Beaten Lengths

[chickenhead]

More basic questions from the chicken:

What distance does a beaten length describe?

I don't care how long a horse is, I care only what I read in the form.

Do chart callers have a fixed way of estimating, rule of thumb? Do they use the uprights, or do they guestimate free form?

I have heard 8ft and 10ft both used, that is a big difference.

Has anyone studied to see that 1 length at Hollywood is equal to 1 length at Belmont, i.e. is there any average reliable difference between the way different tracks (people) chart the race? This would be some wonderfully valuable information if it exists.

Why don't the tracks just clock every horse at every call? Not all that difficult, are they too cheap?

[rokitman]

Only the players would benefit from clocking each horse. So, why would they possibly bother to do THAT?

[GameTheory]

I'm not so sure the players would benefit. Not the smart ones anyway. Remember that we are really playing against the other handicappers, not the horses -- obfuscated data is what allows harder working handicappers to uncover value. By making it easier on the average joe we make it harder on ourselves. Imagine they just told you who the winner was gonna be before the race was run -- you couldn't ask for more accurate data than that, but would it benefit you if everyone else knew it too? (Well, they'd have to pay out on the minus pools, but you get the point.)

In response to the original question, I believe you will find this thread interesting reading:

http://www.paceadvantage.com/forum/showthread.php?s=&threadid=9753

[BillW]

Originally posted by chickenhead
More basic questions from the chicken:

What distance does a beaten length describe?

I don't care how long a horse is, I care only what I read in the form.

Do chart callers have a fixed way of estimating, rule of thumb? Do they use the uprights, or do they guestimate free form?

I have heard 8ft and 10ft both used, that is a big difference.

Has anyone studied to see that 1 length at Hollywood is equal to 1 length at Belmont, i.e. is there any average reliable difference between the way different tracks (people) chart the race? This would be some wonderfully valuable information if it exists.

Why don't the tracks just clock every horse at every call? Not all that difficult, are they too cheap?

The photo finish software I've seen is set up such that the steward scrolls through the data (picture) and clicks the mouse on the nose of each horse. This establishes the distance difference between the finishers at the time each horse crosses the line. The software has a constant built in that converts that value to beaten lengths (this isn't actually beaten lengths which is the measurement at the instant in time the winner crosses the finish line not the point in space of each entrant at the time they cross the line). The stewards weren't able to tell me what that constant was, and since they told me the chart caller didn't use that data anyway, I didn't pursue it any further (pretty long explaination to NOT naswer your question huh? possibly interesting background for some though)

I have had the theory that the chartcaller uses the uprights on the inner fence to estimate lengths. They look to me like they are spaced 10 feet apart. I have yet to confirm this (just haven't talked to the right person yet.) I have seen this substantiated once in an article posted re. Secretariat (just recently) but I have yet to talk to thet person that absolutely confirms that.

I really need to talk to one of the Equibase people as the real answer would lie there.

I'll let you know whan I find out (or someone else on the board already knows and will answer)

You are correct though, it doesn't matter how long a horse is, but only how the lengths are estimated (trust me, you cannot determine the length of a horse using the photo finish data)

BTW clocking each horse at every call results in the same problem that the measurement using the photo finish data I described above causes. It is the measurement at a unique point in space, not in time as beaten lengths is intended to be.

Another point: the company (Equibase) measuring the beaten lengths and the ((company (the track) responsible for the measureing equipment are not the associated. Equibase isn't able to install measurement equipment at the track

Bill

[BillW]

Originally posted by GameTheory
I'm not so sure the players would benefit. Not the smart ones anyway.


Agree, good example is the publishing of Beyer figs.

[JustMissed]

The charts show the lenghts that a horse is ahead of the horse just behind him at each call.

The pp's show the lengths each horse is behind the lead horse at each call except for the lead horse which shows how far ahead of the second horse.

I believe the pp's are produced from the charts and would guess the computer that produces the pp's is programmed to make the conversion from lenghts ahead of next back horse to lenghts behind the leader at each call.

Does anyone know if that's how it works?

JM

[chickenhead]

Great postings on that other thread. Here is a snippet i want to ask about:

"So how do we figure out how long a length is? Well, the way I did it is to hand-time a bunch of runner-up horses and then adjust my time calculating procedure and beaten length value until I was fairly close most of the time. You'll find there really is no good value because the horses are all slowing down at different rates. But I find something between 9 - 10 feet works best for me."

Can you expand on this? You clocked them, so you knew the time delta, then you used you velocity calc's to come up with feet, and compared feet to lengths as recorded by clocker? That would make sense to me, but I have a feeling that is not what you are saying.

It seems to me maybe the only way to really get a measurement is to have the race on DVD, stop it at the calls, get a fixed reference for width, calculate feet, and compare to the charts. Should be able to tell pretty quickly how consistent the internal calls are.

[kenwoodallpromos]

8' is usually used. That is about the standing still measurement of t the length of a horse. But a horse actually runs 11' per 1/5 second, or 55' per second.
I read in Ainslie's book about it, that 1/7 seconds is really about 1 length.
1/5 second is used because the old stopwatches were broken down in seconds and 1/5 second intervals and they still use that; it was close enough then.
A horse actually runs just about 1 length in .25 seconds. Of course, that varies depending on the size, leg action, and stride of the individual horse. My 55' per second is based on 12 seconds per furlong, 1:12 per 6f. Of course at any distance the actual time can be +- 3%.

[chickenhead]

Originally posted by kenwoodallpromos
8' is usually used. That is about the standing still measurement of t the length of a horse. But a horse actually runs 11' per 1/5 second, or 55' per second.
I read in Ainslie's book about it, that 1/7 seconds is really about 1 length.
1/5 second is used because the old stopwatches were broken down in seconds and 1/5 second intervals and they still use that; it was close enough then.
A horse actually runs just about 1 length in .25 seconds. Of course, that varies depending on the size, leg action, and stride of the individual horse. My 55' per second is based on 12 seconds per furlong, 1:12 per 6f. Of course at any distance the actual time can be +- 3%.

plenty of ways to calculate how long it takes a horse to run x amount of feet, I'm looking for how long is one length as recorded in the charts. I guess clocking them yourselves, or measuring on video are the only two ways to find out. And I'm guessing I will find the answer is somewhere between 8 and 10 feet. Seems like well worth the effort just the same.

If you saw what I was getting at in the original question, is how much average variance there is from chart caller to chart caller. That could be a sweet spot that people are not taking into account. A small one, but it all adds up.

[rokitman]

If the fence posts are a length, which they appear to be, then a conscientious chart caller can do a fine job making an accurate chart off them. But the "conscientious" part is the big "if." The beaten lengths on the early pace calls are often ridiculously wrong. That's probably evidence that they are using the fence posts which are not as clearly visible as they are at the finish. Well, they are plenty visible but they are not RIGHT IN YOUR FACE, so, apparently, that's reason enough to do a sloppy job. Get the humans off the job and get it accurate. Consistent, at least. As for the tiny minority who may be profiting from inaccuracies in the beaten lengths- tough luck.

With all the video available now available a nice business opportunity has presented itself here. A beaten-lengths service. There ya go, Chickenhead. A nice, old-fashioned get-rich-slow scheme. But, then again, maybe not so slow.

[BeatTheChalk]

Darn you folks anyway ! I didnt have enough to do ...now I
have to call my smart friends who took calculus and physics ..
while I wasted my youth and Basketball Court and Football
field. ( I shudda played Golf ) If anyone can do all of the math
to get the "" real beaten lenghts"" ...more power to em. Truly
a Business Opp for the right person. :):p :D :(

[kenwoodallpromos]

Go to Cecil Seamen and Co. online- they have a great chart of thoroughbred lengths- up to about 96" or 8'; od course, that is from chest to rump- not counting "by a nose, head, or neck". You should email Equibase or DRF and ask how their chart callers estimate lengths back.

[chickenhead]

I sent an email to Equibase already, I'll post if i hear anything back.

[GameTheory]

You have to understand that beaten lengths don't really make sense until you look at them as units of time rather than length. Beaten lengths as they appear in the charts/pps are NOT a "snapshot" of the race at the finish -- they represent of the amount of space/time as the horses pass the finish line (imagine you can only see the finish line itself -- Horse 1 crosses the line, and 1/2 second later Horse 2 hits it -- that 1/2 second is now represented as "lengths" on the chart).

The hand-timing I referred to in the earlier thread was between the winner and the runners-up to cross the line. I would determine the actual amount of time that needed to be added to the winner's time to get the correct times for the rest of the horses. Then I used that information to calibrate my method of calculating those times based on the beaten lengths as published in the chart/pps so that my results matched reality as close as possible.

[JustMissed]

What about this:

Let's say the A horse crosses the wire 12 lenghts ahead of the B,C,D horses.

Let's say the D horse is a very fast closer and passes the B & C horses after the A horse passes the wire.

What will the pp's say?

JM

[JackS]

I'm With GT. Perspectives from a chart writer can also vary. A length would be a lot more accurate if our and the chart maker were viewing each points of call from a perfect lateral posisition. Total accuracy is impossible because of the differing opinions that individual chart makers may have. When the Racing Times were in print, the slight difference in the DRF charts were obvious. I would never say that one was better than the other, just different. Technology has advanced enough that lengths or parts of lengths could be more accuratly described in 100s of a second versus noses, necks, 1/4s or 1/2s. I personally hope it never comes to this. I will continue to adjust the PPs as printed converting to beaten times. More important than how long a horse may be is how accurate is the timing equipment? I think much of this stuff is not always completely accurate. But then again we have the varient which should put us back in the ball park when estimating times. Handicapping perfection is making a profit and not counting races won or lost. Someone here has already pointed out that regardless of the information that is presented, we will all be working with the samething. The difference- Winners will be using that info a little different or maybe a lot different than the general public.

[BillW]

Originally posted by GameTheory
Beaten lengths as they appear in the charts/pps are NOT a "snapshot" of the race at the finish -- they represent of the amount of space/time as the horses pass the finish line (imagine you can only see the finish line itself -- Horse 1 crosses the line, and 1/2 second later Horse 2 hits it -- that 1/2 second is now represented as "lengths" on the chart).


How does Equibase measure this?

Bill

[Zaf]

When's this GPS stuff coming ? Is this a good or bad thing ?

ZAFONIC

[JackS]

THinking about the 8' horse versus the 10' horse. Would this mean tha the 8 footer had to run 20% faster the the 10 footer just to keep up? Also ,is it possible that a 8 foot horse had legs that were longer the the 10 footer? Maybe the smaller horses stride made him the equivent of 10'6". Questions in futility and I urge everyone to ignore this post.

[chickenhead]

Originally posted by JustMissed
What about this:

Let's say the A horse crosses the wire 12 lenghts ahead of the B,C,D horses.

Let's say the D horse is a very fast closer and passes the B & C horses after the A horse passes the wire.

What will the pp's say?

JM

I'm driving home tonight, and the significance of your posts hits me like a ton of bricks.

YOU CAN'T USE PAST PERFORMANCES, YOU HAVE TO USE CHARTS!

I can't beleive I didn't realize this before, but even if you knew exactly how long long a mythical length was, and the chart callers were perfect, and if they reported it perfectly in the Form, you still would not get the right times, final or fractional, and your velocity ratings would not be right, except in a few cases.

I wish someone had told me this before! And the beauty is they give the charts away for free!

I still need to know how long a length is, but I feel refreshed by this new understanding. I'll get different, better numbers by using the charts and making my own PPs.

Thanks all!

[Tuffmug]

Chickenhead,
U]PP lines are made from the charts.[/U] Simple addition of beaten lengths of each horse who placed in front of the horse of interest.[

[Cratos]

Chickenhead,

From the Glossary on the Daily Racing Form Website a length is defined in the following definition:

“LENGTH- Length of a horse from nose to tail, about 8 feet. Also distance between horses in a race.”

Also please be aware that the past performances lengths comes from the charts, but the distances of the head, nose, and neck are not taken into consideration when the chart data is converted to past performances.

For example if in the charts horse A is 5 lengths in front of horse B and horse B is a neck in front of horse C, the DRF past performances will have both horse B and horse C 5 lengths behind horse A.

[JackS]

Yes I agree. There's no such thing as a length and a nose/hd/nk or even a length and an 1/8th. If a length is described as 8', an 1/8th would be a foot and anything less than 1/4th is not reported. So a length as it appears in the DRF is somewhere between 8' and 10'. For practical purposes, I've always used 10 feet converted to time. I Have experimented with using lesser times when converting earlier fractions as applied to lengths behind. Your call.

[GameTheory]

The pps are simply the lengths from the charts added up, as has been stated. What I was getting at with the "snapshot" stuff is the same thing that JM was -- beaten lengths are about time, not distance. If the winner wins by 10 lengths, that means just means it was 10 lengths = X seconds (however you want to do it) before the next horse crossed the line. It does NOT mean that the second horse was in second at the time the winner crossed the line -- there is no source for that information except the race replay itself. Now at the fractional calls, where all this is done by eyeball with no video, maybe the caller does "scan backward" and doesn't actually watch each horse cross the 1/4 pole or whatever he is looking at. So the earlier calls are really inaccurate.

Bottom line: much of what is in the charts is WILD GUESSES! The only thing you can count on (most of the time) are the horses names and final finishing positions.

[chickenhead]

If a horse hits the second call 5 lengths behind the horse in front of him, but was 10 lengths behind at the first call, he is gaining on the horse in front of him, yes? He gained ground on the horse in front of him during the 5 lengths he ran to reach that second call as well, yes?

Now, if you are using PP's, and we're talking about the lead and second place horse, you're fine. But what if you horse is in third? What if he is running slightly slower than the second place horse? Say he was 1 length behind the 2nd horse at the first call, now he is 2 back.

His PP line would be
3by11 3by7 He picked up 4 lengths, right?

Wrong, he actually picked up more than 4 lengths, because the 2 was gaining on the 1 for those 5 lengths. The further back your horse is at any point in the race, or maybe more precisely, the more spread out the race is in front of your horse, the more exaggerated this will be. Without the chart you don't know this. You have to calculate the velocity for all horses in front of your horse at any point in the race to get it.

Do you understand why I say there is a difference which numbers you use? You don't get the same numbers with the PP as you do with the chart. I understand none of these numbers are perfect, but doesn't it makes sense to be as precise as possible?

[BillW]

If the 3rd horse lost a length to the second, his total gain on the lead should be one less than that of the second. It works out OK.


Bill

[NoDayJob]

If you use 10 feet per length then there are 66 lengths per furlong.
If you use 9 feet per length then there are 73-1/3 lengths per furlong.
If you use 8 feet per length then there are 82.5 lengths per furlong. Let's say the winner finishes 6 furlongs in 70 seconds (1:10)
6 * 66 / 70 = 5.6571428 lengths per/sec.
6 * 73.34/ 70 = 6.2862857 lengths per/sec.
6 * 82.5 / 70 = 7.0714285 lengths per/sec.
The old rule of thumb was 6f. in 1:12 = approximately 5 lengths per/sec.
Not likely a 1:10 horse is traveling 7 lengths per/sec. but, I could be wrong.
Let's say your horse finishes 5.75 lengths behind the winner.
((66 * 6) - 5.75) = 390.25 / 70 = 5.57 lengths per/sec.
If you want a speed/velocity number just subtract 4.5 from 5.57 and multiply by 100.
Your speed/velocity number is 107.
The winner's speed/velocity number is 116.

NDJ

[JackS]

Chicken- I understand what your saying and what you describe is "trip Handicapping". The synopsis below the charts will give you a better picture of what actually happened during a race rather than just the posisition and lengths behind that are the PP's. Not widely used in North America but popular in Europe.

[GameTheory]

Originally posted by chickenhead
If a horse hits the second call 5 lengths behind the horse in front of him, but was 10 lengths behind at the first call, he is gaining on the horse in front of him, yes? He gained ground on the horse in front of him during the 5 lengths he ran to reach that second call as well, yes?

Now, if you are using PP's, and we're talking about the lead and second place horse, you're fine. But what if you horse is in third? What if he is running slightly slower than the second place horse? Say he was 1 length behind the 2nd horse at the first call, now he is 2 back.

His PP line would be
3by11 3by7 He picked up 4 lengths, right?

Wrong, he actually picked up more than 4 lengths, because the 2 was gaining on the 1 for those 5 lengths. The further back your horse is at any point in the race, or maybe more precisely, the more spread out the race is in front of your horse, the more exaggerated this will be. Without the chart you don't know this. You have to calculate the velocity for all horses in front of your horse at any point in the race to get it.

Do you understand why I say there is a difference which numbers you use? You don't get the same numbers with the PP as you do with the chart. I understand none of these numbers are perfect, but doesn't it makes sense to be as precise as possible?

Once again, think time, not distance. Lengths are the TIME it takes for the horse to reach the same point as the horse in front of him. There is no "gaining during those 5 lengths" because those five lengths don't really exist as distance. Imagine watching the point on the track that represents the second call through a pinhole where you can't see anything but that one point (which is exactly how the photo finish camera works, by the way -- it only photographs a space a couple of inches wide). You see the leader pass. 2 seconds later you see another horse pass that same piece of ground. Those 2 seconds will be represented in the chart as something around 10 lengths. That is the only sane way to look at it...

[JackS]

No Day Job- Thanks for the math. Using 10' for one length and the standard 1/5sec divided into 66 lengths would equal 13seconds. This would equal 26 seconds to the first call. If par is 22 seconds, this horse covered more ground than the standard 50ft per second using the 1/5th second per length. The 4 second difference in this example would equal about a 16% error using the common 1/5th to make ones numbers. Converted to 10ths then 100th's the more realistic time goes from 0.20 to 0.17 rounded. This is not a part of my handicapping but conversion to time is. My conversions are much simpler and much faster and in no way accurate. My method attempts to distinguish differences between horses times rather than any true number.

[JustMissed]

chickenhead, Glad my post was usefull to you.

Thanks to you, GT and others for responding. This has turned into a very good thread.

One other thing to think about when looking at pp's and not the chart.

Say your horse is third 5 lengths behind the leader with the second horse 3 lengths behind the leader. Say the lead horse pulls up or goes wide and drops to 4th position just prior to the next split.

If your horse and the second place horse held their positions, the pp's would lead you to conclude your horse made a 3 length move going from 5 lengths behind to 2 lengths behind. How about that?

If all else fails resort to the S.W.A.G. method of handicapping.
Scientific Wild Ass Guess.

Good luck,

JM

[Cratos]

This is all being made too complicated. The only time that is exact is the time for the horse that breaks the timing beam at the start, at each point of call, and at the finish. All other times are just estimates.

Therefore by example, I use the following method of "beaten lengths" to estimate the time of a horse who is behind the leader at a point of call or that finishes behind the winner.

EXAMPLE:

Horse B runs a 6f in 1:10

Horse W finishes 6 lengths behind Horse B

Step 1: Calculate Horse B's average feet per one-fifth second

as 3960/350 = 11.31429 feet per one-fifth second.

Step 2: Calculate Horse W's estimated distance behind Horse B in feet using the DRF metric of 8 feet per length as 8*6 = 48 feet.

Step 3: Convert Horse W's distance to estimated time by dividing (48 feet by 11.31429) times one-fifth second equals .85 seconds.

Step 4 Add Horse W's estimated time to Horse B's actual final time and it is 1:10.85 seconds.

This method can be applied at each point of call also it can be easily put into a spreadsheet and automated for any number of beaten lengths.

[chickenhead]

Originally posted by GameTheory
Once again, think time, not distance. Lengths are the TIME it takes for the horse to reach the same point as the horse in front of him. There is no "gaining during those 5 lengths" because those five lengths don't really exist as distance. Imagine watching the point on the track that represents the second call through a pinhole where you can't see anything but that one point (which is exactly how the photo finish camera works, by the way -- it only photographs a space a couple of inches wide). You see the leader pass. 2 seconds later you see another horse pass that same piece of ground. Those 2 seconds will be represented in the chart as something around 10 lengths. That is the only sane way to look at it...

I may be wrong, but at least I've got everyone talking :)

I understand that you can look at things in the time domain, what I don't know is whether the chart caller does. Unless I am wrong, the chart caller looks at them the way I described above, like they appear in the chart, horse A X lengths in front of next when A hits the call, horse B X lengths in front of next when he hits the call, etc. etc. If this is in fact how the chart caller makes the call, I can't for the life of me see how I'm wrong. Someone with definitive knowledge PLEASE say something if you know...

This all goes back to my original question, how does the chart caller make his chart, and what is a length?

[Tom]

Cratos,
Aren't you using B's feet per fifth as the velocity that W is rrunning?
I would assume the tow horses are traveling at different speeds.
Wouldn't you have to calculate W's time by using 3960-48 = 3912 feet in 350 fifths?
This would have him going at 11.177 feet per fifth. So the value of a beaten length, feet, at 6 furlongs in 1.10 is 11 feet? So if 8 feet = a length, the value of a BL is 1.25 points?

How about first call. 1320 ft in 22 flat. 1320/110=12, so 12 feet per fifth. So if a length is 8 feet, then the value of a BL at 2 furlongs here is 1.5 points?

[GameTheory]

Originally posted by chickenhead
I may be wrong, but at least I've got everyone talking :)

I understand that you can look at things in the time domain, what I don't know is whether the chart caller does. Unless I am wrong, the chart caller looks at them the way I described above, like they appear in the chart, horse A X lengths in front of next when A hits the call, horse B X lengths in front of next when he hits the call, etc. etc. If this is in fact how the chart caller makes the call, I can't for the life of me see how I'm wrong. Someone with definitive knowledge PLEASE say something if you know...

This all goes back to my original question, how does the chart caller make his chart, and what is a length?

You have it correct as far as that goes. But I don't know how you come up with (in your original example) that the 3rd horse makes up more than 4 lengths on the leader from the first to second call. I say it only makes sense to deal with lengths as time and not distance because we really have no information about the positions of the horses relative to each other at any single given instant -- when horse A is X lengths ahead of the next horse at a call, we don't know who that next horse is (it may or may not be the next horse that hits the call). And when the next horse does hit the call, we have no idea of the current position of the first horse that has already passed (he could have dropped dead 3 inches past the call and it wouldn't affect the beaten lengths of the horses behind him for that call).

Now of course you do need a method to convert lengths to time as Cratos has shown above, but once you've done that, forget about distance...

[rokitman]

I think all the responses here that don't specifically answer your question ultimately DO answer your question in a back-handed way.

"How does a chartcaller make a chart?" is a scientific question about something that is not scientific at all. It's an "art." Or more honestly, a sloppy pile of pooh compiled using video by counting posts on the fence when they can see them and throwing out a wild guess when they can't. The "what is a length" question serves perfectly to second the unscientificness of it all.

Why humans are still involved in this tedious job in 2004 is beyond me. I've worked in automation. I estimate about 2 days per track to set this up so humans are only needed for messing up the trouble lines.

[chickenhead]

the horse does make up more than 4 lengths, not from the chart callers perspective, but in real terms of feet behind the leader, he does.

You are right, it is not a snapshot, he was not 7 lengths behind the leader necessarily when the leader past the second call, he might have been, he might not of been. If you model each horses velocity, you can make an extrapolation as to how many feet they lost or gained on the horse in front of them AFTER THE HORSE IN FRONT OF THEM HIT THE CALL. You can then reference everything back to the leader, and try to actually have a snapshot that would be complete as of the time the last horse passes the call, looking forward.

It would be an extrapolation (interpolated as well), but it just might be more accurate as to reality.

I apologize if no one understands what the hell I'm jabbering about. I'm not sure I do.

[Cratos]

Hi Tom,

If you convert 1:10.85 into fifths of a second and divide the conversion into 3960 feet (the distance of 6f) you will get 11.17855 feet per one-fifth second.

[chickenhead]

oh yes, heard back from Equibase, pasted below:

"This information is compiled by the individual tracks and then sent to us. For information on how a specific track charts the information, you'll need to contact that track directly."

I don't have a lot of faith in their email jockey anyway....I also sent one to the VP of Engineering, I'll let you know if he answers (not likely I would guess)

[BillW]

Originally posted by chickenhead
oh yes, heard back from Equibase, pasted below:

"This information is compiled by the individual tracks and then sent to us. For information on how a specific track charts the information, you'll need to contact that track directly."

I don't have a lot of faith in their email jockey anyway....I also sent one to the VP of Engineering, I'll let you know if he answers (not likely I would guess)

hmmm, thats funny. HOU has an Equibase employed chart caller on premises. I guess the "Answer those pesky customer e-mails" dept. at Equbase isn't aware of that. :rolleyes:

Bill

[kenwoodallpromos]

Multiply beaten lengths in sprints and fast routes by .14; in slower routes and turf races by .15. The answer is in seconds behind.

[kenwoodallpromos]

Ainslie's 1/7 second per length was right- he was talking about he was actually talking about length of track covered by the average racehorse- about 7 "lengths" or 7 x 8'=54', close to the 55' per second at 12 seconds per furlong.
The 8' length must have been an estimate of standing chest to rump measurement, used because of stopwatch 1/5 second fractions and because when building the old wood track railings, most wood was (and still is) measured in 8' lengths- 2x4's, plywood, etc.; so posts were at 8' intervals and it all worked out close enough.
To convert pp or chart lengths to seconds, use each 8' length times .14 for dirt sprints and fast dirt routes (12 seconds per furlong); times .15 for slow dirt routes and turf races (13 seconds per furlong). If you want more adjustment, you can adjust 1% for each second faster or slower than 12 and 13 seconnds per furlong.

[chickenhead]

There is no right answer, they are all approximations.

Say we are all watching a mythical race.

Horse A passes the second call and is clocked at 44.00

Horse B is in second place, and is 5 lengths behind horse A as A crosses the call.

Horse C is in third place when horse B crosses the call, 2 lengths behind horse B.

Now, if I asked you to tell me the time each of them ran the 1/2 in, you can all come up with good, close answers..

But I maintain, you will come up with better numbers if you calculate horse B's time using his expected velocity at that point in the race (calculated from his 1st call velocity, the lead horses 1st and 2nd call velocity, and the number of lengths gained or lost by the B horse between the 1st and 2nd call), use that velocity to figure how fast he likely ran those 5 lengths to the call, then use that new time to calculate for C, how long did it take him to close 2 lengths to the call at his expected velocity. This is absolutely different than using "lengths behind the leader" because that does not exist, but that is what everyone bases their numbers on.

The main leap of faith or assumption is that the horse that hits the call in 2nd say, the B horse in this case, is the same horse that was 5 lengths back when A crossed the line. In most cases this will be true, and I think it is less of an assumption than people use now.

This is a bit of a paradigm shift, but mathematically it is correct, or at least more correct than using some fixed constant. It requires some more complexity, because you can't use average velocity, you have to figure what the likely velocity is for any given horse at each call.

Sounds like fun, eh? The numbers don't have to be right, they just have to be more right than the next guys...

[GameTheory]

At the finish, the lengths are calculated from film, and they convert times to lengths via some rule like 1/5 sec = 1 length. At least that is how it used to be done 10 years ago or so, maybe things have changed. In which case calculating velocity, or using 1/6 sec for a length or anything else will be off. However, because each horse is rounded off, any horse more than a few lengths behind will still be wrong if you convert back to time using 1/5. Tricky business...

[JackS]

Horse A wires the field 1/1 ,1/1 ,1/1. Horse B in the same race runs last 12/12, 12/12,12/12. Didn't both of these horse cover the same amount of ground in the exact same time? It would appear the horse B lost the race at the gate. Had both horses broke on top in a dead heat, then the win would also appear to be a dead heat. Had horse B gained one length i.e 12/12, 12/11,12/11 now it appears that except for the bad break, horse B actually covered the same amount of ground faster than horse A These situations occur at multiple points with usually many horses. Races within races. In the above example, horse A wins because perhaps he's a professional speedball or perhaps due to some racing luck, he happens to get the perfect break.

[Big Bill]

NoDayJob,

The example in your post to this thread stated:

"If you want a speed/velocity number just subtract 4.5 from 5.57 and multiply by 100."

What does the number, 4.5, represent?

Big Bill

[chickenhead]

Originally posted by JackS
Horse A wires the field 1/1 ,1/1 ,1/1. Horse B in the same race runs last 12/12, 12/12,12/12. Didn't both of these horse cover the same amount of ground in the exact same time?

No, you have no reason to think so, because you don't really have any reason to think that horse B was ever 12 lengths (as measured in distance) behind Horse A. What you do have is that the distance between 1-2, 2-3, 3-4, 4-5, 5-6, 7-8, 9-10, 10-11, and 11-12 place horse added up to 12 lengths cumulative at the time each of them (except for the 12) passed through each call. It's different. That is, if the chart caller actually makes the chart the way they are supposed to.

[Tom]

And all this is why it is so omportant to measure times in hundreths of a second. :rolleyes:

[cj]

Originally posted by JackS
Horse A wires the field 1/1 ,1/1 ,1/1. Horse B in the same race runs last 12/12, 12/12,12/12. Didn't both of these horse cover the same amount of ground in the exact same time? It would appear the horse B lost the race at the gate. Had both horses broke on top in a dead heat, then the win would also appear to be a dead heat. Had horse B gained one length i.e 12/12, 12/11,12/11 now it appears that except for the bad break, horse B actually covered the same amount of ground faster than horse A These situations occur at multiple points with usually many horses. Races within races. In the above example, horse A wins because perhaps he's a professional speedball or perhaps due to some racing luck, he happens to get the perfect break.

You are off base here. They may have broken right next to each other, but from the start to the first call, the first horse outran the second by 12 lengths worth of time. The rest of the way they ran the same time, assuming the BL measurements are correct. They ran the first part of the race much differently, then the last two calls in the same time.

[JackS]

Chicken- I think your refering to the charts, I'm referring to the PP's. CJ- I think the first call is a better example as you stated. From the gate would represent all beaten lengths fron 0-12.

[chickenhead]

Originally posted by JackS
Chicken- I think your refering to the charts, I'm referring to the PP's. CJ- I think the first call is a better example as you stated. From the gate would represent all beaten lengths fron 0-12.

He could have run 10 seconds slower from the first call to the finish. You have no way of knowing just by looking at either the PP's or the charts at face value (even if beaten lengths are correct and precise). There is a nuance that everyone misses.

[chickenhead]

I shouldn't have said everyone, Game Theory obviously doesn't, I'm sure lots of others on this board don't either.

[cj]

I honestly have no idea what you are talking about, maybe when I wake up tomorrow I'll see it clearer. Have a good night.

[chickenhead]

I hope somebody does, I feel like I'm tilting at windmills.....

[NoDayJob]

Originally posted by Big Bill
NoDayJob,

The example in your post to this thread stated:

"If you want a speed/velocity number just subtract 4.5 from 5.57 and multiply by 100."

What does the number, 4.5, represent?

Big Bill

You could sinply multiply 5.57 times 100 to get a whole number. By subtracting 4.5 (approximate lengths per second) from 5.57 LPS you have a number similar to the speed ratings currently used. It's your choice, however.

NDJ

[Turf2Dirt]

This whole thread is a big advertisement for racereplays.com:D

Don't use numbers, don't use charts. just watch the race for yourself