Predicting Speed Rating from Past Ratings

[Cratos]

Originally Posted by thaskalos

What was tongue-in-cheek...what I said about this game not being a mathematical problem waiting to be solved? I was dead serious. I have seen you post repeatedly about your belief that this game is nothing but a matter of physics, and I've disagreed with this opinion of yours...but you are entitled to it, so I've never stated to you that this opinion "leaves a lot to be desired". And when you finally decide to grace us with a concrete example of what you really mean, I might even embrace this opinion of yours myself.

I may be often wrong...but I am not stubborn...

I don't believe you are 'stubborn" and I enjoy your posts, but for you to say that many believe this. game is a mathematical problem waiting to be solved when the most popular handicapping tool in TB handicapping today, the speedfigures methodology is derived from a mathematical formulation is puzzling.

Yes, animal racing (horseracing), human racing, and machine racing is all supported in analysis and prediction from mechanics, a branch of physics.

[Tom]

Yes, and the races are numbered 1,2,3.....numbers.
But no one needs to use math to win.

[thaskalos]

Originally Posted by Cratos

I don't believe you are 'stubborn" and I enjoy your posts, but for you to say that many believe this. game is a mathematical problem waiting to be solved when the most popular handicapping tool in TB handicapping today, the speedfigures methodology is derived from a mathematical formulation is puzzling.

Yes, animal racing (horseracing), human racing, and machine racing is all supported in analysis and prediction from mechanics, a branch of physics.

I don't know what other players think about speed figures...but I consider them to be only one piece of the handicapping puzzle. And, as Beyer himself has stated, I believe that the over-reliance on the speed figures is a sure pathway to the poorhouse.

[Cratos]

Originally Posted by thaskalos

I don't know what other players think about speed figures...but I consider them to be only one piece of the handicapping puzzle. And, as Beyer himself has stated, I believe that the over-reliance on the speed figures is a sure pathway to the poorhouse.

For clarity I wasn't speaking of "player reliance", but "player populaiity" . However the pusillanimous player might disagree with me.

[whodoyoulike]

Originally Posted by Cratos

For clarity I wasn't speaking of "player reliance", but "player populaiity" . However the pusillanimous player might disagree with me.

Did you just call him a ...?

[HUSKER55]

did I ever tell you guys and gals about the merchant who couldn't sell lemons 3 for $0.26?





Made a new sign that said Lemons $0.10 per pience and sold out in an hour.


that is a class act.

[TrifectaMike]

Originally Posted by thaskalos

IMO, trainer intent plays a big part in the performances of these horses. That's why I find myself shaking my head whenever we start talking about using intricate mathematical calculations to predict energy expenditure, and other things of that nature. This game is not a mathematical problem waiting to be solved, like the theory of gravity, which was always sitting there, waiting for Newton to discover it. These are living, breathing creatures of habit...and the horses have their own idiosyncrasies as well.

If this were really a mathematical equation waiting to be solved...then the mathematicians would own their OWN tropical islands.

Thask, you've made some good points, but not entirely accurate.

The game is a mathematical problem waiting to be solved, "unlike the theory of gravity".

When one, as yourself, speaks of a mathematical solution or equation, you imply what is referred to as a physical theory.

Physical theories allow us to make predictions: given a complete description of a physical system, we can predict the outcome of some measurements. This problem of predicting the result of measurements is called the modelization problem, the simulation problem, or the forward problem. ( "findings of Newton and Galileo...")

Horse racing as a mathematical problem is "viewed" as an inverse problem.

The inverse problem consists of using the actual result of some
measurements to infer the values of the parameters that characterize the system.

While the forward problem has (like in deterministic physics) a unique solution, the inverse problem does not. The inverse problem has multiple solutions (in fact, an infinite number).

Precious Mike

[classhandicapper]

Originally Posted by thaskalos

Whenever I see a horse get trounced in a race when going up one level in class, even though the horse's pace and speed numbers indicate that it was the fastest horse in the race...then I don't believe that the horse lost because it was victimized by those lightening-quick mini-fractions that horses supposedly throw at one another during the "classier" races. Nor do I suspect that the horse lost because it was intimidated at a glance by the supposedly classier competition that it faced. I think the defeat is explained by the fact that the horse was not WELL MEANT in the supposedly classier race.

The pace and the speed figures may indeed say that a horse is the fastest in a given race...but those figures won't be of any help if the horse's trainer thinks that the horse doesn't belong there...and is using the race as a conditioner for a future endeavor.

Your latter point (trainer intent) is relevant, but let's leave that one aside because I am not a mind reader.

What if the data shows that horses that otherwise look equal disappoint way more often on the class rise than they do if they stay in the same class?

What if the data shows that horses that were well beaten improve their figures on the class drop more often than horses that stayed in the same class?

Then, when you looked at the average paces, final times, and other qualitative measures you determined that there were differences between those class levels?

Wouldn't that convince you to consider class moves no matter what the cause?

If you want to remove trainer intent, then I would say you should focus on class moves at the Graded Stakes level. Most of those horses are trying because it costs money to enter, ship etc.. IMO, if anything, the significance of class is even easier to appreciate at the stakes level. Horses that don't look much slower than the best horses in the race get crushed in major Grade 1 races all the time.

[JohnGalt1]

Originally Posted by cj

Since this thread has turned towards the elusive topic of "class", I'll challenge the advocates to point out a few plays (and why) before the races. The results don't matter, just want to see how class is applied. No retro fits.

I use a numerical measurement of class that was created by William L. Scott in the book "Total Victory at the Races." He called it Performance Class Ratings.

It measures how a horse competed or performed at the various classes and against the number of opponents it faced, (a horse beating a field of 12 is more impressive than if he beat fields of 5 or 6.)

I have modified his method and more structured class rankings for todays racing climate. I developed a chart for easy comparison.

It is just after 8 AM CT.

I will post my class ratings for Saturday 12/13 GP race 8 7.5 f turf 100k open Stakes. Ranked in their running styles with abbreviated names.

E
none

EP
1-Excaper-----------283(-3) ratio of higher class races in past
3-Asset--------------373
5-Tiger---------------187
10-Howe-------------215(-3)
11-Fredericks--------223

P
2-Vinny--------------230
4-Breitling-----------187
6-Bad Debt----------224
7-Mshwish----------(230) 3 American races
8-Midnight-----------167
9-Strong-------------184

The winner will be among the top 3 PCR horses about 67% of the time.

In this race there are 2 standout ratings but third place is a tie between the 2,6,7,10, and 11. 10% is roughly the separating factor.

PCR is only about 1/3 of what I do. Condition and speed/pace is just as important depending on the race. Turf and dirt routes PCR is slightly more important and speed/pace can be more important in sprints.

In this race I will bet the to win and play over the in an exacta. Other bets will depend on scratches or odds.

Race 9 1 1/16 dirt $100k Stake

E
1-Sr Quis------------142

EP
2-Csaba-------------200
3-Rose---------------225 (if male the rating would be 274)
4-Parks--------------340
6-Schivarelli---------320
7-Valid---------------349
8-Liam's-------------(383) 3 races

P
9-Hy-----------------283

S
5-Hampstead--------one/dirt race

PCR's can only be computed if 3 or more races and if 3or 4 races I put in parentheses to be used as advisory ratings.

The Filly is rated lower because races against males I rate higher than against females. I rate equal a female open 50k claiming race to a male 32K open claiming race.

In the above race the 6,7 and 8 are the top 3, but because the 8 is based on only 3 races I would also consider the 6.

I am passing this race because most horses are too close in pace/speed ratings.

I've posted this because CJ asked if class could be quantified, and to me it does, and it works for me.

The above 2 races are meant to show what these numbers look like.

You can consider them when you handicap these races with your form and speed figures.

THESE ARE NOT STAND ALONE FIGURES, and must be used with other factors.

Obviously they are not computed in races where no one has at least 3 races, or even when most don't have 3 races.

I only rate dirt races if the race is on dirt and turf races for a turf race.

The numbers look like a baseball batting average. Who do you want to bat clean up (or bet your money on) the .305 with some pop in his bat (speed) or the .212 hitter who "might" drive in the winning run (win the race?)

The absolute best performance of these rating is traditional starter allowance races--raced for claiming price of x since y.

They do best on Claiming Crown, since they are just starter allowances with huge purses.

[classhandicapper]

Originally Posted by thaskalos

I don't believe those par charts that I see...which have speed and class peacefully coexisting all the way from the top to the bottom of the class ladder. I honestly believe that these par charts have been massaged, in order to give the impression that the equine world is much more orderly than it really is.

They are definitely massaged because the samples are way too small in many cases and the differences between some classes are small enough that even large samples won't produce a nice neat chart. That doesn't mean an informed massaging won't yield very useful results.

I consider that more of an indictment on speed figures than class handicapping.

Speed figures are a function of those charts. Class handicapping is an attempt to bypass issues like parallel time charts, run ups, rails up or down, timer malfunctions, wind, changes in track speed during the day, track maintenance during the day etc...

[Greyfox]

Originally Posted by JohnGalt1

I use a numerical measurement of class that was created by William L. Scott in the book "Total Victory at the Races." He called it Performance Class Ratings.

Years ago I developed a methodology and program based on that book of Scott's.
It worked very well but was very tedious to do.
I really respected this late author's work.
However, when I was in The Gambler's Bookstore in Las Vegas, I noticed that there were several copies of it that buyers had turned back to them.
I guess the amount of work his ideas required wasn't very popular with horse players.
I ultimately gave up using the model I had developed because of the tediousness of it.
Yet the one thing it demonstrated to me was it is quite possible to win at the races without using Pace and Speed figures.

[Tom]

Yet the one thing it demonstrated to me was it is quite possible to win at the races without using Pace and Speed figures.

One of the options in HTR is what pace line method to use. I frequently turn it off so that no speed or pace at all is used, so I am only looking at stuff not mainstream.

I call it the No Math Method!

[classhandicapper]

CJ asked that a "class handicapper" review a "class play". Rather than do that, I'm going to illustrate an example of where "class handicapping" can get exceedingly complex and confusing.

The 9th race at GP today is the 100K Harlan's Holiday stakes. It's ungraded, but there are a few horses in the race that have been competitive in graded stakes company and others that may have their best days ahead of them. So IMO, it's a solid field for this level.

For me, the key horse is Liam's Map because he's the horse that is tough for me to classify. He beat a NW1 ALW field by 10 lengths while being loose on the lead on a track, that, if anything, was tilted towards speed. He appeared to win with something in reserve. It was a very good effort, but he didn't beat much and it wasn't a tough trip.

A move from NW1 to this level is not massive, but it's certainly large enough that most horses that try it will fail miserably. The difference of course is that we know this is no ordinary NW1 winner. He crushed back to back fields. Even in his loss as a maiden he was well clear of the 3rd horse and the winner came back to win again.

Generally, I will try to beat horses like these. They will come in with flashy figures, take a lot of money, but not be up to the challenge against better horses and the tougher trip (at least yet). There are exceptions though. Some horses like this are simply exceptional horses that haven't revealed what's in the tank yet (Graydar and Cross Traffic are 2 recent examples from Pletcher).

This is an example where raw class handicapping is telling me to play against this horse if he gets bet heavily, but practical experience is telling me I shouldn't be so anxious. This is a lightly raced, dominant, well bred horse, from the Pletcher barn. He might actually be the class of the race.

This may seem like a very unsatisfactory analysis because I'm expressing no real opinion on the horse or the outcome. But it shouldn't be. I'm illustrating one of the complexities of handicapping a race without figures and an awareness that sometimes you CAN'T know how classy a horse is until after the fact. For figure handicappers, the analysis is probably as simple as does he have the top number and will he duplicate it or improve.

There's no way I'm going to bet on Liam's Map. But it going to take a bias or some other advantage to get me to pull the trigger against him.

[cj]

Originally Posted by classhandicapper

CJ asked that a "class handicapper" review a "class play". Rather than do that, I'm going to illustrate an example of where "class handicapping" can get exceedingly complex and confusing.

The 9th race at GP today is the 100K Harlan's Holiday stakes. It's ungraded, but there are a few horses in the race that have been competitive in graded stakes company and others that may have their best days ahead of them. So IMO, it's a solid field for this level.

For me, the key horse is Liam's Map because he's the horse that is tough for me to classify. He beat a NW1 ALW field by 10 lengths while being loose on the lead on a track, that, if anything, was tilted towards speed. He appeared to win with something in reserve. It was a very good effort, but he didn't beat much and it wasn't a tough trip.

A move from NW1 to this level is not massive, but it's certainly large enough that most horses that try it will fail miserably. The difference of course is that we know this is no ordinary NW1 winner. He crushed back to back fields. Even in his loss as a maiden he was well clear of the 3rd horse and the winner came back to win again.

Generally, I will try to beat horses like these. They will come in with flashy figures, take a lot of money, but not be up to the challenge against better horses and the tougher trip (at least yet). There are exceptions though. Some horses like this are simply exceptional horses that haven't revealed what's in the tank yet (Graydar and Cross Traffic are 2 recent examples from Pletcher).

This is an example where raw class handicapping is telling me to play against this horse if he gets bet heavily, but practical experience is telling me I shouldn't be so anxious. This is a lightly raced, dominant, well bred horse, from the Pletcher barn. He might actually be the class of the race.

This may seem like a very unsatisfactory analysis because I'm expressing no real opinion on the horse or the outcome. But it shouldn't be. I'm illustrating one of the complexities of handicapping a race without figures and an awareness that sometimes you CAN'T know how classy a horse is until after the fact. For figure handicappers, the analysis is probably as simple as does he have the top number and will he duplicate it or improve.

There's no way I'm going to bet on Liam's Map. But it going to take a bias or some other advantage to get me to pull the trigger against him.

All that for the usual "passhandicapper"?

[cj]

Originally Posted by JohnGalt1

I use a numerical measurement of class that was created by William L. Scott in the book "Total Victory at the Races." He called it Performance Class Ratings.

It measures how a horse competed or performed at the various classes and against the number of opponents it faced, (a horse beating a field of 12 is more impressive than if he beat fields of 5 or 6.)

I have no doubt it can be measured. I like that you are using it as just part of the equation. That is how any handicapping factor should be used. The toughest part of handicapping is knowing which factor is most important in each race. Actually, I should say which factor offers the best value in each race.