Will Benny the Bull's BSF be higher than R1 winner Desert Key's?

[Bobzilla]

The fractions in Belmont 6/7/08 races 1 and 6 were as follows:

R1: 22 2/5, 45, 56 3/5, 1:08.80
R6: 22, 44 4/5, 56 3/5, 1:09.06

Though I believe BTB ran the better race in terms of performance, the BSFs aren't really suppose to take trip, race dynamics, ground loss, etc.,etc. into consideration. It's basically a figure assigned for the final time of the race itself normalized for the varying speed of the surface.

I've been doing my own figs for years. More often than not my figures are within a few points of what the Beyer Boys assign. I think they do admirable work. I do feel, however, they sometimes adjust the numbers disproportionately higher if the race was a graded stakes event, such as Benny's win in the True North yesterday. They probably do this to make the numbers seem more believable. I was at Belmont yesterday, as I am every year on this day, and I'm pretty confident it was the same track, for the most part, between 1202 EDT and 1502 EDT.

I would have given the Alw N1X Race 1 a 101 number, and the True North a 97 or 98. Because that suggests, to many, that Desert Key ran a better race than Benny The Bull did in the stakes race I think it's probable the official figures will have BTB's number much higher than DK's and the asumption of a higher variant between 1202 and 1502 will be the argument as to why. The 98 number would be more in line with with what Man of Danger probably would have earned yesterday while only losing by a neck to a courageous Benny. It was very tough to make any kind of move from anywhere in the pack on the lead yesterday, at all distances. In terms of my perfromance figs I have BTB's number vary high. But I'm only talking BSFs here.

Anyway, what do you think? Will Race 1 or the True North have a higher BSF?

[john del riccio]

I love questions like this.....

Race 1 77 pace 79 final
Race 6 78 pace 80 final

So the answer to your question is YES, on my figs.

John

[Robert Fischer]

Will Benny the Bull's BSF be higher than R1 winner Desert Key's? (showthread.php?p=538930#post538930)

Yes.

When a well regarded horse with big BSFs runs slower, even in a better overall performance, then you are supposed to falsify the track variant. :)

Splitting Variants is not reserved for drastic shifts in track weight. Splitting can be utilized when star horses run in slower final times.

You have to pervert the adjusted final time figure in order to meet an expectation from a projection. The track weight has nothing to do with the track, it is determined by how a few horses meet their BSF projection expectations.

[cj]

Will Benny the Bull's BSF be higher than R1 winner Desert Key's? (http://showthread.php?p=538930#post538930)

Yes.

When a well regarded horse with big BSFs runs slower, even in a better overall performance, then you are supposed to falsify the track variant. :)

Splitting Variants is not reserved for drastic shifts in track weight. Splitting can be utilized when star horses run in slower final times.

You have to pervert the adjusted final time figure in order to meet an expectation from a projection. The track weight has nothing to do with the track, it is determined by how a few horses meet their BSF projection expectations.

I haven't done the day yet. Are you saying the track speed never changes during a race card?

[john del riccio]

I haven't done the day yet. Are you saying the track speed never changes during a race card?CJ,

I didnt have the track changing speed. The dynamics between the two races was very different.

John

[cj]

CJ,

I didnt have the track changing speed. The dynamics between the two races was very different.

John

I understand, though like I said, I haven't looked at the card. However, a "pure" speed figure should not take the race dynamics into account.

[john del riccio]

However, a "pure" speed figure should not take the race dynamics into account.

CJ,

This is where I think we will have to agree to disagree...it is a very important
aspect of how I make figs.

John

[cj]

CJ,

This is where I think we will have to agree to disagree...it is a very important
aspect of how I make figs.

John

I'm not disagreeing. I use race dynamics as well.

The thread is about Beyers though. If the track didn't change speeds, then the Beyer should be higher for the faster raw time. Beyers measure nothing but track speed and the final time.

[john del riccio]

I'm not disagreeing. I use race dynamics as well.

The thread is about Beyers though. If the track didn't change speeds, then the Beyer should be higher for the faster raw time. Beyers measure nothing but track speed and the final time.

I understand, I woul dbe interested in hearing what Beyer did...
did you see the question I asked you about the NY bred race on Met Mile day ? What fig did you give the other 1TM ?

John

[cj]

I missed the question, sorry. I'll have to go back and check.

[Stareagle]

I use two ratings - a pure speed rating and one that is adjusted for pace and other things. I've got DK ahead 95-94 on speed rating and a 95-95 tie on total performance.

I gave Da'Tara a 73 performance rating, which has to be the worst I've ever seen for a Triple Crown race. I can't even think of a worse one for a Grade I win.

[ezpace]

now to adjust for the dope factor.. should we say add 7pts in Ky..3 in NY,FL AND

CA off the charts

[Pace Cap'n]

I don't do figures, but just going by the final time, would not Secretariat have won by almost the same margin had he been in the race?

[Skanoochies]

I don't do figures, but just going by the final time, would not Secretariat have won by almost the same margin had he been in the race?
No he would be 38 years old now. Probably still been in the money tho. :lol:

[Bobzilla]

Looks like what most of us suspected was right, the final time of The True North yielded a higher BSF for its winner Benny The Bull than Desert Key's quicker N1X Race 1 this past Saturday. From what I read in Steve Crist's DRF blog today, BTB's race earned a 106, while DK's race earned a 104. The blog made mention of the track being judged faster for races 1-4. Though I suspect the actual figures probably should have been lower by 5-10 points, I'm at least encouraged to see that the delta between the two figs wasn't as wide as I had expected it to be, I was thinking the Beyer people would have the difference more like 10 points. Though I greatly respect the work of the Beyer Boys, I do think they take into account their own projected figs for top graded stakes performers like BTB a bit too much in regard to assigning the final figure, rather than treating the number as a pure speed figure arrived at from raw final time. I would only guess that if this is done, it's done so that the general public who use their numbers will not lose confidence.

[Cratos]

I use two ratings - a pure speed rating and one that is adjusted for pace and other things. I've got DK ahead 95-94 on speed rating and a 95-95 tie on total performance.

I gave Da'Tara a 73 performance rating, which has to be the worst I've ever seen for a Triple Crown race. I can't even think of a worse one for a Grade I win.

Races are independent events that occur in a moment in time that might never exist again or have ever existed before.

Therefore for you to say: “I gave Da'Tara a 73 performance rating, which has to be the worst I've ever seen for a Triple Crown race. I can't even think of a worse one for a Grade I win.” Shows a lack of understanding on your part of how events occur.

In the 1973 Wood, the great Secretariat ran a final time of 1:50.47 for the 1 1/8m race and finished third. At the time, some of his detractors were saying he was just another over-hyped Bold Ruler colt. I believe that rhetoric allowed him to go off in the Kentucky Derby at the largest odds of his 3yo career and his closing odds would never approach his Derby odds the rest of his career. By the way he turned in three straight stellar performances after his dismal Wood performance by winning the Triple Crown and setting track records in each of the TC races.

Lastly, if Da’Tara’s Belmont was the worse for you, then how do you rate Savara’s 2002 Belmont performance which was run in a slower final time than Da’Tara’s Belmont final time and run also over a “fast” surface?

[classhandicapper]

I recently had an email conversation with Beyer on this whole race dynamics/pace issue and the Beyer figures that get assigned.

I believe Beyer feels he wants to give out figures to the DRF audience that approximate the ability of the horses. If the pace may have impacted the final time, we should expect that he's often going to build the impact right into the figure via the track variant rather than give out a very slow or fast figure.

This is a complex issue because the interrelationships between pace and final time are not 100% clear, measuring pace accurately is difficult, and there are problems with building the impact of pace into the figure via the variant (it assumes that all the horses were impacted equally and that is clearly not the case).

For himself, I suspect that he makes careful notations of very slow/fast paces and does not adjust the figures.

Of course, he also believes that tracks change speed and makes adjustments for that too.

[cj]

If that is truly what he is doing, and he is, then they aren't really Beyer SPEED figures.

Also, as you know, the pace doesn't impact all horses the same in a single race.

[Cratos]

I recently had an email conversation with Beyer on this whole race dynamics/pace issue and the Beyer figures that get assigned.

I believe Beyer feels he wants to give out figures to the DRF audience that approximate the ability of the horses. If the pace may have impacted the final time, we should expect that he's often going to build the impact right into the figure via the track variant rather than give out a very slow or fast figure.

This is a complex issue because the interrelationships between pace and final time are not 100% clear, measuring pace accurately is difficult, and there are problems with building the impact of pace into the figure via the variant (it assumes that all the horses were impacted equally and that is clearly not the case).

For himself, I suspect that he makes careful notations of very slow/fast paces and does not adjust the figures.

Of course, he also believes that tracks change speed and makes adjustments for that too.


The relationship between pace and the final time of a race are interrelated with the final time being the dependent variable when the final race time is calculated. The so called “track variant” is nebulous in this discussion because you can’t derive a constant from two variables. The track surface varies with respect to environmental, maintenance, and use conditions. The horse’s performance varies with respect to class, physical ability/health, and pace.

Therefore how can the time of a race indicate or measure the variant of a track surface using those variables?

[Tom]

A lot of people do it very successfully everyday. While what you say might be technically true, there are proven methods that imporve the situation without ever finding a magic number of, say, Fast 4 for today. And after all, the object is not to describe a track variant number but to provide predcitive numbers. And that is really old hat. This isn't an exact science, which is why computer generated figs, like BRIS, ore DRF TV+SR will generally be not as good as one made by objective people.

[classhandicapper]

The relationship between pace and the final time of a race are interrelated with the final time being the dependent variable when the final race time is calculated. The so called “track variant” is nebulous in this discussion because you can’t derive a constant from two variables. The track surface varies with respect to environmental, maintenance, and use conditions. The horse’s performance varies with respect to class, physical ability/health, and pace.

Therefore how can the time of a race indicate or measure the variant of a track surface using those variables?

Believe me, I've been studying and debating this issue for decades.

To sum up, let's pretend there was 3 horse race and they all typically run a Beyer figures in the high 90s or low 100s. They get into a massive and suicidal duel that causes them all to collapse.

When the figure maker calculates the figure for the day based on the track speed all day, he comes up with an 80. It's obvious that the 80 does not reflect the ability of the horses. It reflects how fast they ran after engaging in a suicidal duel.

Most very sophisticated handicappers would prefer getting a speed figure of 80 because they would realize the pace was a major factor in that slow final time and figure. Perhaps they use their own pace figures or visual skills, but either way they would know.

Many casual players would look at the figure and think either the figure is wrong or that the horse's went off form.

The figure maker then has a bit of a problem.

Does he want to give out a figure that better reflects the ability of the horses for his mass audience (perhaps a 98-102) or does he want to appeal to the most sophisticated handicappers and give out an 80?

That is the ongoing debate and question.

IMO it depends on your audience and what they want.

However, making adjustments for pace has some very big problems because anyway:

1. It assumes that all the horses within the same race are equally impacted by the pace. That is clearly false because they have different levels of ability and run different fractions.

2. In some situations it's very difficult to seperate changes in track speed from the impact of pace and understand what really happened.

[Tom]

And it is all acedemic. Do the figs you come up with work?
That is what matters, not the philosophy behind making figures. No one ever makes a 100% accurate number - any time you apply a variant, you are altering reality to what you think it should be. What counts is if you make an assumption that pays off at the windows. CJ handles your situation quite nicely.

[classhandicapper]

CJ handles your situation quite nicely.

Tom,

I agree. That's because CJ understands the issues, has fairly sophisticated customers, and also gives out a performance figure (PF) that combines the impact of pace and the final time.

So in my example above, he might give out a pace figure of 130, final time figure of 80, and PF of 100 and all his customers would understand what happened.

[Tom]

It would seem the to be the responsiblilty of the handicapper to understand the numbers he is using.

[Cratos]

Believe me, I've been studying and debating this issue for decades.

To sum up, let's pretend there was 3 horse race and they all typically run a Beyer figures in the high 90s or low 100s. They get into a massive and suicidal duel that causes them all to collapse.

When the figure maker calculates the figure for the day based on the track speed all day, he comes up with an 80. It's obvious that the 80 does not reflect the ability of the horses. It reflects how fast they ran after engaging in a suicidal duel.

Most very sophisticated handicappers would prefer getting a speed figure of 80 because they would realize the pace was a major factor in that slow final time and figure. Perhaps they use their own pace figures or visual skills, but either way they would know.

Many casual players would look at the figure and think either the figure is wrong or that the horse's went off form.

The figure maker then has a bit of a problem.

Does he want to give out a figure that better reflects the ability of the horses for his mass audience (perhaps a 98-102) or does he want to appeal to the most sophisticated handicappers and give out an 80?

That is the ongoing debate and question.

IMO it depends on your audience and what they want.

However, making adjustments for pace has some very big problems because anyway:

1. It assumes that all the horses within the same race are equally impacted by the pace. That is clearly false because they have different levels of ability and run different fractions.

2. In some situations it's very difficult to seperate changes in track speed from the impact of pace and understand what really happened.
Let’s be real about this topic. It not about horseracing, it is about how to use normalized data to assess future performance.

The two most popular attempts at using that statistical methodology are personal credit ratings and scholastic aptitude tests. Both have failed miserably because of the variability in the data which cause too many false assumptions to be made.

Therefore when we look at speed figures we must clearly understand the slope of the curve from which those figures were derived and put that curve in a confidence interval which will tell us just how realistic was the performance or performances from which the curve was derived. This is a difficult task because generally a horse’s performance curse is downward sloping and non linear.

Can a statistical methodology be use in horserace handicapping? The answer is a resounding yes, but it is a lot more complex than speed figures. There are many horseplayers (many of whom post on this site) do very well and never use speed figures.

[Tom]

While that is certainly true, a great many here use speed and pace figs extensively and successfully. To us, it is about horse racing, not statistics.
46 doesn't believe in class, I weigh it heavily. Dave Schwartz bets as many races as he can, I bet a comparatively small number.

It's not the journey that counts, it's the destination (Win-town! Woo Hoo!)

[bobphilo]

I think there is some confusion over the role of statistics in handicapping, and even just what a statistical analysis is.

Handicapping is simply the process of predicting the value of a target variable (A), such as horse’s performance in a given upcoming race, from another, or usually several other, predictor variables (X,Y,Z, etc.) Whether one realizes it or not, that is a statistical analysis and it behooves a handicapper to learn how to do it right.

Statistics is not just some abstract exercise, but the practical process or tool we use to make real world predictions. No more, no less.

Bob

[Cratos]

While that is certainly true, a great many here use speed and pace figs extensively and successfully. To us, it is about horse racing, not statistics.
46 doesn't believe in class, I weigh it heavily. Dave Schwartz bets as many races as he can, I bet a comparatively small number.

It's not the journey that counts, it's the destination (Win-town! Woo Hoo!)


Not unless you know something that no one else knows, horseracing is profoundly bounded by two disciplines: statistics and physics.

You can argue with the wind about the validity of that assertion, but any knowledgeable person will tell you it is true.

Therefore it more about your understanding and application of those disciplines that will yield winning results and you don't have to apply complicated formulas to use them, however to win you will and must use them even if you don’t realize it.

You can’t derive a speed figure without using statistics unless it happens through serendipity

[classhandicapper]

It would seem the to be the responsiblilty of the handicapper to understand the numbers he is using.

Tom,

I agree with you, but let's be realistic. The average person that picks up the DRF (even if they do so fairly often) has never made figures, does not understand many of the relationships between pace and final time etc.... They probably don't even really understand the issue.

Many of the figure makers and some of the customers do, but they aren't going to explain it all to those that don't. The figure maker will make a choice one way or another to provide the data in whatever form they decide is best.

I have been lobbying for a notation of some sort in the DRF (in Simulcast Daily) to reduce the work load on people that would like to know when a figure was broken from the rest of the day. That would save time for a lot of people.

[Tom]

To the degree an average is statistics, yes, you are correct. But that is about the extent of it. When I made figs, I compared actual times to expected times, averaged the differences, and made a couple of variants.

I don't know how I can use stats and not know it????

[classhandicapper]

Therefore when we look at speed figures we must clearly understand the slope of the curve from which those figures were derived and put that curve in a confidence interval which will tell us just how realistic was the performance or performances from which the curve was derived.......

......Can a statistical methodology be use in horserace handicapping? The answer is a resounding yes, but it is a lot more complex than speed figures. There are many horseplayers (many of whom post on this site) do very well and never use speed figures.

We are mostly discussing the choice of methodology in presenting figures, not the figures, their accuracy or their usefullness.

Should a speed figure reflect the final time the horses ran or should it reflect the performance of the horses?

IMO, it should reflect final time. If you want to incorporate other factors like pace into the figure, it should be a seperate number and it should be called a "performance figure".

The problem with this is that many customers of figures do not really understand the questions and issues, let alone the answers. So if you gave them just final time speed figures, they would do a lot worse than if you gave them performances figures and simply never told them.

[Cratos]

We are mostly discussing the choice of methodology in presenting figures, not the figures, their accuracy or their usefullness.

Should a speed figure reflect the final time the horses ran or should it reflect the performance of the horses?

IMO, it should reflect final time. If you want to incorporate other factors like pace into the figure, it should be a seperate number and it should be called a "performance figure".

The problem with this is that many customers of figures do not really understand the questions and issues, let alone the answers. So if you gave them just final time speed figures, they would do a lot worse than if you gave them performances figures and simply never told them.


Let me understand this, “the means are justified by the end result?” That is what is implied in your posting: “We are mostly discussing the choice of methodology in presenting figures, not the figures, their accuracy or their usefulness. “

Therefore it is to you the presentation, not the accuracy or the usefulness of the speed figure.