What is REAL Handicapping and So what?

[VigorsTheGrey]

Originally Posted by ultracapper

As long as there is an inability to quantify the subjective, there is a chance for all handicappers to find quality winners.

I'm not even slightly familiar with the advanced programs available, or that even may be possibly created, but I'm reasonably confident that what I do on a daily basis would take a monumental effort to quantify accurately. There are many, many subtleties that just can not be effectively quantified. If they could be, they already would be.

I agree, many subtleties. This is especially true for physicality handicappers...everyone ‘sees’ different things and they often say ‘that horse LOOKS GOOD....but unless one points out specifically WHAT makes GOOD, then what good is that...? Just another’s guys opinion....there are certain things I look for but this process rightfully takes longer to access than a 10 second peek on a video board can offer...what would help would be a comprehensive list of said subtleties and a list of all the tools ‘in the toolbox’...

[Robert Fischer]

Fundamental Analysis = OK, I claim to have some competence here.


Technical Supercomputer Data A.I. WHALE = I don't necessarily want your prime power' or your 'odds line' algorithm, - but I DO WANT I want to buy your info, your products, books, Formulator, TFUS, Pace Makes the Race, Pace-view, the clocker guy, the guys on the grounds, the guy in the office, the ipad, the PC etc etc...

I want to use good instruments for my fundamental analysis

[ultracapper]

Originally Posted by VigorsTheGrey

I agree, many subtleties. This is especially true for physicality handicappers...everyone ‘sees’ different things and they often say ‘that horse LOOKS GOOD....but unless one points out specifically WHAT makes GOOD, then what good is that...? Just another’s guys opinion....there are certain things I look for but this process rightfully takes longer to access than a 10 second peek on a video board can offer...what would help would be a comprehensive list of said subtleties and a list of all the tools ‘in the toolbox’...

There is so much that you can see in a previous race, in the paddock, or in the post parade, that is nothing but meaningless, and again, so much that has significance to varying degrees. Very, very difficult to measure it in a numerical fashion.

I read a lot of the chart notes. A lot of them are very good, and very accurate, but absolutely meaningless in handicapping a horse's future races.

[Dave Schwartz]

Originally Posted by sjk

I am still making odds lines and betting overlays with success (happily for both of us since I have bought your data for at least a dozen years).

I can see a difference in the plays available. They are fewer for me and much more slanted towards high priced horses. I don't know if the whales are reluctant to bet the hoses with little other money in the pool or if they will take away that value in time.

With high odds plays the losing streaks are longer and since I see things evolving I can have no certainty that good results will return until they do.

If it is true that whales are operating on slim after rebate margins and the competitive landscape is changing, I would think it would be very difficult for them to distinguish a short term streak from a change that would dictate significant adjustments or retreat.

Yes, Steve. Happy that this arrangement has worked so well for both of us.

Of course, you are using an approach that rivals the whales in terms of fit-test-fit-repeat. That puts you in a group of maybe 1/10000? Certainly 1/5000.

Perhaps I should have said that making an oddsline is dead for the guy who is staring at a racing form. BTW, the prime reason for that is that the tote is such a moving target.

In fact, the first order of business for any handicapper (in my opinion) is to find a way to remove the last-minute-based-on-tote decisions. They will almost always work against you.

Whales... margins. I do not know what the current model is, but it used to be that the target was to lose 3.5% before rebate and make it up in volume.

BTW, I recently saw what the rebates look like for a really big player today. I was shocked at how high they actually are. I assumed that the tracks were being successful at getting the manure back into the horse. Apparently, that is only for the far-below-whale-level players. (Of which I am one.)

Back on track - lose 3.5%. So, with even their rebates dropping, how are they keeping up the profits? Assuming that they are (keeping profits up) it is likely because of sharper models.

As I showed last year, the $Nets (i.e. return per $2) have gotten very flat all the way down to 3rd choice and perhaps 4th choice. That means they are successfully optimizing.

That also explains why you (SDK) are having a more difficult time finding races and why those are mostly on longer prices.

Another BTW... The typical whale betting model allows them to "roll the bankroll" between 100 and 125 times a year.

Put another way... When Peter Wagner's group was reputed to have won $27m while betting $300m - a 9% advantage...

... Reality is that he won $27m playing out of a $3m bankroll! Looking at true return-on-investment, that is a 900% profit. Good work if you can get it.

[VigorsTheGrey]

My instincts inform me that Whale entities have:

1. Known and well-established and stable relationships with their host tracks that involve some sort of mutuel benefit.

2. They must have cut a deal with their host tracks whereby last second cancellation of large bets are allowed and routine, unlike the general public.

3. They must have the ability to apply track specific selection strategies that continuously update, test and re-test, allowing them to switch back and forth between the various strategies when and where applicable.

[coachv30]

Please excuse the ignorance here but i know nothing about these whale models...what is exactly being gauged through their high tech programs? Is it gauging the same handicapping factors that you and I are, but just to a broader scale or is it something else? In other words, what lands them on a particular horse in a particular race? Are they strictly toteboard based? Will an AI model identify a horse because it gets out of the gate and is dropping down today?

[sjk]

I don't know any whales but my program uses the information from the charts in every possible way I can think of (which would be your old school handicapping) to evaluate each horses performance and to predict what sorts of outcomes are likely with today's fields and convert that into odds.

There are items that are not contained in the past performances that are in the chart data when you have charts for all of the races (how did his competitors run back) and calculations sufficiently complex that a person could not do them (pace adjusted sectional time adjustments).

A computer can do things that I could never do in person even though I wrote the program and am capable of looking up all of the ingredient pieces.

[Appy]

Good post Dave. I can relate to several points you made.
My apologies for this otherwise worthless post, which focuses on "Whales... margins. I do not know what the current model is, but it used to be that the target was to lose 3.5% before rebate and make it up in volume."
That reminded me of a joke I heard years ago about a couple of boys out in the hills trying to make a living cutting, slitting, and delivering firewood at a price point undercutting the competition.
During another long, dirty, exhausting day of work they took refuge in the shade to have a few beers and try to solve their profitless dilemma.
One asked the other, "What are we gonna do?"
After some thought his partner replied "We just need to buy a bigger truck!"

[Dave Schwartz]

Originally Posted by VigorsTheGrey

My instincts inform me that Whale entities have:

1. Known and well-established and stable relationships with their host tracks that involve some sort of mutuel benefit.

2. They must have cut a deal with their host tracks whereby last second cancellation of large bets are allowed and routine, unlike the general public.

3. They must have the ability to apply track specific selection strategies that continuously update, test and re-test, allowing them to switch back and forth between the various strategies when and where applicable.

I'd say, you went 0-for-3.

1. Not with tracks. With ADWs.

2. Perhaps slightly easier than most, but most ADW customers have that.

3. Absolutely not. Global strategy is the way it works. Simply put, nobody will build specific systems for many tracks. Just impractical to manage.

BTW, the answer to #3 is why we can still win.

...used to be that the target was to lose 3.5% before rebate and make it up in volume."

I had a "whale-guy" tell me that playing 4 races per track across several tracks at -3.5% was worth at least $4m per year.

Might need to sharpen that to -2.5% now.

Please excuse the ignorance here but i know nothing about these whale models...what is exactly being gauged through their high tech programs? Is it gauging the same handicapping factors that you and I are, but just to a broader scale or is it something else? In other words, what lands them on a particular horse in a particular race? Are they strictly toteboard based? Will an AI model identify a horse because it gets out of the gate and is dropping down today?

Same handicapping factors. In fact, think of it as simply pushing the winners to the top to produce an accurate probability for each and producing an accurate final odds prediction, which (logically) includes the wagering in the pool thus far.

BTW, there is no A.I. It is generally some form of pseudo regression analysis.

[VigorsTheGrey]

Dave,

Re:

I'd say, you went 0-for-3.

1. Not with tracks. With ADWs.
2. Perhaps slightly easier than most, but most ADW customers have that.
3. Absolutely not. Global strategy is the way it works. Simply put, nobody will build specific systems for many tracks. Just impractical to manage.

BTW, the answer to #3 is why we can still win.
————

1. Ok, I got who the host was wrong, but the idea is right: That it is a stable relationship with mutual benefits...the players here look more like business partners than a traditional business/ customer relationship.

2. I could be wrong, and it may vary between tracks, but I think at most track-contracted facilities, very near off-time win wager cancellations are not allowed, or at the very least watched and frowned upon (for pool manipulation reasons, I guess)...I heard that the wager machines are programmed to reject large cancellations and thereby force personnel transactions in these cases...I don’t know...

...but with the ADW as host, this appears to be no problem so #2 is moot but not really wrong...oh well.

3. Your answer here is good to know and makes sense...the one person I read, who has posted here, that goes GLOBAL , is NorthAmericanClassRankings...

....He has some 35,000? North American Horses ranked by Class...I don’t know how he calculates that but I really like that he does it...I review it from time to time just to freshen up my memory on some of racing’s top performers...it is updated on a weekly basis...really wonderful to have THAT KIND of DataBase going for you...

...I can only envy the others who have built these kind of GLOBAL methodologies...I am a pen and form guy, who listens also to guys like Matt Bernier, Dan Illman, Mike Beer, Jude Feld and a few others...for additional insights...outside of that I mark up my form highlighting all the top metrics so I’m not totally in the dark...

...I also look for early gains for form reversals, post position reversal factors, and pace advantages....I do ok but consider myself a dinosaur...I also look at good wager structure in exotics, but have found that the exotics are getting hammered with favorites in the mix and often do not outperform strictly betting to WIN...

[JerryBoyle]

Originally Posted by Dave Schwartz

I had a "whale-guy" tell me that playing 4 races per track across several tracks at -3.5% was worth at least $4m per year.

Might need to sharpen that to -2.5% now.

$4mil a year in profit or handle?

[Seabiscuit@AR]

In some cases the big whales don't need to have a good relationship with an ADW. This is because the whale owns the ADW they bet with. So the whale is the ADW. You can get some very big rebates this way

[Dave Schwartz]

Originally Posted by JerryBoyle

$4mil a year in profit or handle?

Profit

[Saratoga_Mike]

Originally Posted by Dave Schwartz

Profit

Which would require $100 mm+ in annual handle, correct? That works out to almost $300k/day (in reality, higher on days that deep pools are available, meaning you can't bet as much on a Monday as a Saturday). Let's say such a bettor plays $10k into a typical race and plays 30 races/day. Here's the question: what's the variability for such a player? In different terms, in a one-year period, what do you think the worst losing streak looks like?

[JerryBoyle]

Originally Posted by Dave Schwartz

Profit

That's pretty incredible. My total guess was that the best teams could maybe earn $3-6mil/year at this point, so I'm surprised he can do this playing only a few races at each track. I would have thought it required playing most races at most/all tracks (most/all pools).