B/Ls mutiplier

[raybo]

I would like to get an idea of what the handicappers here use when adjusting times by beaten lengths. .2 seconds per length has been the standard for years but I've heard there are other multipliers that are more accurate.

[Johno]

None of the above is correct. It depends on how fast the horse and race is running that determines how much time equals a length.

Johno

[robert99]

If the course uses the photofinish camera to measure time-lengths it is common that each 0.20 seconds late in finishing (passing the winning post) equals one time-length. There is no standard definition of a length and it is better to keep to using time ie horse is beaten beaten by exactly 2.40 seconds, say, (12.0 time lengths) than any other measure using physical guess at "lengths" in distance terms.

[raybo]

None of the above is correct. It depends on how fast the horse and race is running that determines how much time equals a length.

Johno

Ah-Haaaa! I was wondering if someone was doing what I have been thinking about for quite sometime. Do you adjust by the leader's pace at each call (using distance traveled by the leader versus distance traveled by each horse in the same amount of time)?

One can figure the leader's average speed per fraction easy enough. Based on the distance traveled by the leader vs. the distance traveled by each horse (using 8', 9', 10' per length, shouldn't matter as long as you apply the same distance per length to all horses, you can compare distances traveled by all horses and therefore the average time per length).

It's hard to explain my thinking but I've come up with a formula that varies according to the fraction leader's average pace rather than using a static time per length regardless of the pace or distance traveled in any particular fraction, and the race distance itself.

[raybo]

If the course uses the photofinish camera to measure time-lengths it is common that each 0.20 seconds late in finishing (passing the winning post) equals one time-length. There is no standard definition of a length and it is better to keep to using time ie horse is beaten beaten by exactly 2.40 seconds, say, (12.0 time lengths) than any other measure using physical guess at "lengths" in distance terms.

This would be great if all tracks actually timed each horse at each fraction and then converted the individual timings into a standard beaten length (whether it be 8', 9' 10' or whatever, shouldn't matter as long as we know what the length measurement is).

[cj]

Ah-Haaaa! I was wondering if someone was doing what I have been thinking about for quite sometime. Do you adjust by the leader's pace at each call (using distance traveled by the leader versus distance traveled by each horse in the same amount of time)?



I do something similar, but I use how fast the horse itself is traveling (best estimate of course) at each call rather than the leader. Of course you have to know the leader's fraction to get this info.

[Jeff P]

Speed and pace figs are problematic. We've talked about the effect of track speed, bias, wind, weight, trips, ground loss, the effect of pace on final time, riders rating horses differently in the early stages of races depending on distance and perceived track bias, and more.

Despite that, speed and pace figs, or incremental velocity, or beaten lengths and time - no matter what you call it... it's THE starting point for the vast majority of players. I'll offer up a few random thoughts on figure making.



I think one of the first things a would be figure maker needs to do is ascertain is the length of a horse. After all, if we need to use X seconds per length when making figures - don't we also need to know how long a length is?

Is the widely held 10 ft per length accurate? FWIW, I've seen different authors use different values for this. In Modern Pace Handicapping Brohammer used 10 feet per length. Nick Borg used 9 ft. And Charles Carroll used 8 ft. If you did your own sample - if you busted out a tape measure and start recording actual measurements of say 1000 horses; how would that relate to the typical beaten lengths distance used by a chartcaller at a NYRA track? And would beaten lengths for the same horses in the same race be different if the NYRA caller called in sick and a chartcaller from KY or FL sat in that day?




The second thing you need to know as a figure maker is that X seconds per length can vary based on distance and class of horse.

I recently started a project where I began making my own velocity figures. Whether or not that project pans out and those figures make their way into JCapper is still yet to be determined. But after coding out logic to record the fractional times for every race winner in every race run in North America over the past two years a number of things became apparent. Aside from discovering significant differences (track pars) for the various track layouts at identical distances - I was able to reinforce my own undertsanding of the basics behind figure making.

Final fraction velocity in fps for horses in routes is much slower than final fraction velocity for horses in sprints. So a relationship exists for F3 time in fps based on distance travelled. The same type of relationship exists among the horses of different ages and class levels. For example, at identical distances 3 yr old maiden claiming fillies run much slower F3 numbers in fps than do 5 yr old male stakes horses. That's not really a revelation and should come as no surprise to anybody.

But it suggests to me there might be a better way than universally applying X seconds per length in all situations for the purpose of figure making.



-jp

.

[robert99]

This would be great if all tracks actually timed each horse at each fraction and then converted the individual timings into a standard beaten length (whether it be 8', 9' 10' or whatever, shouldn't matter as long as we know what the length measurement is).

A few use Trakus and In UK we used to have Turftrax but professionally you have to get all "reliable" sectional timing and any other timing data from video software. If you are interested in time and speed why do you ever need to go into "lengths" then back into time. Forget "lengths" completely - they are historical artifices, only necessary before the days of stopwatches and video etc.

[raybo]

I do something similar, but I use how fast the horse itself is traveling (best estimate of course) at each call rather than the leader. Of course you have to know the leader's fraction to get this info.

That's what I'm doing.

[raybo]

I agree with all of your post. Using a static time per length for all races and all tracks is, IMO, inaccurate at best. We can't change the way individual timers or track timing equipment produce their beaten lengths numbers so there will always be inaccuracies/inadequacies involved. However, using .2 seconds per length universally seems to be, simply, wrong.

[raybo]

A few use Trakus and In UK we used to have Turftrax but professionally you have to get all "reliable" sectional timing and any other timing data from video software. If you are interested in time and speed why do you ever need to go into "lengths" then back into time. Forget "lengths" completely - they are historical artifices, only necessary before the days of stopwatches and video etc.

Using lengths, in my case, is required by the handicappers I am designing spreadsheets for.

As for my own personal requirements, I have no need for lengths other than the necessity of using them to adjust the leader's fractional times to obtain individual horses' fractional times.

[raybo]

Here's the formula I'm playing with to vary the time per length based on the fractional leader's time:

"L1Fr" = Leader's 1st call fraction. 9' is the length of a horse, but could be 8' or 10' or 11', doesn't matter too much.

"LTPL1Fr" = Leader's Time per length, in seconds, for 1st fraction.

"BLs" = Horse's beaten lgths end of 1st fraction.

"D1fr" = distance of 1st fraction in feet.

"LL1FR" = Leader's lengths run during 1st fraction time.

"HL1Fr" = Horse's lengths run during 1st fraction time.

"Adj1Fr" = Horse's Adjusted 1st call fraction


LL1Fr = D1FR/9
LTPL1Fr = L1Fr/LL1Fr

HL1Fr = LL1Fr-(BLs*9)

Adj1Fr = HL1Fr*LTPL1Fr

[bobphilo]

None of the above is correct. It depends on how fast the horse and race is running that determines how much time equals a length.

Johno

That's why I chose "Other" as the best answer. The time it takes a horse to cover a length depends entirely on how fast he's moving.
Given that the average horse is about 8 feet long, the traditional .2 answer is the least likely. Horses just don't run that slow unless, as Charles Carroll says in Handicapping Speed, you're handicapping harness horses or Clydesdales. Clydesdales are about 10 feet per length so the tradirional measures of 10 feet per length and .2 lengths per second would be more accurate for them.

Come to think of it, the traditional measures might work for t-breds too since the 2 errors in measurement might cancel each other out. I doubt that it would be as accurate as using the accurate measures of a length and how fast the horse is actually moving.

Bob

[kenwoodallpromos]

Ideally I would like to have all horses timed at each 1/4 pole and along with the final times that are actually done now, give only me the information.

[raybo]

Ideally I would like to have all horses timed at each 1/4 pole and along with the final times that are actually done now, give only me the information.

Yeah,that would solve a lot of things, for you, anyway. :)

[JPinMaryland]

For horses moving 24 1/2 sec per quarter, use 0.14 sec per length

For those moving 25 1/2 sec per quarter, use 0.15 sec per length. You could extrapolate these two numbers to arrive at other speeds.

THis is based on 8 ft per horse length which we've had another confirmation of this from someone else in another discussion here.

[ranchwest]

This would be great if all tracks actually timed each horse at each fraction and then converted the individual timings into a standard beaten length (whether it be 8', 9' 10' or whatever, shouldn't matter as long as we know what the length measurement is).

While the value assigned to a length affects computations for trailers equally, it does affect the comparison to the leader differently. I'm not trying to argue which is best, only that they ARE different. I use 8 feet. Then, I don't try to estimate the time to the fractional distance, I compute the speed to the distance travelled.

[raybo]

While the value assigned to a length affects computations for trailers equally, it does affect the comparison to the leader differently. I'm not trying to argue which is best, only that they ARE different. I use 8 feet. Then, I don't try to estimate the time to the fractional distance, I compute the speed to the distance travelled.

I split the difference between the shorter horse (8') and the longer horse (10') and come up with 9' as the average ft per lgth..

[ranchwest]

I split the difference between the shorter horse (8') and the longer horse (10') and come up with 9' as the average ft per lgth..

I don't think actual length matters. All that matters is what works for you.

Really, I think having confidence in the number is significant. If you believe in it, that's a big help in itself.

[Robert Goren]

The finish time is a function of the photo finish camera. The picture you see is not a snapshot of the field when the winner crosses the the finish lines. If know the speed of the camera then you know time of the horse. The fractional times are a completely different story. There you have to rely on the chart maker's call. It has been my experience when comparing charts to replays that you are lucky if they get the position of a horse right let alone the lengths back. JMO

[Cratos]

I would like to get an idea of what the handicappers here use when adjusting times by beaten lengths. .2 seconds per length has been the standard for years but I've heard there are other multipliers that are more accurate.

The beaten length is based on the measurement of the chart caller or the person who constructs the charts.

To that end I use 0.165893 seconds per length and I came up with that number from the following calculation:

• The DRF states that they use 9 feet as the measurement of a length

• The average horse speed is 36.00 MPH or 10.85040/.2 second

Therefore taking the DRF’s 9 feet/10.85040 you will get 0.829463 and multiply that by .2 seconds the resultant will be the 0.165893 seconds per length.

However this is a general static measurement and can be made dynamic by the recalculation using the final time or internal fractions of the race(s) under analysis.

For instance, if a race was run at 1 mile in a final time of 1:33 the feet per one-fifth second would be 11.35484 which translates into a beaten length conversion of 0.158523 seconds or less than an inch difference between using 0.165893 as the time increment per beaten length.

[fmolf]

Yeah,that would solve a lot of things, for you, anyway. :)
unless trakus is in use you will always be at the mercy of the accuracy of the chartcallers estimation!

[HUSKER55]

when a horse runs they "stretch out" and some horses don't "stretch out" as far as others do.

that difference would have a huge impact on feet per length, I would think. I am assuming the horse does not "2 step".

I have always wondered just how accurate any figure is for beaten lengths.

It seems that there is a lot of "average figs" being used.

[ranchwest]

Only LA breds can 2 Step. :lol:

Because I know that beaten lengths are not recorded precisely, I try to not get too caught up in the exactitude of speed numbers. It's ball park numbers and I try to treat them that way.

[Bobby Seller]

8'

[raybo]

I started this thread as a test, to see what the consensus for this forum is.

While not many here voted, from the results of those who did it is clear that the age-old 1/5 second per length is still most popular.

The 2nd most popular category, "Other", shows that many questioned that age-old standard and went in a different direction.

At the time the thread was started, I already knew that any "static" multiplier, no matter what that value is, is simply, inaccurate.

A "dynamic",pace variable, multiplier, based on how fast the horses are traveling, at the internal fractions, IMO, is far more accurate.

With present technology being what it is, for the number of beaten lengths, the best we can do is to use the internal fraction leader's time, as the key, to becoming most accurate, in adjusting individual horses' internal and final B/Ls.

The math is basic and the results, surprising for many, I'm sure, are quite different. All of a sudden one realizes, or should realize, that the previously popular way of adjusting times, "static" times per beaten length, was/is wrong.

[Tom]

The problem is, Raybo, it appears the people betting the horses put far more thought, work, and technology into the races that those who run the show.
This game, overall, is a farce run by people who have no clue what horse racing is. Anyone who believes 100% anything they see in the Form is a fool.
At best, the PPs are rumors of what happened.:D

[raybo]

The problem is, Raybo, it appears the people betting the horses put far more thought, work, and technology into the races that those who run the show.
This game, overall, is a farce run by people who have no clue what horse racing is. Anyone who believes 100% anything they see in the Form is a fool.
At best, the PPs are rumors of what happened.:D

Agreed!

However, all of us are faced with the same "inaccuracies", and all of us are forced to, either, deal with them in the most efficient manner possible, or totally ignore the subject of beaten lengths. I'm not one to "ignore" how badly a horse was beaten by others in a race. Therefore, I will deal with them as accurately as possible.

[Robert Goren]

Only LA breds can 2 Step. :lol:

Most Tx breds think they can too. I know because every time I bet on one they try it coming out of the gate.:bang: :lol:

[Bobby Seller]

Most Tx breds think they can too. I know because every time I bet on one they try it coming out of the gate.:bang: :lol:
LOL, at least yours come out the gate, mine won't run a step :liar: :p

[Steve 'StatMan']

Horses just don't run that slow unless, as Charles Carroll says in Handicapping Speed, you're handicapping harness horses or Clydesdales. Clydesdales are about 10 feet per length so the tradirional measures of 10 feet per length and .2 lengths per second would be more accurate for them.

I imagine Clydesdales are about 10 feet long, but since I've never seen them pulling anything other than a beer truck for show, I don't think they're ever going fast enough to be .2 lengths per second (that's still about 30 mph).

I'm being a little silly to point out it isn't just how long the horse is, it's also how fast they are going when one measures them, and it is likely worth more (be faster) when the pace is faster, and slower (take longer) when they are running slower. It can could depend on whether it is a slow paced turf race that is faster late than a fast paced dirt sprint that is slower at the end. And of course, the photo finish is based on what they have the camera set at - and they can speed or slow that depending on the expected finishing speed of the class level, so the image isn't too distorted, but likely uses a standard measurement for the beaten length. So we truly never know how fast a horse is traveling unless we only use info from tracks with a precise measuring system, such as a Trackus system in place at Woodbine and Del Mar, and maybe a rare one or two more. Or switch and just bet Quarter Horses.

[raybo]

I imagine Clydesdales are about 10 feet long, but since I've never seen them pulling anything other than a beer truck for show, I don't think they're ever going fast enough to be .2 lengths per second (that's still about 30 mph).

I'm being a little silly to point out it isn't just how long the horse is, it's also how fast they are going when one measures them, and it is likely worth more (be faster) when the pace is faster, and slower (take longer) when they are running slower. It can could depend on whether it is a slow paced turf race that is faster late than a fast paced dirt sprint that is slower at the end. And of course, the photo finish is based on what they have the camera set at - and they can speed or slow that depending on the expected finishing speed of the class level, so the image isn't too distorted, but likely uses a standard measurement for the beaten length. So we truly never know how fast a horse is traveling unless we only use info from tracks with a precise measuring system, such as a Trackus system in place at Woodbine and Del Mar, and maybe a rare one or two more. Or switch and just bet Quarter Horses.

What you're saying is true, but, the length of every single horse in every race cannot be obtained, so we have to choose something to use. Whether you choose 8', 9', 10' or whatever, makes no difference in regards to how much time per length you use, whether it is a static time (.2 seconds for example) or whether you use a dynamic, variable time per length based on the fractional leader's time. The important thing here is whether you use a static time or a dynamic time. Either way you're going to have error but you will have more error using static time than using dynamic time, because the time per length is dependent on how fast the horse is moving. Using a static time ignores that fact.

[harness2008]

I've always used dynamic time, never thought that the 1 length = 1/5 second worked for the T-Breds. Here's my example using feet-per-second and 9 feet equaling a length.

6 furlong race run in 112:0, horse finished 5 lengths behind.

6 furlongs = 3960 feet.

5 BL times 9 = 45

3960 minus 45 = 3915

3915 divided by 72 seconds = 54.375 fps

3960 divided by 54.375 = 72.828 seconds (final running time of 112:828)

[cj]

I've always used dynamic time, never thought that the 1 length = 1/5 second worked for the T-Breds. Here's my example using feet-per-second and 9 feet equaling a length.

6 furlong race run in 112:0, horse finished 5 lengths behind.

6 furlongs = 3960 feet.

5 BL times 9 = 45

3960 minus 45 = 3915

3915 divided by 72 seconds = 54.375 fps

3960 divided by 54.375 = 72.828 seconds (final running time of 112:828)

That works, but you can go even further by breaking it down to the last fraction and working with that instead of the time over the whole distance.

[raybo]

That works, but you can go even further by breaking it down to the last fraction and working with that instead of the time over the whole distance.

Exactly!

I break all call times down into fractional dynamic times per beaten length, also using fractional variants to adjust them further.

[Trotman]

Very interesting thread here regarding BL. What would you say the BL would be for a Harness horse.

[raybo]

Very interesting thread here regarding BL. What would you say the BL would be for a Harness horse.

Don't you harness players discuss this sort of thing? Do the harness results even state B/Ls? I've never even looked at harness data.

I assume harness horses are traveling at much lower velocities than thoroughbreds, in which case, the time per beaten length, for harness racing, would be quite a bit longer than a thoroughbred beaten length time.

[HUSKER55]

formulator publishes the finish times of each horse or you can display the winners time.

if the winner ran 3960 in 72 then you have 55.00 that means your horse ran it in 73 second for 54.25


what am I missing?

thanks

[raybo]

formulator publishes the finish times of each horse or you can display the winners time.

if the winner ran 3960 in 72 then you have 55.00 that means your horse ran it in 73 second for 54.25


what am I missing?

thanks

How'd you get the "54.25"?

Was it timed or does Formulator calculate it? If Formulator calculates it, how does it do that, what is the formula? Are there fractional times in harness? If so, does Formulator also adjust the times for the fractions? If so, how do they adjust them, what is the formula?

[HUSKER55]

3960/72 =55
3960/73 = 54.25

IT IS VELOCITY. RATE X RIME = DISTANCE IF YOU ARE USING 5 LENGTHS PER SECOND THEn FIVE LENGTHS IS ONE SECOND. DistNCE COVERED IS 3960 Horse a did it in 72 and horse B did it in 73

[raybo]

3960/72 =55
3960/73 = 54.25

IT IS VELOCITY. RATE X RIME = DISTANCE IF YOU ARE USING 5 LENGTHS PER SECOND THEn FIVE LENGTHS IS ONE SECOND. DistNCE COVERED IS 3960 Horse a did it in 72 and horse B did it in 73

Are all harness horses timed individually? No, I don't think so. So, you must be trusting Formulator's calculations. Is that correct?

[Cratos]

Earlier in this thread I posted a calculation using a static model or .166 seconds per length and it would yield a time for the second place horse based on the information given of 1:12.83.

However the dynamic model is more efficient and is calculated as follows:

• 3960 (distance in feet) ÷ [72 (race time in seconds)/.2]
= 11 feet per 1/5 second

• Length assumed to be 9 feet

• Distance behind in feet = 5*9 = 45 feet

• Dynamic time lengths = 45 ÷ 11 = 4.090909

• Time behind in 1/5 seconds = 4.090909 *.2 = .818180

• 2nd place horse’s time = 1st place horse time + .818180 = 1:12 + .818180 = 1:12.82 which is slightly better than the static model

• But the real value of the dynamic model comes as the final time changes.

• If the 6F race was run in 1:09 we would see the following:

• Static model = 1:09.83

• Dynamic Model = 1:09.78

Additionally, if the dynamic model is calculated by race segments (e.g. quarter times), pace would be included and the model would be more efficient.

[raybo]

Earlier in this thread I posted a calculation using a static model or .166 seconds per length and it would yield a time for the second place horse based on the information given of 1:12.83.

However the dynamic model is more efficient and is calculated as follows:

• 3960 (distance in feet) ÷ [72 (race time in seconds)/.2]
= 11 feet per 1/5 second

• Length assumed to be 9 feet

• Distance behind in feet = 5*9 = 45 feet

• Dynamic time lengths = 45 ÷ 11 = 4.090909

• Time behind in 1/5 seconds = 4.090909 *.2 = .818180

• 2nd place horse’s time = 1st place horse time + .818180 = 1:12 + .818180 = 1:12.82 which is slightly better than the static model

• But the real value of the dynamic model comes as the final time changes.

• If the 6F race was run in 1:09 we would see the following:

• Static model = 1:09.83

• Dynamic Model = 1:09.78

Additionally, if the dynamic model is calculated by race segments (e.g. quarter times), pace would be included and the model would be more efficient.

I thought I covered all this earlier, the math, and the fractionals.

[raybo]

Also, it appears that with all your above "dynamic" calculations, you're still using 1/5 of a second. Why would you want to go through all that trouble and still use something as antiquated as 1/5 second?

Here's how I would do it:

Distance = 3960'
Winner's time = 1:12 (72 seconds)
Average horse length = 9'
2nd place horse's beaten lengths = 5

3960'/72 seconds = 55 fps (feet per second) winner's avg. velocity
5 B/Ls X 9' = 45' (2nd place horse's distance behind)
45' / 55 fps = .81818 seconds behind
72 seconds + .81818 = 72.81818
or 1:12.82

And, of course, I adjust the fractions the same way, based on the fractional leader's avg velocity for the fraction.

[Cratos]

Also, it appears that with all your above "dynamic" calculations, you're still using 1/5 of a second. Why would you want to go through all that trouble and still use something as antiquated as 1/5 second?

Here's how I would do it:

Distance = 3960'
Winner's time = 1:12 (72 seconds)
Average horse length = 9'
2nd place horse's beaten lengths = 5

3960'/72 seconds = 55 fps (feet per second) winner's avg. velocity
5 B/Ls X 9' = 45' (2nd place horse's distance behind)
45' / 55 fps = .81818 seconds behind
72 seconds + .81818 = 72.81818
or 1:12.82

And, of course, I adjust the fractions the same way, based on the fractional leader's avg velocity for the fraction.

The one-fifth second metric is used because it is an accepted metric to normalize the relative (and that is what the calculated time is, relative) time too. What is funny is that both calculations (yours and mine) yielded the same relative time and you are arguing; go figure?

[raybo]

The one-fifth second metric is used because it is an accepted metric to normalize the relative (and that is what the calculated time is, relative) time too. What is funny is that both calculations (yours and mine) yielded the same relative time and you are arguing; go figure?

I'm not arguing, I just don't see the reason for including 1/5 seconds in the calculations. If you want to include fifths, why not just convert the adjusted times from digital to fractional (hundreds to fifths), like 1:12.82 = 1:12.4 or 1:12 4/5?

[Cratos]

I'm not arguing, I just don't see the reason for including 1/5 seconds in the calculations. If you want to include fifths, why not just convert the adjusted times from digital to fractional (hundreds to fifths), like 1:12.82 = 1:12.4 or 1:12 4/5?

I am quite sure you know this; it is called relative formulation and as long as 1 second is the base time number, which segments of a second that is used as the relative metric shouldn’t make any difference.

[raybo]

I am quite sure you know this; it is called relative formulation and as long as 1 second is the base time number, which segments of a second that is used as the relative metric shouldn’t make any difference.

Ok, not familiar with "relative formulation" but I get it. Just seems more complicated and confusing to formulate than the way I do it.

Bottom line, I guess, is the result is the same.

[ranchwest]

I am quite sure you know this; it is called relative formulation and as long as 1 second is the base time number, which segments of a second that is used as the relative metric shouldn’t make any difference.

Varying the measurement of distance or time does not affect the relationship among losers, but it does affect the relationship to the winner.

[Cratos]

Varying the measurement of distance or time does not affect the relationship among losers, but it does affect the relationship to the winner.

I think I understand what you are inferring, but please explain.

[ranchwest]

I think I understand what you are inferring, but please explain.

If you adjust your standard for time or distance, all of the horses will still be in sequence and will have the same relative relationship of speed measurement among all losers.

However, if you're sliding the time or distance, then there is a difference for every horse relative to the winner because the winning speed is fixed.

I hope this explanation is easier to understand.

[Cratos]

If you adjust your standard for time or distance, all of the horses will still be in sequence and will have the same relative relationship of speed measurement among all losers.

However, if you're sliding the time or distance, then there is a difference for every horse relative to the winner because the winning speed is fixed.

I hope this explanation is easier to understand.

You are correct there is a difference for every horse in a given race relative to the winner’s time.

But what change is that the horses that are behind the horse in front, either run faster or slower. The reason for this is that the interval as given by the chart caller doesn’t change, but the time to cover that interval changes.

What this is being said is that a 6 length interval in terms of speed of a horse finishing second to a winning horse in a mile in 1:34 is difference than a 6 length interval in terms of speed of horse finishing second to a winning horse in a mile in 1:40.

The reason being is that the relative time to cover the 6 length interval at the 1:34 time metric faster than at the 1:40 time metric.

[HUSKER55]

Cratos,

Thanks for clearing that up for me.

I appreciate it.

[ranchwest]

You are correct there is a difference for every horse in a given race relative to the winner’s time.

But what change is that the horses that are behind the horse in front, either run faster or slower. The reason for this is that the interval as given by the chart caller doesn’t change, but the time to cover that interval changes.

What this is being said is that a 6 length interval in terms of speed of a horse finishing second to a winning horse in a mile in 1:34 is difference than a 6 length interval in terms of speed of horse finishing second to a winning horse in a mile in 1:40.

The reason being is that the relative time to cover the 6 length interval at the 1:34 time metric faster than at the 1:40 time metric.

Not quite what I was saying.

If a horse goes a mile in 1:40 flat, that is his time no matter how long you say a length is -- 8, 9, 10, 11 ft., whatever. But if you're comparing a loser to a winner, then you are calculating different speeds for the loser when you vary the distance of a length. All losers are on the same scale, but the scale for the winner is not really the same and is fixed.

It's the same if you vary the time measurement.

[Cratos]

Not quite what I was saying.

If a horse goes a mile in 1:40 flat, that is his time no matter how long you say a length is -- 8, 9, 10, 11 ft., whatever. But if you're comparing a loser to a winner, then you are calculating different speeds for the loser when you vary the distance of a length. All losers are on the same scale, but the scale for the winner is not really the same and is fixed.

It's the same if you vary the time measurement.

You are correct; all losers are on the same scale with respect to distance, but with respect to speed the scale should be in my opinion the winner’s time scale. This is based on how fast the race was run.

For example if a horse finishes second to the winner by 3 lengths and the distance behind would be 24 feet for a length of 8 feet. If 9 feet is used the distance is assumed to be 27 feet.

However if the distance of the race is 5f and the winner’s time is 1 minute then you have the following times for the second place finisher:

At 1 length = 8 feet the second place finisher finished .44 seconds behind the winner or a time of 1:00.44 seconds

At 1 length = 9 feet the second place finisher finished .49 seconds behind the winner or a time of 1:00.49 seconds

When it comes to time estimation for the horses behind, the handicapper is at the mercy of the chart caller, but the general accepted measurement of a length is 9 feet

[ranchwest]

Well, there you go. I use 8 feet. Ha! To each his own.

[classhandicapper]

This is very important issue for figure handicappers that I've often brought up. Here's an example of what I mean.

Assume a 5 horse field where 4 of the horses typically run a BEYER around 90. They finish in a bunch and another horse beats them by ten lengths at 6F.

Using Beyer's beaten length chart suggests the winner probably ran about a 104. (10 lengths = 24 points)

Using the beaten length chart I use suggests the winner probably ran a 100 (I use a different number of feet for one length).

Turn the margin into 12-15 lengths etc... and you get ever larger discrepancies. This is one of several reasons I have suggested that some big winners may be overrated by some figure makers based on final time.

[classhandicapper]

Just a note:

The chart caller DOES NOT estimate beaten lengths at the finish. Each horse is timed as it crosses the finish line and the differences between their times are converted into beaten lengths via a formula that is not public.

[JPinMaryland]

how is it that the formula is not known? Why cant it be reverse engineered by looking at the BL and comparing them on the video via a stopwatch or by whatever method you used to come up with the ky derby final fractions in the other thread. (or did you get those simply from the official BL?)

Related questions: Does this formula vary from track to track?

What do they use at the mile call and other calls prior to the finish? Simply the race call?

For how long have they been doing it this way?

Looking at the final fractional times posted by DePalma, it would seem that BL according to the official beaten lengths in the chart is more like 1 BL = 1/4 second. At least that is what it seems like quckly comparing Ice Box vs Make Music and Super Saver vs Paddy....DePalma used a "timeline" and possibly stop watch as well.

I am not sure how you came up with your final fraction for Ice Box so I cannot comment on that. Did you use something different?

tx

[Tom]

Why don't they just report the times and forget the BLs?

[classhandicapper]

how is it that the formula is not known? Why cant it be reverse engineered by looking at the BL and comparing them on the video via a stopwatch or by whatever method you used to come up with the ky derby final fractions in the other thread. (or did you get those simply from the official BL?)

Related questions: Does this formula vary from track to track?

What do they use at the mile call and other calls prior to the finish? Simply the race call?

For how long have they been doing it this way?

Looking at the final fractional times posted by DePalma, it would seem that BL according to the official beaten lengths in the chart is more like 1 BL = 1/4 second. At least that is what it seems like quckly comparing Ice Box vs Make Music and Super Saver vs Paddy....DePalma used a "timeline" and possibly stop watch as well.

I am not sure how you came up with your final fraction for Ice Box so I cannot comment on that. Did you use something different?

tx


I know several people have asked for information on this issue and no one (including me) ever got a satisfactory response. If someone does know the answer he is keeping it to himself.

I backed into an answer that I think is close enough, but only after CJ enlightened me on this and we had a few conversations about it.

[gm10]

Ah-Haaaa! I was wondering if someone was doing what I have been thinking about for quite sometime. Do you adjust by the leader's pace at each call (using distance traveled by the leader versus distance traveled by each horse in the same amount of time)?

One can figure the leader's average speed per fraction easy enough. Based on the distance traveled by the leader vs. the distance traveled by each horse (using 8', 9', 10' per length, shouldn't matter as long as you apply the same distance per length to all horses, you can compare distances traveled by all horses and therefore the average time per length).

It's hard to explain my thinking but I've come up with a formula that varies according to the fraction leader's average pace rather than using a static time per length regardless of the pace or distance traveled in any particular fraction, and the race distance itself.

Yep that is the way to do it.

[JPinMaryland]

I know several people have asked for information on this issue and no one (including me) ever got a satisfactory response. If someone does know the answer he is keeping it to himself.

I backed into an answer that I think is close enough, but only after CJ enlightened me on this and we had a few conversations about it.

Yes I had heard that before about the BL and a formula. I assume they would do it all the tracks as you cant simply go by beaten lengths when the leader hits the wire, they could change as horses pass each other. So it made sense when I first heard it and I would assume most/all tracks do it that way.


which leads me to think it shoud not be too hard to reverse engineer the numbes and see what formula they are using. There should be plenty of data pts out there.

[michiken]

Why don't they just report the times and forget the BLs?Go to the front of the class for best idea of 2010!

http://paceadvantage.com/forum/images/SCNums/1.jpg

[Cratos]

how is it that the formula is not known? Why cant it be reverse engineered by looking at the BL and comparing them on the video via a stopwatch or by whatever method you used to come up with the ky derby final fractions in the other thread. (or did you get those simply from the official BL?)

Related questions: Does this formula vary from track to track?

What do they use at the mile call and other calls prior to the finish? Simply the race call?

For how long have they been doing it this way?

Looking at the final fractional times posted by DePalma, it would seem that BL according to the official beaten lengths in the chart is more like 1 BL = 1/4 second. At least that is what it seems like quckly comparing Ice Box vs Make Music and Super Saver vs Paddy....DePalma used a "timeline" and possibly stop watch as well.

I am not sure how you came up with your final fraction for Ice Box so I cannot comment on that. Did you use something different?

tx

This is an age old dilemma which has its genesis in 1 length equals one-fifth second in a horse race.

That assertion cannot be true because the distance system of measurement for one-fifth second per unit of time varies with respect to the final time of the race.

However a length can be standardized, but the one-fifth second metric will just be the denominator for normalization.

For example if a horse ran 6 furlongs in 1:12, it covered the 3,960 feet in 360 fifths of a second or 11 feet/ one-fifth second. If another horse ran the 6 furlongs in 1:09, it would have covered the 6 furlong distance at a rate of 11.45 feet per one-fifth second.

Therefore if we standardized a length at 9 feet we would have the first metric being 1.22 lengths per one-fifth second and the second metric would be 1.27 lengths per one-fifth second. That is why I suggest that beaten lengths should vary with the final time of the race.

[JPinMaryland]

Shouldnt they vary based on the final fraction of the race or perhaps the final velocity of the leader?

[ranchwest]

A length is a distance. A second is time. The two are NEVER the same (except in the very misguided viewpoint of some people in horse racing).

[raybo]

A length is a distance. A second is time. The two are NEVER the same (except in the very misguided viewpoint of some people in horse racing).

Are you saying that we should not try to adjust fractional times and final times to reflect what each horse ran at those points?

[ranchwest]

Are you saying that we should not try to adjust fractional times and final times to reflect what each horse ran at those points?

No, not at all. I'm saying that considering a length to be a unit of time is inherently flawed.

[JPinMaryland]

oh I think I see what Ranch is saying, anytime you use length to indicate time it is going to be messed up because unless everyone agrees on a standard velocity there is no way to convert length to time and make it accurate for every horse in the race.

I would think the best choice would be to use the winner's velocity in the last fraction if you are going to measure say the final fractions of the runners up. .

[Tom]

Is it worth it for fractional times?
Take 4 furlongs in 44 flat and 48 flat.
Assume 10 feet per length, the value of a length in time is .167 and .183.
Do you gain that much considering your 1 beaten length itself might be one and a half or three quarters to begin with?

[46zilzal]

THINK ABOUT the velocity of the assembled group of horses as they move in relation to one another. Dick Mitchell suggests using a variable measure depending upon what fraction the horses were in at the time,

i.e in a sprint of 22 1/5 are the beaten lengths as dramatic as in the final fraction of that same sprint where the field was moving at 25 3/5?

[ranchwest]

If the leader goes a quarter in 24 and another horse is 2 lengths off, that horse's time is also 24, but his distance is a quarter less 2 lengths. He hasn't run a quarter yet.

[JPinMaryland]

I use 7 lenght/sec when the fraction is e.g. 24 sec and 6/sec when it is slower but after reading the fractions that DePalma posted Im wondering if there is some inherent error in all this..

[Cratos]

If the leader goes a quarter in 24 and another horse is 2 lengths off, that horse's time is also 24, but his distance is a quarter less 2 lengths. He hasn't run a quarter yet.

A very good reply

[raybo]

If the leader goes a quarter in 24 and another horse is 2 lengths off, that horse's time is also 24, but his distance is a quarter less 2 lengths. He hasn't run a quarter yet.

I agree with this statement as fact.

Now, how do we use this fact, in determining another horse's performance vs the leader's clocked time? Should we assign a standard horse length to all horses in the race. Should we calculate the time it took to run an assigned horse length, using the leader's clocked time and the number of horse lengths that were run to the clocking point?

Any way you do it there are inherent errors, like acceleration or deceleration of horses after the fraction was clocked, and, the clockers' accuracy in determining how many lengths each horse was behind, and , what is that clocker using for the distance or time of a horse length.

It just seems to me that the very best we can do, with present technology and what is made public regarding the measurements/lengths behind, is to use the clocked time of the leader of each fraction, which is probably the most accurate information for each fraction, then try to figure out how fast the field is moving at the point the leader was clocked. I realize that some horses are accelerating, trying to get closer to the leader, while others are decelerating, for various reasons, while still others are moving at the leader's velocity. How do we take as much error out, of these scenarios, as possible?

[Bruddah]

Let me first say, this subject has had my attention and interest for years. For many many years the best we could convert beaten lenghs was by using the old figure of 1 lgth = 1/5 sec. With the advent of the computer and calculators, we were able to refine it to a better number. In essence we could refine the time to 10th's of a second as compared to fifths of a second. We can now extrapolate that number in thousandths.

But, I ask, does any of this give us an edge, or is it just busy work for mathmatical minds. I have yet to discover the edge of refining this number further. I just ask someone to show me the edge, or is this just a SNIPE HUNT. :confused:
(If you're not from the South, you probably won't understand what a SNIPE HUNT is.

[ranchwest]

I agree with this statement as fact.

Now, how do we use this fact, in determining another horse's performance vs the leader's clocked time? Should we assign a standard horse length to all horses in the race. Should we calculate the time it took to run an assigned horse length, using the leader's clocked time and the number of horse lengths that were run to the clocking point?

Any way you do it there are inherent errors, like acceleration or deceleration of horses after the fraction was clocked, and, the clockers' accuracy in determining how many lengths each horse was behind, and , what is that clocker using for the distance or time of a horse length.

It just seems to me that the very best we can do, with present technology and what is made public regarding the measurements/lengths behind, is to use the clocked time of the leader of each fraction, which is probably the most accurate information for each fraction, then try to figure out how fast the field is moving at the point the leader was clocked. I realize that some horses are accelerating, trying to get closer to the leader, while others are decelerating, for various reasons, while still others are moving at the leader's velocity. How do we take as much error out, of these scenarios, as possible?

Let's go through the simple case of computing f1, f2 and f3, the three basic race segments. Let's assume 6f, where the segments will end at 2f, 4f and 6f.

Let's follow up on the example mentioned before:

24.00 and 2 beaten lengths
48.11 and 3 beaten lengths
1:12.75 and 4 beaten lengths

So, our distances:

((440 yards * 3 feet_in_a_yard) - (2 lengths * 8 feet_per_length)) ==
1320 feet - 16 feet == 1304 feet

((880 yards * 3 feet_in_a_yard) - (3 lengths * 8 feet_per_length)) ==
2640 feet - 24 feet == 2616 feet

((1320 yards * 3 feet_in_a_yard) - ( 4 lengths * 8 feet_per_length)) ==
3960 feet - 32 feet == 3928 feet

f1_feet == 1304
f2_feet == 2616 - 1304 == 1312
f3_feet == 3928 - 2616 == 1312

f1_time == 24.00
f2_time == 48.11 - 24.00 == 24.11
f3_time == 1:12.75 - 48.11 == 72.75 - 48.11 == 24.64

f1 == 1304 / 24 == 54.33 feet per second
f2 == 1312 / 24.11 == 54.41 feet per second
f3 == 1312 / 24.64 == 53.24 feet per second

So, why did I use 8 feet per length? Some people use 9, 10, 11 or other values -- some people even use a sliding scale. I settled on 8 because I'm thinking that a smaller value may be more predictive. The real velocity figure might be closer to 9 or 10. I am de-emphasizing the "penalty" for beaten lengths. Now, I don't have any studies behind this thought. I just know I'm satisfied with what I'm doing.

There's two keys here.

First, I think the above provides a valid formula as far as keeping distance related to distance and such. This allows for a better foundation for understanding more advanced studies.

Second, a consistent formula allows for getting a grip on the figures, whether they are technically exact or not. If you always do the same thing, you're likely to at some point grasp what the numbers are telling you.

[raybo]

Let me first say, this subject has had my attention and interest for years. For many many years the best we could convert beaten lenghs was by using the old figure of 1 lgth = 1/5 sec. With the advent of the computer and calculators, we were able to refine it to a better number. In essence we could refine the time to 10th's of a second as compared to fifths of a second. We can now extrapolate that number in thousandths.

But, I ask, does any of this give us an edge, or is it just busy work for mathmatical minds. I have yet to discover the edge of refining this number further. I just ask someone to show me the edge, or is this just a SNIPE HUNT. :confused:
(If you're not from the South, you probably won't understand what a SNIPE HUNT is.

I hear what you're saying, but, if we're going to use beaten lengths, at all, shouldn't they at least have some qualitative value that takes into account the velocity at which they were attained? I'm sorry, but, 1/5 second or .2 seconds, or .23 or .225 seconds just doesn't get it for me.

Even Brisnet, uses different lengths multipliers, for their speed figures, for different distances (due to the differing velocities typically run for those distances), and that makes sense, so, wouldn't the same line of thought apply to pace figures, and ultimately, to fractional adjusted times?

[ranchwest]

Let me first say, this subject has had my attention and interest for years. For many many years the best we could convert beaten lenghs was by using the old figure of 1 lgth = 1/5 sec. With the advent of the computer and calculators, we were able to refine it to a better number. In essence we could refine the time to 10th's of a second as compared to fifths of a second. We can now extrapolate that number in thousandths.

But, I ask, does any of this give us an edge, or is it just busy work for mathmatical minds. I have yet to discover the edge of refining this number further. I just ask someone to show me the edge, or is this just a SNIPE HUNT. :confused:
(If you're not from the South, you probably won't understand what a SNIPE HUNT is.

You are exactly right that the old 1 length = 1/5 second was contrived to make it easier to figure velocity without a computer or calculator. We don't need that any more.

How much of an edge do we get? Well, to be honest, probably the biggest edge is that we're doing these computations at all. A lot of people don't.

If we get the best number we can get, then we have a foundation for more advanced studies. We can use the raw numbers or we can adjust for track variance, distance, surface, track, etc. or any combination. We can also combine the segments (f1, f2, f3) into various complex numbers.

Keep in mind that a lot of people take a look at these segment numbers from a pace line race -- a race the handicapper feels represents the basis for the evaluation he seeks.

This is all something you can do or not. Maybe try it and see if you like it. If you can't ever get an advantage out of it, you can always drop it at any time.

[ranchwest]

I hear what you're saying, but, if we're going to use beaten lengths, at all, shouldn't they at least have some qualitative value that takes into account the velocity at which they were attained? I'm sorry, but, 1/5 second or .2 seconds, or .23 or .225 seconds just doesn't get it for me.

Even Brisnet, uses different lengths multipliers, for their speed figures, for different distances (due to the differing velocities typically run for those distances), and that makes sense, so, wouldn't the same line of thought apply to pace figures, and ultimately, to fractional adjusted times?

Lengths are not seconds. Forget that you ever heard they were.

We're adjusting the distance to compare to the lead time. That tells us the velocity for a length. We never have to worry about inventing a time value for a length. Try to wrap your head around this different concept -- the old one was contrived to make computations without a calculator or computer.

[raybo]

Let's go through the simple case of computing f1, f2 and f3, the three basic race segments. Let's assume 6f, where the segments will end at 2f, 4f and 6f.

Let's follow up on the example mentioned before:

24.00 and 2 beaten lengths
48.11 and 3 beaten lengths
1:12.75 and 4 beaten lengths

So, our distances:

((440 yards * 3 feet_in_a_yard) - (2 lengths * 8 feet_per_length)) ==
1320 feet - 16 feet == 1304 feet

((880 yards * 3 feet_in_a_yard) - (3 lengths * 8 feet_per_length)) ==
2640 feet - 24 feet == 2616 feet

((1320 yards * 3 feet_in_a_yard) - ( 4 lengths * 8 feet_per_length)) ==
3960 feet - 32 feet == 3928 feet

f1_feet == 1304
f2_feet == 2616 - 1304 == 1312
f3_feet == 3928 - 2616 == 1312

f1_time == 24.00
f2_time == 48.11 - 24.00 == 24.11
f3_time == 1:12.75 - 48.11 == 72.75 - 48.11 == 24.64

f1 == 1304 / 24 == 54.33 feet per second
f2 == 1312 / 24.11 == 54.41 feet per second
f3 == 1312 / 24.64 == 53.24 feet per second

So, why did I use 8 feet per length? Some people use 9, 10, 11 or other values -- some people even use a sliding scale. I settled on 8 because I'm thinking that a smaller value may be more predictive. The real velocity figure might be closer to 9 or 10. I am de-emphasizing the "penalty" for beaten lengths. Now, I don't have any studies behind this thought. I just know I'm satisfied with what I'm doing.

There's two keys here.

First, I think the above provides a valid formula as far as keeping distance related to distance and such. This allows for a better foundation for understanding more advanced studies.

Second, a consistent formula allows for getting a grip on the figures, whether they are technically exact or not. If you always do the same thing, you're likely to at some point grasp what the numbers are telling you.

So, your fps velocities are strictly distance oriented and do not attempt any qualifications regarding class of horses, speed of surface, variances from track to track, etc.(?)

If so, then, I assume you use additional methods to answer those other questions, as I'm sure you are very interested in knowing how those velocities were attained, ie., breezing, rating, pressuring or being pressured, etc., etc..

I know, for me anyway, those velocities mean very little, without at least some knowledge of how much energy was being expended in each fraction.

[raybo]

Lengths are not seconds. Forget that you ever heard they were.

We're adjusting the distance to compare to the lead time. That tells us the velocity for a length. We never have to worry about inventing a time value for a length. Try to wrap your head around this different concept -- the old one was contrived to make computations without a calculator or computer.

Ok, after rereading your example I see that you're adjusting lead times and arriving with the same adjustments, in average velocities, that I am. I call it a "variable time per beaten length multiplier", which is based on the leader's fractional time (not a static multiplier like 1/5 second), along with the fractional distance in feet, and then calculating his average velocity for the fraction.

We're using different calculation methods but arriving at the same adjustments.

[ranchwest]

So, your fps velocities are strictly distance oriented and do not attempt any qualifications regarding class of horses, speed of surface, variances from track to track, etc.(?)

If so, then, I assume you use additional methods to answer those other questions, as I'm sure you are very interested in knowing how those velocities were attained, ie., breezing, rating, pressuring or being pressured, etc., etc..

I know, for me anyway, those velocities mean very little, without at least some knowledge of how much energy was being expended in each fraction.

I do most of the Brohamer/Sartin computations. I use raw times for these. I also do a lot of other evaluations, including using BRIS figures, both fractional and final. I actually look at the BRIS fractional figures in several ways.

A lot of people do adjust their figures and I can see where that probably works for them.

I don't claim that my way is best or great. I just know it is working for me. I've been using f1, f2, f3 for about 27 years. I've only been using the Brohamer/Sartin computations about a month. I've also been working in a lot of aspects of the Sartin Methodology, though not nearly all. It's been the most consistently good month of my life.

[raybo]

You are exactly right that the old 1 length = 1/5 second was contrived to make it easier to figure velocity without a computer or calculator. We don't need that any more.

How much of an edge do we get? Well, to be honest, probably the biggest edge is that we're doing these computations at all. A lot of people don't.

If we get the best number we can get, then we have a foundation for more advanced studies. We can use the raw numbers or we can adjust for track variance, distance, surface, track, etc. or any combination. We can also combine the segments (f1, f2, f3) into various complex numbers.

Keep in mind that a lot of people take a look at these segment numbers from a pace line race -- a race the handicapper feels represents the basis for the evaluation he seeks.

This is all something you can do or not. Maybe try it and see if you like it. If you can't ever get an advantage out of it, you can always drop it at any time.

Well stated!

I used these velocity figures (incorporating a fractional track variant) combined into "composite" figures" (EP, AP, SP, TE,etc) and futher combining those "composite" figures into what became a performance rating, during PA's "Black Box Challenge", resulting in a 38% hit rate and a 24% overall ROI, totally calculated and picked by the program. Not too bad for 6 different tracks on 6 days.

I believe there is definite value in using these "pace variable" velocities vs the traditional static B/L method in adjusting fractional and final times.

[raybo]

I do most of the Brohamer/Sartin computations. I use raw times for these. I also do a lot of other evaluations, including using BRIS figures, both fractional and final. I actually look at the BRIS fractional figures in several ways.

A lot of people do adjust their figures and I can see where that probably works for them.

I don't claim that my way is best or great. I just know it is working for me. I've been using f1, f2, f3 for about 27 years. I've only been using the Brohamer/Sartin computations about a month. I've also been working in a lot of aspects of the Sartin Methodology, though not nearly all. It's been the most consistently good month of my life.

I've seen some of the Brohamer computations but have never bought the entire methodology and have thus far used only what I have stated, regarding velocity analysis. I'm sure as time goes by some of the Brohamer stuff will get into Alldata, once we get the database setup so these things can be better analyzed.

I have a "client"/collaborator who I helped with spreadsheet design and he, for all intents and purposes, introduced me to Sartin velocities. Once I got involved in the calculations and saw what they were telling me, I kinda got hooked and am still involved in "work in progress" with enhancing the basic figures.

[shouldacoulda]

This is quite a thread. I don't think it is that critical actually, as long as it is a constant number. The key becomes was he getting closer to, or further away from the leader. You have to remember that a length is an interpreted measurement and is not always accurate in the past performances. The number of lengths shown are only as good as the trackmen that call it. All the calculations in the world won't be accurate if the baseline data is askew.

[gm10]

If the leader goes a quarter in 24 and another horse is 2 lengths off, that horse's time is also 24, but his distance is a quarter less 2 lengths. He hasn't run a quarter yet.

True, but if it wasn't true, it'd be because we have individual times for each horse and there wouldn't be any problem!!

[ranchwest]

This is quite a thread. I don't think it is that critical actually, as long as it is a constant number. The key becomes was he getting closer to, or further away from the leader. You have to remember that a length is an interpreted measurement and is not always accurate in the past performances. The number of lengths shown are only as good as the trackmen that call it. All the calculations in the world won't be accurate if the baseline data is askew.

Yes, but it does make sense to do the best we can with what we have.

Two screwey approaches won't make things right.

[ranchwest]

True, but if it wasn't true, it'd be because we have individual times for each horse and there wouldn't be any problem!!

The objective in this thread was to work with what we've got.

[Tom]

Lengths are not seconds. Forget that you ever heard they were.



But, if that is how the lengths are arrived at by EB to begin with, they really are equal to seconds.

[ranchwest]

But, if that is how the lengths are arrived at by EB to begin with, they really are equal to seconds.

That's only true at the finish. If you want to work from that approach at the finish, I cannot disagree with your logic. I wish that we could get a confirmation of how that determination is made at each track.

Can you imagine if each NFL team kept stats with their own guidelines, you couldn't find out what those guidelines were and you had to pay for the information to get it? There would probably be fewer people watching football. Hmmm.

[Tom]

If you were brand new to the game, and the first thing you learned was how they collect and manage the data, I doubt you would continue.

[ranchwest]

And, let's not forget how it is distributed. Different prices, different data. Nobody sells all of the data. Some data is cleaned, some is not.

[shouldacoulda]

Yes, but it does make sense to do the best we can with what we have.

Two screwey approaches won't make things right.

I agree, but my point is tolerances. How important is difference in unit of measure of a length from .2 to .176? If your unit of measure (lengths) is off by 1 or 2 the .024 difference in a multiplier becomes insignificant IMO. Pace for me is done at the end of the handicapping chain between the contenders, usually 3 or 4 horses. The rest are "filler". If there are too many contenders it's a pass.

[ranchwest]

I agree, but my point is tolerances. How important is difference in unit of measure of a length from .2 to .176? If your unit of measure (lengths) is off by 1 or 2 the .024 difference in a multiplier becomes insignificant IMO. Pace for me is done at the end of the handicapping chain between the contenders, usually 3 or 4 horses. The rest are "filler". If there are too many contenders it's a pass.

I guess you could watch the replays and make your own beaten length figures.

The ones in the data files seem to work for me.

[kenwoodall2]

"Length= Unit of measurement in racing and charting terminology. The length of an American Quarter Horse, the distance from the horse's nose to the tip of his flying tail, is .16 seconds (16/100ths), while the length of a Thoroughbred is .20 seconds (20/100ths)."
Actually, Quarterhorses run about 66 feet per second, 11 feet in .16 seconds; TBreds run just under 55 feet per second, 11 feet in .20 seconds.
But the real amswer is that since the accurate finish chart figuring is only listed on the AQHA website under "Racing Glossary", then time for beaten lengths only count for QTRS and TBREDS who race 4 1/2f or less on the same card as QTRS!
On mixed cards, chart callers switch from .2 to .16 using the same camera, depending on what breed is running. Arabians, Appys, or Mules I assume depends on the distance run?!!
Anyway, after finding out all that, I just use NO time difference if under 2 beaten lenghths, and for over 2 beaten lengths if I have to use a calculation I reduce B/L by 25% and use .2 for TBreds. But if I have to use beaten lengths to pick I usually just pass the race or bet longshots anyway!

[gm10]

I agree, but my point is tolerances. How important is difference in unit of measure of a length from .2 to .176? If your unit of measure (lengths) is off by 1 or 2 the .024 difference in a multiplier becomes insignificant IMO. Pace for me is done at the end of the handicapping chain between the contenders, usually 3 or 4 horses. The rest are "filler". If there are too many contenders it's a pass.

I was playing around with this a few months ago. I calculated the finishing speed of a number of races, and used those to convert B/L into time.

The result weren't spectacular but they are significant all the same. Let's take 6F sprints at Keeneland as an example. In 2010, the range of finishing speeds was 15.52 m/s to 17.69 m/s.

So for a horse beaten in the 'quickest' race (15.52 m/s), each length would equal 2.7 / 15.52 = 0.174 seconds. In the 'slowest' race (17.69 m/s), each length would be 2.7 / 17.69 = 0.153 seconds. (This assumes that a length is 2.7 meters).

So let's see what this does to B/L's:

1 length behind: time difference of 0.02 seconds.
2.5 lengths behind: time difference of 0.05 seconds.
5 lengths behind: time difference of 0.11 seconds
7 lengths behind: time difference of 0.15 seconds

Since 1 point on the BSF scale is approximately 0.05 seconds, it is fair to say that the finishing speed can make a BSF inaccurate by 1 point for each 2.5 B/L. (This is not a criticism of BSF btw.)

Finally, I must add that the inaccuracies become larger for longer distances (especially on the dirt where finishing speeds vary more and winning margins are larger).

[ranchwest]

gm10,

Good post.

[Cratos]

I was playing around with this a few months ago. I calculated the finishing speed of a number of races, and used those to convert B/L into time.

The result weren't spectacular but they are significant all the same. Let's take 6F sprints at Keeneland as an example. In 2010, the range of finishing speeds was 15.52 m/s to 17.69 m/s.

So for a horse beaten in the 'quickest' race (15.52 m/s), each length would equal 2.7 / 15.52 = 0.174 seconds. In the 'slowest' race (17.69 m/s), each length would be 2.7 / 17.69 = 0.153 seconds. (This assumes that a length is 2.7 meters).

So let's see what this does to B/L's:

1 length behind: time difference of 0.02 seconds.
2.5 lengths behind: time difference of 0.05 seconds.
5 lengths behind: time difference of 0.11 seconds
7 lengths behind: time difference of 0.15 seconds

Since 1 point on the BSF scale is approximately 0.05 seconds, it is fair to say that the finishing speed can make a BSF inaccurate by 1 point for each 2.5 B/L. (This is not a criticism of BSF btw.)

Finally, I must add that the inaccuracies become larger for longer distances (especially on the dirt where finishing speeds vary more and winning margins are larger).

Your analysis of 8.85825 feet (2.7 meters) is very close to the accepted length of 9 feet which was once posted in the DRF.

In my static calculations I assume that the chart caller is measuring the distances between horses in 9 feet increments, but that distance unit has be converted into distance/time units which I found to be on average a factor of 1.204 for every .167 seconds.

[shouldacoulda]

I guess you could watch the replays and make your own beaten length figures.

The ones in the data files seem to work for me.

You absolutely could. I am having trouble wording this properly. I know I am going to catch flack for this, but pace is not the end all be all in handicapping. If it was, this game would be a snap. I think too much emphasis is put on pace/speed. There is much more that goes into the dynamics of winning a race. My handicapping improved greatly when I stopped putting a pen to paper and just evaluated the race at a glance. Then when I have the race shape and contenders identified I will try to see who ran the fastest. I do the calculations in my head. Then you can usually narrow it down to two real contenders. At that point I play the one with the higher odds as long as it is higher than 2 or 3 to 1. There is too much randomness in racing to live and die by the numbers alone. If there was a mechanical approach to racing my Dad (may he rest in peace) would have found it. I watched him devote 50 years of his life to figuring out this game. He kept his own jockey/trainer numbers in the 70's way before computers. I wish he was still around so he could see how the knowledge he passed on to me has evolved. If pace works for you then use it. If the color of the jockeys jacket works or picking by name works, then use it. I developed my own style after many painful losses and let down wins (poor pay offs). It works for me and pace does play a role, but it's just another spoke in the wheel. In sales they say many small yeses add up to a big yes (sale made). It's the same principle same with handicapping.

[cj]

Since 1 point on the BSF scale is approximately 0.05 seconds, it is fair to say that the finishing speed can make a BSF inaccurate by 1 point for each 2.5 B/L. (This is not a criticism of BSF btw.)



Maybe I'm confusing things here, but 1 point on the Beyer scale is not equal to 0.05 seconds. Since the post talks about a 6f race, is it safe to assume you are talking about the Beyer scale at 6f?

If so, I think your calculations are off. 1 point at 6f is about 0.07 seconds. 0.05 seconds is worth about 0.7 points on the Beyer scale.

[JPinMaryland]

If you were brand new to the game, and the first thing you learned was how they collect and manage the data, I doubt you would continue.

Assuming you didn't know about the 20% takeout.. :eek:

[gm10]

Maybe I'm confusing things here, but 1 point on the Beyer scale is not equal to 0.05 seconds. Since the post talks about a 6f race, is it safe to assume you are talking about the Beyer scale at 6f?

If so, I think your calculations are off. 1 point at 6f is about 0.07 seconds. 0.05 seconds is worth about 0.7 points on the Beyer scale.

Beyer himself writes that 3 points equal 1 length in sprints. Since 1 length is about 0.15-0.18 sec, it's not a stretch to use the equation 1 point = 0.05 secs.

Anyway, that's not the point.

[gm10]

Your analysis of 8.85825 feet (2.7 meters) is very close to the accepted length of 9 feet which was once posted in the DRF.

In my static calculations I assume that the chart caller is measuring the distances between horses in 9 feet increments, but that distance unit has be converted into distance/time units which I found to be on average a factor of 1.204 for every .167 seconds.

But surely you must have a range of factors for difference surfaces, going conditions, and distances?

[windoor]

Why don't they just report the times and forget the BLs?

I use Post Time Daily 2.0 for my PP's, and in the "View Menu" I have the option of showing, Race times, Show Horses time, Show horses pace rating, Show race internal fractions, Show horses internal fractions.

I use "Show horses time" and Horses "internal fractions" all the time and put great faith in the numbers I'm seeing.

I have no idea how accurate they really are, but when used with my form cycle data, I usually do very well with them.

Regards,

Windoor

[cj]

Beyer himself writes that 3 points equal 1 length in sprints. Since 1 length is about 0.15-0.18 sec, it's not a stretch to use the equation 1 point = 0.05 secs.

Anyway, that's not the point.

I know this isn't true for a fact. Just take a look at his speed chart. Can't anyone here just says "Oops, I was wrong!"?

Actually, it is a very important point. You are using an erroneous assumption to then say " it is fair to say that the finishing speed can make a BSF inaccurate by 1 point for each 2.5 B/L."

[gm10]

I know this isn't true for a fact. Just take a look at his speed chart. Can't anyone here just says "Oops, I was wrong!"?

Actually, it is a very important point. You are using an erroneous assumption to then say " it is fair to say that the finishing speed can make a BSF inaccurate by 1 point for each 2.5 B/L."

Listen, buddy.

Beyer writes: "In sprint races, three points translates into a margin of a length or more" (Beyer On Speed, p. 33).

So I used one length = 3 points.

If you've got issues with that, then frankly, you should introduce a more accurate definition instead of nitpicking.

My point was that using a static definition of B/L causes inaccuracies that are significant but not spectacular. I don't care if a BSF can be wrong by 1 point at 2.3 B/L or 2.8 B/L. I never use BSF, but if you want to create a full correction chart, I'm sure people will be very appreciative.

[Cratos]

But surely you must have a range of factors for difference surfaces, going conditions, and distances?

Correct, and because the speed/time unit curve is downward-sloping you will find less feet per unit of constant time as the distance gets longer. For instance at the sprint you might see 11 feet/.167 seconds, whereas in the route it might be 10.60 feet/.167 seconds.

[cj]

Listen, buddy.

Beyer writes: "In sprint races, three points translates into a margin of a length or more" (Beyer On Speed, p. 33).

So I used one length = 3 points.

If you've got issues with that, then frankly, you should introduce a more accurate definition instead of nitpicking.

My point was that using a static definition of B/L causes inaccuracies that are significant but not spectacular. I don't care if a BSF can be wrong by 1 point at 2.3 B/L or 2.8 B/L. I never use BSF, but if you want to create a full correction chart, I'm sure people will be very appreciative.

Too funny. Just admit you were wrong. I don't care what Beyer said, it is all in print if you bother to look it up. Your statement was blatantly false.

When Beyer says one length is 3 points or more, it is because he is defining a sprint as any distance less than one mile. On average, it probably is 3 points. But we all know that at each distance the value is different. You used 6f, and at 6f 3 points is not only wrong, it is wrong by a lot. Your "it is fair to say" comment is very unfair. Notice he used "or more", clearly showing he wasn't trying to break things down to the 1/100th second as you were.

As for me introducing a more accurate definition, I already did for 6f. I'm not going to repost the whole Beyer speed chart to make you happy. Try Google! There is a novel idea. Of course had you done that in the first place you wouldn't have made the mistake you did.

We all make mistakes, at least I know most of us do, and I make more than my fair share. When I am wrong, I admit it instead of deflecting blame to other people. This is precisely what you are doing when you blame Beyer for your use of a statement taken completely out of context.

[Tom]

For those interested......

http://www.angelfire.com/la2/LouisianaRacing/SPRChartsBeyer.html


http://www.drf.com/products/beyers/beyers.html

http://www.drf.com/help/help_speedrate.html

[fmolf]

whatever measurement you do use just make sure you apply it to all horses times and bl's in the same exact way for all past races and comparisons horse to horse.

[gm10]

Too funny. Just admit you were wrong. I don't care what Beyer said, it is all in print if you bother to look it up. Your statement was blatantly false.

When Beyer says one length is 3 points or more, it is because he is defining a sprint as any distance less than one mile. On average, it probably is 3 points. But we all know that at each distance the value is different. You used 6f, and at 6f 3 points is not only wrong, it is wrong by a lot. Your "it is fair to say" comment is very unfair. Notice he used "or more", clearly showing he wasn't trying to break things down to the 1/100th second as you were.

As for me introducing a more accurate definition, I already did for 6f. I'm not going to repost the whole Beyer speed chart to make you happy. Try Google! There is a novel idea. Of course had you done that in the first place you wouldn't have made the mistake you did.

We all make mistakes, at least I know most of us do, and I make more than my fair share. When I am wrong, I admit it instead of deflecting blame to other people. This is precisely what you are doing when you blame Beyer for your use of a statement taken completely out of context.

Why are you so pig-headed???

This was never about Beyer. My only point was that converting B/L into time is not a static conversion. The only reason why I mentioned BSF was to illustrate the possible inaccuracies. I explicitly said it was not a criticism of BSF.

You are taking this into a whole different direction for reasons that are totally beyond me. Why are you so defensive about BSF? Take a chill pill, dude. This was a general observation, not a nuanced critique of how BSF can be affected by different finishing speeds.

[cj]

Why are you so pig-headed???

This was never about Beyer. My only point was that converting B/L into time is not a static conversion. The only reason why I mentioned BSF was to illustrate the possible inaccuracies. I explicitly said it was not a criticism of BSF.

You are taking this into a whole different direction for reasons that are totally beyond me. Why are you so defensive about BSF? Take a chill pill, dude. This was a general observation, not a nuanced critique of how BSF can be affected by different finishing speeds.

If I'm being pig-headed here I am not alone. While you said it was not a criticism, you also said it was fair to say the false data you provided led to the figures being wrong by a point for every two and half lengths. That is totally untrue and destroys the whole crux of your argument.

[HUSKER55]

I believe Beyer used 5 lengths per second in his book. He tells you how to build your own and how his method works..

Wouldn't it be possible to use 6,7,8 lengths per second or what ever you wanted to make your own scale?

would that work or not?

[cj]

I believe Beyer used 5 lengths per second in his book. He tells you how to build your own and how his method works..

Wouldn't it be possible to use 6,7,8 lengths per second or what ever you wanted to make your own scale?

would that work or not?

Beyer uses time to construct his speed charts. Each distance has a value for a fifth of a second. I don't think he goes into details about how his beaten lengths chart is constructed, but it is based on time as well from what I can tell. He never goes into what defines a length, but I'm sure it is consistent. I'm about 99.9% sure he does not use the 5 lengths per second rule.

[gm10]

If I'm being pig-headed here I am not alone. While you said it was not a criticism, you also said it was fair to say the false data you provided led to the figures being wrong by a point for every two and half lengths. That is totally untrue and destroys the whole crux of your argument.

OK, no problem.

Mr. P. Advantage, please delete the post with all the false data, so CJ can relax again.

[Cratos]

For those interested......

http://www.angelfire.com/la2/LouisianaRacing/SPRChartsBeyer.html


http://www.drf.com/products/beyers/beyers.html

http://www.drf.com/help/help_speedrate.html

Dean Keppler provided an adequate overview of the Beyer Speed Figure methodology even though he made some outlandish comments.

[Cratos]

I believe Beyer used 5 lengths per second in his book. He tells you how to build your own and how his method works..

Wouldn't it be possible to use 6,7,8 lengths per second or what ever you wanted to make your own scale?

would that work or not?

If I remember correctly, Andy Beyer stated in one of his earlier books (I don’t remember which) that there are 6 lengths to one-fifth second. However he might have later changed or modified that metric.

[cj]

If I remember correctly, Andy Beyer stated in one of his earlier books (I don’t remember which) that there are 6 lengths to one-fifth second. However he might have later changed or modified that metric.

He said about, but his charts clearly indicate this changes with distance.

[Cratos]

He said about, but his charts clearly indicate this changes with distance.

Okay, “about” is the adjective, but it doesn’t the intent of his assertion not unless you are seeking precision.

[Tom]

Tom Ainslie offered a chart that used 60 points per second and 10 per beaten length in his Complete Guide.

[ranchwest]

Maybe I'm confusing things here, but 1 point on the Beyer scale is not equal to 0.05 seconds. Since the post talks about a 6f race, is it safe to assume you are talking about the Beyer scale at 6f?

If so, I think your calculations are off. 1 point at 6f is about 0.07 seconds. 0.05 seconds is worth about 0.7 points on the Beyer scale.

I agree. You are confusing things here.

[Nitro]

I agree. You are confusing things here.

Would you like some more confusion (or perhaps some answers as to why its confusing)?
Check out a thread I started back on 11-08-2018 called:
A Critique of Speed: Fallacies in Creating Accurate Speed Figures

http://www.paceadvantage.com/forum/showthread.php?t=148582

[cj]

Would you like some more confusion (or perhaps some answers as to why its confusing)?
Check out a thread I started back on 11-08-2018 called:
A Critique of Speed: Fallacies in Creating Accurate Speed Figures

http://www.paceadvantage.com/forum/showthread.php?t=148582

You are replying to a post from 2010. I think we can go ahead and kill this thing off.