9/2.
61% of all winners are under 7/2.

What about 8+ horse fields?
81% of all winners are under 8/1.
73% of all winners are under 6/1.
63% of all winners are under 9/2.
53% of all winners are under 7/2.


So, with all this bad news, where is the good news?

That's coming as soon as I can get it written, which could take me a couple of hours.

Meanwhile, what do YOU think a GOOD answer could be?

It is quite possible that the game at any point will become unbeatable; the takeout, the inside information, the unfair advantage that large bettors seem to enjoy and the Darwinism that forms today’s crowd are strong reasons to believe that the game can no longer be beaten and has already reached an equilibrium that converts it to some type of a Lotto that is played with horses.

Still, my studies do support this conclusion.

Although I do not fully understand the point you are trying to make, I assure you that a relative simple machine learning approach can easily convert the game to a form of BlackJack (as opposed to Lotto) where the EV ranges somewhere between -1% to +1% ( I have absolutely no trouble to explain and demo how this can be done if there is some interest on this).

What might be more interesting in regards to what you are saying here, is that the related models are targeting lower prices that are usually indicated by composite ratings like prime power. The existence of these models are enough for me to believe that although what you are saying can very well be true (meaning that the specific metric is gradually losing its value) this does not mean that there are not straight forward responses to the improved awareness of the betting crowds.

Interesting but not really all that terrible.

Assuming you are a competent handicapper and can weed out all the really bad contenders it can still be viable. These quick numbers include entries, and all field sizes for 2018 so far.


20195 races 2018

mutuels
3258 races $4 payout or less
10003 between $4 and 10.01
4430 between $10 and $20.01
14433 between $4 and $20.01
1793 between $20-40.01
445 between $40.01-62
237 greater $62

>$20 mutuel is approximate 12%







48657 races year 2010

mutuels
6022 $4 payout or less
23840 between $4 and 10.01
35029 between $4 and $20.01
11630 between $10 to $20.01
4621 between $20-40.01
1132 between $40.01-62
783 greater $62

>$20 mutuel is approx 13%

Dave your info is a lot to take in but much appreciated. I tend to bet longshots more at tracks that have a higher overall win payoff. Usually that means larger fields and most of the time Gulfstream Park. Definately not NYRA tracks.

Well, my guess is that with the proper analysis and the right tools, we're seeing changes in the wagering patterns of the whales which may introduce some opportunities for the minnows.

I interpret your data and come to the following conclusion - the whales are being pressured to generate the same profits (or greater) than in the years past. Nobody likes to see profits shrink, and the world is run by bean counters, so this is hardly surprising. The whales have adjusted to this pressure by zeroing in and hammering the most likely winners, and not dutching the wagers as equally in the past. This is my observation this year, as I've seen numerous late moves on the tote for the winning horse being crushed like a red lamper in Hong Kong.

So perhaps the trick would be to find not just the vulnerable favorite, but to find the vulnerable favorite that the whales will hammer but then run out?

Since when do we automatically assume that whales win? Las Vegas loves Whales, because they lose. I realize that some people assume that there are whales who pound favorites and then get rebates but the rebates for win bets aren't that good. And the computer generated betting syndicates are betting exotics, aren't they?

Delta mentioned that the whales may have inside information, but you can't rely in inside information. Sure, there are so-called "hot" horses that are tipped and win but inside information is generally bad information. I personally know someone who lost a tremendous amount of money paying off jockeys and trainers for inside info in New York.

That does not even come close to what my database says.

You say 12% and I say under 3%. How can we be that far apart?

Was there ever a time when the bettor, who "pretended that all the horses under 8/1 are not in the race" ...WASN'T done before he got started? When was such a bettor ever profitable?

There is no way only 3% of races pay over $20. Are you sure you're reading his post correctly?

Only 3 out of every 100 races pays over $20? No way.

I think you mean 20-1 or higher, correct?

He posted over $20, not 20-1.

The other poster was wrong too. If the $20+ winners hit at a 12% clip...then they would be a winning system all by themselves.

Yes it is true....

ALSO...

Horses paying over $10 come in at 5%
This has been true for a long time

Mike

Out of 100 races run every day, you don't think there are around 12 that pay over $20? That doesn't sound farfetched to me...but it's only my gut backing me up.

And how would it be a winning system? You'd be betting every horse going off at 10-1 or higher...