Good pace figures

[GameTheory]

I want to move this discussion to its own thread.

CJ said something about pace figures being hard to make (compared to final time figures) and about BRIS & TSN pace figures being lousy.

I don't know anything about BRIS or TSN figures, but I don't get it. What is special/different about pace figures? And if you make a final time figure, aren't you in effect making pace figures too?

There are two basic components to most figures:

-- track/surface equalization (a variant)

-- distance equalization (so you can compare figures from different distances on the same scale)


What am I missing?

[cj]

Time for bed, but I'll throw out one short thing on the subject until tomorrow.

Most horses are only capable of running within a certain final time range depending on the track conditions. These same horses, however, are not restricted to this to the pace calls. It is more a reflection of the rating tactics of the rider in combination with ability. This is especially true in route races.

Just some food for thought til morning...

CJ

[GameTheory]

Yeah, but aren't we still just looking for a number that reflects how fast the horse actually ran at each call (regardless of how much or how little the horse was rated)?

Are you talking about making a pace variant to compensate for the effects of rating by the jockey? If so, that does sound difficult! I don't think I would even call it a pace figure anymore...

[Chico]

[i]

I don't know anything about BRIS or TSN figures, but I don't get it. What is special/different about pace figures? And if you make a final time figure, aren't you in effect making pace figures too?
What am I missing?

Pace and final time figures are adjusted for a track's par figures (based on 10K horses) and then given a daily variant. A race may run in any of 9 ways based upon the actual pace and final time numbers for a given class of horse:

Pace Final Time

Slow Slow
Slow Average
Slow Fast
Average Slow
Average Average
Average Fast
Fast Slow
Fast Average
Fast Fast
As you can see, the pace and final times distinctly differ from each other. A deep discussion of this can be found in one of the "All-Ways" newsletters which can be found at frandsen.com.

Regards,
Chico

[Tom]

Not the easiest things to make, but when they are good, they are very very good.
One problem is wind. You might have 30-40 mph wind blowing into the backstretch for the first 3 or 4 races, which slows them up, and then the wind stops and now you have normal times. What to do? Race by race wind data is hard to get. Also, some pace times are affected by one horse going out and burning rubber so to speak, and then is nowhere around at the wire. You can compensate for that by using the pacetimes of the winners.
Some races just plain defy logic. Is it the timer is tripped too soon, by a bird, or is it that they start the run up a varying distances?
The big problem is that you are dealing with a very small sample of sprints and routes. I try to lump days together whenever I can, but that is not always possible. Some days, you have only one route race or only one sprint race. Or a mix of really cheap races and one stake.
I never take one pace number alone as being meaningful - I always wnat to se either a familiar pace pattern or previous races with similar pace numbers. BRIS/TSN are volitile, but if you use several lines, they make a lot more sense.
I find the most useful plays come from some kind of pace advantage (no, not sucking up to PA, hehe) or pace key race.

[GameTheory]

Here's what I do. It is completely mechanical.


First, I establish a variant, which is based on final times only. I'm not going to go into how I do that, but it is also mechanical (the computer does all my figures automatically for me -- but I wrote the code).

So now I've got an adjustment that I'm going to apply to the final time. Let's say it is +1 second. (Note to Karl: I've studied this, and the variant *DOES* improve the figures by a measurable margin.)

Here are my original times for a 6 furlong race:

22.0 46.0 72.0

So my adjusted final time is 73.0.


Then I do distance equalization, which I won't go into either, but let's say my final figure for 73.0 at 6f is an 80.

I make the pace figures "Jim Cramer"-style. That is, I give them the figure that would gotten if they had run the whole race at the same speed they ran the fraction.

But first I've got to adjust the fractions according to the variant. To do that, I simply preserve the original ratio the raw fractions had to the raw final time. So,

22.0 / 72.0 = .305555 (ratio of original 2f time to final 6f time)
.305555 * 73.0 = 22.3 (multipy ratio by adjusted final time to get adjusted 2f time)

Then the 4f time:

46.0 / 72.0 = .638888 * 73.0 = 46.6

So my adjusted times are:

22.3 46.6 73.0


Now to turn them into figures I make final times for the fractions based on them running the same speed over the whole race that they ran the fraction:

22.3 * 3 (to make a projected 6f time) = 66.9
46.6 * 1.5 = 69.9

Then I do distance equalization on those two times -- 66.9 & 69.9 -- as if they were truly 6f final times and I'd get final figures something like:

110 95 80

I like to do them this way because I can easily see the way the horse distributes energy (and notice I took care to preserve the original energy ratio distribution).

I also make non-distance equalized velocity figures for the early fractions which are more useful for figuring out who's going to get the lead. I haven't decided whether raw or adjusted times are better for this.

And the other thing I do is make a set of pace pars "Randy Giles"-style (all of my ideas are stolen as you can see) where the pace par is the average ratio of the pace time to the final time for that particular track & distance. Then I can judge whether the pace was fast or slow given this particular final time.

All completely mechanized.


So what would I do to make them better? If it involves wind, I'll pass. I'm interested in this rating stuff though...

[cj]

GT,

I think we do things very similiar. I also use pace pars that are a % of the final time. The only difference I can see is that I use the pace time of the winner, not the leader, when making a pace variant. This is where the rating tactics come in that I talked about earlier. I jockey could gun some sprinter for 6f and set an incredibly fast pace time, but I don't want that calculated in my pace variant. I have found the winner's pace time to be much more reliable. I also try not to use Maiden Claiming races either. They are also generally skewed much too fast in relation to overall track speed.

Does any of this make sense? :confused: :confused:

CJ

[GameTheory]

So basically the only thing I might do differently is use the average pace ratio of the winner to the final time instead of the pace of race when making a pace par (ratio). That makes sense.

Any idea how big a difference that is in pace times? I'll have to look into that -- maybe there is infomation to be gleaned from comparing the two...

[cj]

One other thing, I don't make my own variant for the final figure, I extract the Beyer variant and use it MOST times, but not all. Sometimes the Beyer version is a "projection" as they call it, and I don't like those.

Another part of the pace being tough is similiar to final figures. There just aren't enough races run in the day at similiar distances. Especially California, you will get many days with one route race only. Makes it tough for final figs, really tough for pace figs. Aqueduct's main track is the same way, very few 2 turn races. It almost always happens on Wood Memorial day, making pace figs very unreliable for that race.

CJ

[cj]

Originally posted by GameTheory


Any idea how big a difference that is in pace times? I'll have to look into that -- maybe there is infomation to be gleaned from comparing the two...

You will notice huge differences in races where closers win, and, of course, these are usually the fastest paced races.

CJ

[Tom]

GT...I like your ideas - the only problem you might hit with your method is when the pace is very slow or very fast due to wind or somehting else. Yoou can compensate easily just by handicapping.
If a horse get your figs of:

100 85 78
102 84 80
95 79 79
101 86 81

I would be suspect of the 95-79 race and not use it. I would suspect your figs will be right most of the time and at most tracks, plus you have the edge of having them for many tracks(?) without all the hours of work trying to do them by hand.

[GameTheory]

I actually make only one daily variant for all races over the track (same surface, same condition), route and sprint. Most people consider this a big no-no, but I compared tons of data, and it was always better to include as many races as possible. My variants are "per furlong", so I take the difference between the actual time & the par, and then divide by the distance. This also removes the problem that the absolute differences are usually greater at longer distances. Then I take the average difference in time per unit of distance and that's the variant (sometimes there are minor adjustments if there is a real screwy race in there). For the times, I use the average time of the first few finishers, not just the winner.

So my variant is actually a multiplier that I multiply by the distance when I go to adjust the raw time. I completely ignore the one turn/two turns issue because the research showed that the variant was more reliable the more races I've got of any distance. Possibly there should be some tweak in there for the extra turn in routes, but it works pretty well anyway...

[Fastracehorse]

You said this: What is special/different about pace figures? And if you make a final time figure, aren't you in effect making pace figures too?

That is myyyyyyyyyyyyyyyyy bull-ieffff.

fffastt

[cj]

I guess I'm wrong, where do I turn in all the money I've won with pace figures?

CJ

[MV McKee]

Over the past few years I have attempted to use a couple of methods to try and extrapolate pace variants from my final time variant.

1) Use a fraction of the final time variant that represented the proportion of race distance that the fractional time represents.
Example: 6f 1/4, 1/2, 3/4 = .33, .33, .33

2) Use a fraction of the final time variant that represented the proportion of race time that the fractional time represents.
Example: 6f 22 46 72 = .306, .333, .361

I have never achieved my desired results when using ratings and variants derived from these formulas in practice.

In my opinion, this is because in using either of the methods mentioned above (in particular method 1) you are utilizing the final time variant as an expression of the speed of the track (and possibly the wind) only. I feel that a final time variant is a product of not only the effect of a racing surface on a horse's speed, but also upon his stamina.

I actually came to the conclusion that it was best practice to calculate the individual fractional variants on a day by day, track by track basis, but extrapolating these variants from the final time variant is generally quite accurate, and in cases where I cannot get a high confidence level for a given day using the former method, I will use the final variant(s) as the basis for their fractional counterparts.

However, I have found that the following proportions are far more accurate for doing this.
In the case of 6 Furlong races:
1/4 (1) .20
1/4 (2) .32
1/4 (3) .48
1/2 (1) .52
1/2 (2) .80

In other words, if I had a final variant that made the track 1 second fast, I would (on average) have the 1st 1/4 of a 6 furlong race adjusted .20 seconds fast. Of course this varies slightly from track to track and day to day, but from my research and practical use, this is a very solid average.

To me, the reason for this distribution became obvious (after about 3 years of having no clue). Like pace, a track variant (be it fast or slow) has a cumulative effect. When I am saying that the final 1/4 of a 6 furlong race is (.48 X Final variant) seconds fast/slow, I am not saying that the section of the track surface from the quarter pole to the wire is that fast/slow. Rather I am stating that the track speed/energy returned have a cumulative effect of X on the final 1/4. In the case of the First 1/4, there is no cumulative effect (unless you want to make a very fine-line distinction and argue that the first 1/8th affects the 2nd 1/8th, etc.)

Did any of this make sense?


Michael

[Tom]

I agree that the final variant usually takes into the account the pace variant, but not always and not always ditributed the same.
I see these times at AQU/AQI frequently

45 112
47 112

Same horse.

One was run into a strong headwind the other ws aided by a strong tail wind. The variant fo rthe day was X but how much of X was the first 4 furlongs and how much was the last fracton? It varies.
March29 at Aqu I made the variant F10 S5 for sprints and F15 and S5 for routes. there was a strong tailwind blowing in fast pace times and stiffling good stretch runs. To use S5 for the daily variant, and then spreading it out over those fast early splits would make them look even faster. The last race had a par of 106 and ran raw 116-108. Using my two variants, I gave the race a 106-113 AF for the class. Using only the S5 and putting 2/3 of it on the pace would make the race a 119-113 against a 106 par.
That is Derby -type numbers for a dimsal NY finale.

[Tom]

Todays races at AQU make no sense whatsoever as far as variants go. Sprints range from S11 to F5 for pace and Routes range from S9 to F20! This is a day you mark down as F'd up and not use any paceline from any of the races. Why this happens I can only guess. My thoughts are the teletimes was screwed up or the times were transposed by a blind drunk. You get this a lot, especially at those crazy tracks like NYRA - one turn routes for 8 miles or whatever it is....just ridiculous. It oly gets worse when Bel opens. This is one of the more powerful uses pace figs - knowing when the pacelines are not worth looking at.

[cj]

You made me have to go look...based on my numbers, the sprints (including 1 turn miles) had figures as follows based on the pace time of the winner and the final time:

-14, -19, -7, -1, -1, -7

Averages out to a -7.

Routes as follows:

-3, +7, +5

Averages out to a +3.

I really don't see anything too out of the ordinary.

FYI, these are Beyer style figures, so the sprint pace call variant was about 2 lengths slower than the final time, while the route pace call variant was about 1 length faster than the final time. I haven't done a final variant yet, but it won't change the relationship to the pace variant.


What am I missing?

CJ

[GameTheory]

No one is gonna yell at me for mixing sprints & routes in my variant? When you guys make a variant is it a fixed amount of time for similar distances? (i.e. 1 second slow would be applied to both 6f & 6.5f races) Mine are different for every distance since the variant is expressed as something like ".2 seconds per furlong of distance".

Originally posted by Tom@HTR
March29 at Aqu I made the variant F10 S5 for sprints and F15 and S5 for routes. there was a strong tailwind blowing in fast pace times and stiffling good stretch runs. To use S5 for the daily variant, and then spreading it out over those fast early splits would make them look even faster. The last race had a par of 106 and ran raw 116-108. Using my two variants, I gave the race a 106-113 AF for the class. Using only the S5 and putting 2/3 of it on the pace would make the race a 119-113 against a 106 par.
That is Derby -type numbers for a dimsal NY finale.

I'm wary of changing the structure/relationship of the figures to each other (pace to final) because that is one of the primary things I use them for. Also in your example, if you applied the S5 to the pace call, you wouldn't really be making it faster (relatively), because you're also making the final figure higher. It maintains the same relationship, even though it has a higher absolute number.

It is probably worth it to do it both ways depending on what you're looking for -- anytime you use a variant you're making up a fictional time anyway. I'm not sure you can call one variant more "accurate" than another since the whole point of using a variant is to move away from reality -- the actual raw time. You can call something more useful, but whether it is or not depends on what you are trying to achieve.

I would say the same thing about MV's method -- more "accurate" how? They're all fantasy times no matter how you do them. Maintaining the ratios of the fractional times to the final time is generally cumulative anyway because the later fractions are slower and therefore represent a bigger chunk of the total time. You might magnify that and generate better predictions, but who's say whether it better represents what happened in that past race?

Lots to research...

[Tom]

GT- I use your method at times, so I don't disagree with it at all. I use whatever seems to fit depending on the track and data.
I totally agree about all variants being fantasy - it is our attempt to explain the unexplainable.

CJ - Using race pars, for sprints I got
1 s11 s4
2 s4 f5
3 f1 0
6 f5 f3
7 f5 f7

Routes

4 s3 f1
5 s9 s9
8 f20 f11
9 f16 f9

I can only explain this if I conclude that races 1-5 had slower pace variants than the later ones. I hate to do this unless I can find a good reason - like rain happening mid-card. Yesterday, I see nothing to explain it. Normally, unsing HTR pars and HSH partimes, the variants are very close and very reaonable for most races - my thoughts here are that yesterday's card was very poor quality and the horses that won the early races were, shall we say, of questionable lineage (oink oink, bark bark). If I don't split the variant, I will give to much credit the early winners.

[Tom]

I assume from your data you count 1 turn miles as sprints?
If so, do you do the same thing for Belmont with 1 turn up to 9 furlongs?

[jotb]

Hello Tom:

In regard to the 29th of March, I totally agree with you that pace times were much quicker than par due to the tailwind. Here is my problem calculating the pace variant for this day. All of the races on this day except for the 2 turn race start down the backside. The four 8F races show the dramatic difference in pace times because of the greater run down the backside. In the 8f races the run to the turn is 2,639.5 feet which is 4F but the 6F races only run 1,320 feet to the turn which is 2F. Within this day there were two races at 6.5F and one race at 7.0F. These three races make it extremly difficult to accurately measure the racing strip because we now have the 6.5F races running an extra 330ft before the turn and the 7.0F race an extra 660 feet before the turn. Now if we take the 1/2 mile times on this day for the following distances( 6F, 6.5F, 7F and 8F) we have 1/2 the races already running around the turn and the other 1/2 running to the turn. I would imagine there is a need for 2 pace variants because of this obstacle but only 1 variant for the first 2F since all of the races start on the backside. The problem I have with this is the 6.5F and 7F races. These races have additional footage running to the turn but less footage than those 8F races. A small percentage of the 1/2 mile time is down the backside and the majority of the 1/2 mile time is on the turn. How do you calculate this mess?

Best regards,
Joe

[jotb]

Originally posted by Tom@HTR
Todays races at AQU make no sense whatsoever as far as variants go. Sprints range from S11 to F5 for pace and Routes range from S9 to F20! This is a day you mark down as F'd up and not use any paceline from any of the races. Why this happens I can only guess. My thoughts are the teletimes was screwed up or the times were transposed by a blind drunk. You get this a lot, especially at those crazy tracks like NYRA - one turn routes for 8 miles or whatever it is....just ridiculous. It oly gets worse when Bel opens. This is one of the more powerful uses pace figs - knowing when the pacelines are not worth looking at.

Tom:

There is no doubt in my mind that the maintance crew played a major role yesterday. They either watered down that track or sealed it after the 6th race. The 8th race time were extremly fast. How fast is fast? The final time was three quarter's of a tick slower than the best time at this distance for the last 3 seasons on the main track. The time was 148.10 and the best time in the last 3 seasons on the main is 147.96 at 9F. I certainly would not use this race in the sample to calculate the track variant for yesterday nor would I use the sprint race (7th race) or the route race (9th race). If you must, then I suggest seperate variants for race 1-6 and 7-9.....JMO

Joe

[cj]

Originally posted by Tom@HTR
I assume from your data you count 1 turn miles as sprints?
If so, do you do the same thing for Belmont with 1 turn up to 9 furlongs?

Tom,

Yes, I do count 1 turn routes as sprints, and at Belmont too. Same thing at Arlington and Colonial as well.

The only track I do separately is Ellis Park, where I make a sprint variant, mile variant, and a route variant (when they run any, that is.)

It also looks like you use the leader to establish your variants, not the winner. Am I correct?

CJ

[Fastracehorse]

You said this:

I guess I'm wrong, where do I turn in all the money I've won with pace figures?

Milk does a body good!

fffastt

[Fastracehorse]

Did any of this make sense?

I'm not sure but I like Game Theory's question: And if you make a final time figure, aren't you in effect making pace figures too?

That's what speed figs do: Pace, variant, class, and, final time.

I use the Beyer as my base but I imagine you non-DRF users could find other suscribers of decent speed figs such as Bris.

fffastt

[Tom]

Yes, I use pace of race times.
I am willing to try pace of horse times - I have everything in Excel so I can easily add another column or two.

[sjk]

McKee,

I use a table of values which is directionally similar to yours. At 6 furlongs it uses 15% of the variant to the 1/4 and 48% of the variant to the 1/2. In route races, the fractions are even lower (8% for the 1/4;25% for the 1/2;55% for 3/4).

I agree with the idea that tracks are fast or slow because they are less or more tiring and the effect becomes more pronounced in the later part of the race.

That is the adjustment due to variant. When I started making pace figures years ago that was the only adjustment I made and some days (high wind I guess) seemed to throw a wrench into the usefulness of the figures. Since then I have also used the average race segment times over each portion of the track to figure out if that part of the track is playing faster or slower than usual and make a second adjustment on that basis.

[oldstuff]

Many of us "use the figures" put out by Bris, Tsn, etc. I am one of them and absolutely have no time to make or adjust figures. I have to use 'what I am given'
Using TSN 2nd fraction 'times' and 'pars' I ended up with one 5 horse array in one position (1,2,3,4,5) in 'times' and another position (1,3,5,2,4) in pars.
My question? Which is correct? Why the difference? It's the same race. In short am I to believe the pars or the times?

As a former owner of a slew of horses, I always knew horses were rated "hence pace" to the 2nd call and from then on it was a "matter of time". Has anyone thought of just using pace up to the 2nd call only (for I said it is only for a horse being rated to that approximate point) ......and then rely on raw ability for the rest of the race? I for one would like to see two sets of completely different figures for the final pace fraction seems quite inappropriate to me. One reason is that you tell the jockey to ease up if the horse is not in winning or place contention at the end. Another is if the pace is too fast and your horse just throws the bit. And what if the pace is horribly slow....everyone ends up with a great late pace figure. I want to learn to use pace figures but just can't seem to relate. I am tremendously interested in this thread.

This may be a dumb post when you all are so eloquent in your computations and so knowledgable of pace figures but I think it has to be said.

[Fastracehorse]

Not even close.

<As a former owner of a slew of horses, I always knew horses were rated "hence pace" to the 2nd call and from then on it was a "matter of time".

And you left out that jockeys can wait on other horses even until mid- stretch - happens alot - further making pace figs 'rough' in many situations.

I'm not saying that adjd speed figs aren't rough in many situations but rough for one whole # is good enough. You will never achieve bulls-eye accuracy with any pace or speed fig. What you do want is that it will hit the target reliably.

What we are really measuring in a round-a-bout way is class. Class is also vulnerable to flaws in definitions of the abilty of different classes of race horses. That is, horses of an apparent weaker class often do beat higher.

We are measuring class when we make pace and speed figs because the pace is theoretically, and in most cases practically, stonger in higher classes.

I beleive that speed figs measure pace. Final time is a product of internal fractions. Fractions are reasonably consistent within every class level. Unusaually fast first quarters are very rare despite what the tele-timer says. That is why we have variants.

So, what I am saying is that trying to find 'exactness' is impossible. 'Round-a-boutness' is good enough.

There is more to the game than just #'s ( pace or speed figs ) and if the limitations of such # crunching hasn't taught you that, then you are limited.

Finally, since 'round-a-boutness' is potent, I 'm glad I only have one # to crunch. Alas, the adjusted speed fig.

fffastt

[anotherdave]

Originally posted by Fastracehorse@DRF
I 'm glad I only have one # to crunch. Alas, the adjusted speed fig.

fffastt

I know that the mystical ffffffffffast adjusted speed figure will be kept a secret like the Colonel's herbs and spices, but

alas?

Why such sadness and grief? I do my own numbers and they usually only bring me happiness.


AD

[sjk]

fastracehorse,

I would agree that you need to adjust speed ratings and that early pace rating is a significant element of the proper adjustment. At the same time, early pace in relation to the rest of today's field is a significant and independent factor. Anyway it works for me.

[TenZin]

My brother in law whose a member here and likes this room
says he knows fast really is a guy with another name here.

[Fastracehorse]

I used to do laborious chart work - now I am thankful for the one # 'thang.'

It sounds like you do pace #'s too - that may be a problem if you enjoy it. Why is that a problem lllastt?? Because you shouldn't really spend too much time on things just because they are enjoyable. I really beleive that is a problem for alot of people - they luv the 'old mighty # too much'. Are you addicted to # crunching??

Like hockey, and you are probably a Canucks fan, me an Oiler, you have to play in the trenches to get the job done. In otherwords, you can't be periphery type players like Naslund and Morrison. But at least you guys have scoring. We have Comrie - yikes.

I took my hits in front of the net and have been rewarded with the adjd speed fig. But enjoyment for me is analyzing a race from a horse-form perspective, of which a speed fig is only a part of the puzzle.

Lllasttly,

I thought it was funny when Vancouver lost first place. I'm not trying to be mean-spirited. You see, I was the biggest Canucks fan on the planet until the new ownership. Now it is best I don't think about them.

I'm looking forward to sharing insights on Hastings - if you aren't too mysterious - LOL.

lllastt

[Fastracehorse]

_________speed figures??

You said:I would agree that you need to adjust speed ratings___

You also said:
early pace in relation to the rest of today's field is a significant and independent factor.

I whole-heartedly agree. I do this by consisitency of the horse's placing at the first call. That is, lengths in front of or behind at the first call. Never use fractional times at the first call, my experience deems them inconsistent.

Unless of course you are using a fig outside of DRF.

fffastt

[Fastracehorse]

I can assure you your brother-in-law is mistaken.

My home forum is DRF - also Fastracehorse - and I enjoy Pace Advantage because of all the insightful people.

Thanx,

Fastracehorse.

[sjk]

fastracehorse,

Maybe I should say "speed figures" rather than "speed ratings". I mean the same thing. I make my own (as do several of the other contributors to the thread); I mainly do it for the sake of convenience and cost. I don't think they are significantly different or better than the ones in the publications.

Whatever your ratings (figures), I believe the adjustments we talked about above are important.

[anotherdave]

Originally posted by Fastracehorse@DRF
It sounds like you do pace #'s too - that may be a problem if you enjoy it.

Actually I do speed figures still, but I use pace as a factor.

As far as hockey, not a Canuck fan ; more of an Oiler fan to tell the truth!

And as far as Hastings goes - figure out who will get the lead on the rail and bet. Worked for me for years. Wish I could figure out some of the other tracks!

AD

[Fastracehorse]

It's actually quite an advantage to make your own figs because you have st different than everybody else.

Great game because it is rarely easy.

fffastt

[Fastracehorse]

Stampede Park in Calgary just started up.

In case you bet it through simulcast trainer Stuart Simon has 5 wins already - in 2 weekends.

Yah, I like the speed biased nature for Hastings - those cheapies can get a piece.

Well, let's go Oilers then.

fffastt

[GR1@HTR]

For the most part, I'm a pace capper...One of these days I would like to fiddle with my own pace and final time figs.

Question: What is the best book out there to teach me about creating variants and how to grow my own numbers?

BTW, the long lost RidersUp was pretty good at that. My master plan was to ask him about the above question when I was ready to move forward but it looks like I missed the boat.

[Fastracehorse]

__________I think all successful handicappers are pace junkies.

I think pace is a function of alot of handicapping styles.

Class handicappers: Those who bet droppers notice that they are often successful - because the pace is less taxing.

Those who bet risers know a sharp horse can handle the pace.

Speed figures encompass pace.

Bias believers know that no pace is going to stop their speed.

fffastt

[Tom]

Originally posted by GR1@HTR

Question: What is the best book out there to teach me about creating variants and how to grow my own numbers?

BTW, the long lost RidersUp was pretty good at that. My master plan was to ask him about the above question when I was ready to move forward but it looks like I missed the boat.

For making variants, check out Beyer's first book and the $50,000
Year at the Races book. Tom Brohamer's revised MPH is good for pace/speed figures. Quirin's second book is the one for pace-final figs. And talk to CJ - he makes good pace figs and his method is sound.
As for Ridersup, I made him an Excel spreadsheet to collect Tam data this season and we were going to look at the results after the track closed, with the eye towards figures. I am truly sorry I missed out on the chance to work with him - he was a good person.

BTW, There is a revised editon of Andy Beyer's MY $50,000 Year at the races out this spring - it is about his foray to the Canadian tracks last year. It is titled "My $26,900 Year at the Races" <G>

[cj]

Originally posted by GR1@HTR
For the most part, I'm a pace capper...One of these days I would like to fiddle with my own pace and final time figs.

Question: What is the best book out there to teach me about creating variants and how to grow my own numbers?

BTW, the long lost RidersUp was pretty good at that. My master plan was to ask him about the above question when I was ready to move forward but it looks like I missed the boat.

I think the following have helped me the most as far as pace numbers go:

Thoroughbred Handicapping: State of the Art, by William Quirin
Picking Winners, by Andy Beyer
Modern Pace Handicapping, by Tom Brohamer
randygiles.com (Old web site, not around anymore)

Quirin helped me understanding the relationship between pace and speed figures for front running type horses. Beyer taught me the concept of making variants, which I then applied to pace numbers also. Brohamer taught me about running styles (EPS), which I now use heavily in conjunction with the numbers. Giles enlightened me to what he called "multipliers", which is how I establish the track "pace pars" I use to make my numbers.

I also use one thing that may be original, I use the winner's time, not the leaders, in making pace variants. Its probably written in someone's book, I'm just not familiar with it.

I'll be happy to go into detail on any of the above, love the topic.

CJ

[MV McKee]

I think the final form your pace and final time figures take is a matter personal choice, so when I used the term "more accurate" earlier I was misstating my opinion.
I do not like to "roll up" too many factors (i.e. pace) into a single numeric expression. Whether it is because of my real world job, or simply because my mind works in a particular way, I get far more utility out of looking at pace and ultimately final ratings in an array.
I only produce figures for commonly run distances, so I deal with a limited range 5f - 7f for sprints and 1 - 1 1/8 for routes.
In a sprint I have 5 pace figures:
@1/4's - 1, 2 and 3 (#3 is 3/8's in 7f, etc.)
@1/2's - 1 and 2

In a route I have I have 7 pace figures:
@1/4's - 2, 3 and 4
@1/2's - 1, 2 and 3
@3/4's - 1 and 2

(My) Experience using numbers like this has taught me to recognize a few patterns within them, patterns that I really don't feel I could express in a single number.

In the case of younger, developing horses, I use the array as an indicator of what, over time, a horse may eventually be able to do. Generally speaking, a younger horse that is still learning his racing lessons will have a distinctly different array than his more experienced counterparts.

In the case of older horses, I am sometimes amazed at the consistency exhibited in the relationship among the various pace numbers. (Again) Generally speaking, I see the numbers improve across the array as a horse improves. But there are a couple of very strong plays that over time I have identified.

1) Is obvious, the lone speed horse. Usually what pace numbers are used for. I generally find that (easily identifiable) lone speed is somewhat overplayed.

2) The (last race) disadvantaged closer. In the case of non-pace type horses, I see a few patterns. True off the pace horses are creatures of habit. In trying to determine where an off the pace type will be in the early to mid stages of a race today, there are 3 PP (previous races) indicators.

In descending order of reliability

1-Relative Position to rest of field
2-Lengths Behind
3-Horses Actual Fraction

I have found the first two be reliable indicators of where a horse is likely to be in today’s race. The last has proven to be almost worthless. While I see horses consistently suck themselves into their usual mid/back of pack position, their customary 5-8 lengths off the pace, in doing so their early and mid pace figures jump all over the place. But many times do their final fractions markedly improve. I have seen time and time again, arrays where a late/mid running sprinter (or more commonly router) runs wildly fluctuating early fractions, but throws essentially the same final 1/4 or 1/2 each time.

These two scenarios are fairly well documented, and are visible to and accounted for by most good handicappers and their software. If these were the only common pace scenarios I wanted to express, than I would be quite satisfied with producing a single figure that incorporated both final time and pace.

There are a couple of other common scenarios that were the bane of my horseplaying existence for years that I have not been able to express in a single number. In fact these situations are the inverse of the more traditional pace effects described earlier.

1) The Speed Horse who can't rate.
2) The MidPack/Closer who can't run into a fast pace.

Regarding #1
A good friend of mine (who also posts on this board) and myself have had this discussion many times. Horses who do not gain anything appreciable by being reserved early. We joke about how differently jockeys will ride a speed horse in a 5.5f vs. a 6f race in the Bay area, or a 6f vs. a 7f race in Social. Ride the hair off the horse early in the shorter race, and then nurse him in the longer, as if the extra 330 feet have some kind of magical stamina sapping qualities that will slow him down to the 8-second final 1/16th that would result in him being caught on the wire.

EASILY identifiable lone speed is, IMO, one of the worst plays in racing. Often over bet, and with a jockey who is riding not to lose.
Common perception is if a rider opens up 6 lengths and loses, it's because he used the horse up early. If he nurses him on a 1-length lead and loses, it's because the horse is a quitter.

I don't know how I would express this type of pattern in a single number, while at the same time using pace figures to identify and rate more legitimate types. If I look at an array, I can usually identify this type relatively early on, perhaps a race or two before most of my "competition". It also helps me to realize that I need not upgrade this type of horse just because he ran a big pace figure. In that case, he merely got his perfect trip.

On the big circuits, with top-notch trainers, I have seen horses train out of this pattern over time, and gain the ability to ration their speed. But on the blue-collar tracks that constitute the majority of my play, I don't see this too often. This is a type I love to play against, as the pattern is exhibited time and time again, although I am not certain I will ever regain what I lost playing these types my first 15 years.

For a classic example of this type of horse in this type of race see
30Mar03- 4GG Greek Authority

Regarding #2

I have seen this type of pattern many times also. Some closers run better into a slow paced race then they do a fast to normal paced race. These are the true "one-run" types that need to be reserved until the final 1/4 to 3/8's to be effective. If you look at an array of pace numbers, you can usually identify these types by the fact that one of their 1/4 mile figures will be abnormally high related to their other numbers, yet neither of their 1/2 mile figures will jump accordingly.

A couple of notes regarding the above however: If I bet against these types, I always have to be aware that the jockey (or the trainer in giving instructions to the aforementioned) will actually do the right thing and let Type 1 (the unrateable speed horse) roll to a big lead, or take back Type 2 (the one-run closer) and lag him 15-20 lengths off the fast pace. But like I said, I play primarily blue-collar tracks, and it's generally safer to bet on a strategical mistake being made, potential payoffs considered. Some connections obviously make these errors more often than others.

IMO, the most important thing to do with an array of pace figures, is to relate them to final time. In essence this "smoothes" them. When I first started doing this, I left the individual pace figures as stand alones, and did not buffer them by relating them to the final time. I would constantly encounter situations like this:

Pace Final
Horse A 105 97
Horse B 101 100

I would have trouble with comparisons like this. As far as who would win the race, I could make a case either way. I just had the indication that A would lead and B would try and run him down. But very often, in cases like this, Horse B would not only win, but would engage Horse A from the bell.

While I was aware that Pace Figures alone were poor indicators of the potential outcome (finish) of a race, I also came to the realization that Pace Figures alone were of dubious value when predicting the pace scenario of a given race. To varying degrees, most every horse is running at something less than 100% speed in the earlier stages of a race. Unless a pace figure is related to the final time, and the number adjusted accordingly, IMO you are using only the X of what should be an XY function.

Ultimately, Game Theory is correct, these are "fantasy" times. I just prefer to view my fantasy(s) in multiple dimensions. It’s nice to see patterns from fantasy race fig to fantasy race fig. But I like to see the pattern within the pattern, rather than attempting to roll up too many factors into a single quantifiable rating.

I guess it's all a matter of personal preference.

[anotherdave]

Originally posted by Tom@HTR


BTW, There is a revised editon of Andy Beyer's MY $50,000 Year at the races out this spring - it is about his foray to the Canadian tracks last year. It is titled "My $26,900 Year at the Races" <G>

Hey that was a few months ago! With the rise in the Canadian dollar it is now my $34,000 year at the races.

AD

[Jed]

Originally posted by MV McKee

IMO, the most important thing to do with an array of pace figures, is to relate them to final time. In essence this "smoothes" them. When I first started doing this, I left the individual pace figures as stand alones, and did not buffer them by relating them to the final time.



IMO, that was a very nice post. Was wondering if you have had any thoughts about "pace" variants?


-Jed