Joe Takach's error/ambiguity

[speculus]

Originally Posted by Hosshead

The horse on the rail is using more %E to fight centrifugal force than the outer horse.

I am sorry, but I really don't know what is this %E?

Can someone explain me so that I can follow the thread properly?

[PaceAdvantage]

Actually, I can't post any HTML without leaving that option set permanently to ON. If I post some HTML, then turn HTML off, it disables the HTML output of stuff previously posted while HTML was ON, ergo, no HTML on the board....sorry....

[BillW]

Originally Posted by speculus

I am sorry, but I really don't know what is this %E?

Can someone explain me so that I can follow the thread properly?

%E is a Sartin calculation representing "percent early energy" expended by a thoroughbred. It is calculated as follows:

%E = EP / (EP + LF ) where EP = early pace and LF = last fraction.

EP is the velocity in feet per second at the second call (distance in feet to 2nd call divided by time in seconds to 2nd call) and LF is in feet per second of the last fraction (dist of last fraction in feet divided by the time in seconds of the last call).

Speculus, paste your HTML table into a private message to me. I'll try to convert it to text for display.

Bill

[Hosshead]

As Bill described above, %E is an attempt by some SW to quantify the horse's proportion of race energy used early. But I was referring to %E in more of a basic way. For example you could calculate the %e of a horse Uncontested, running x-time EP, x-time LP etc. and get the %E.
Then take the Same Horse, Running Head and Head running the same times,= same %E, BUT I Guarantee that the Contested Pace will take more (Energy) out of the horse, leaving less for the stretch. Of course some will say that if it is a contested pace, the EP will be faster, and it often is, But I'm saying that Even If It Is The Same Time, it will cause both horses to use a higher % of there (finite) total Energy.

This is true on the turns, as some horses like to run on the turns,some don't.
Some can run good on the inside when not having a horse breathing down there neck, then use the 'Slingshot Effect" when straighting out into the stretch. But with a horse on the outside, they fold.

I'm just pointing out that it's fine to make all the calculations you want (I use them too), but HOW the Race is run, and the position of the horses, can make those figures work or not.

[speculus]

Gosavi's Formula
for Accurate Timing

T = t ÷ { 1 - (G* × L)}

where

T = Accurate timing upto 1/100th of a second
t = time of call (leading horse)
L = lengths behind* (see table 1.2),
and G* = Gosavi's constant

*Lengths behind is prone to human error unless
it is recorded from a frozen video frame

TABLE 1.1

Gosavi's Constant for
various distances G*
=====================================

Mile, furlongs, value of G*
3/16 1-1/2 0.00928

1/4 2 0.00696
3/8 3 0.00464
7/16 3-1/2 0.00398
1/2 4 0.00348
9/16 4-1/2 0.00309
5/8 5 0.00278
11/16 5-1/2 0.00253
3/4 6 0.00232
13/16 6-1/2 0.00214
7/8 7 0.00199
15/16 7-1/2 0.00186
1 8 0.00174
[1 mile & 70 yards] [8 fur & 70 yards] 0.00167
1-1/16 8-1/2 0.00164
1-1/8 9 0.00155
1-3/16 9-1/2 0.00147
1-1/4 10 0.00139
1-3/8 11 0.00127
1-1/2 12 0.00116
1-5/8 13 0.00107
1-7/8 15 0.00093
2 16 0.00087
2-1/8 17 0.00082

TABLE 1.2
Table of decimal values for verdict

Verdict Value of "L"

Deadheat 0

Short Head 0.1

Head 0.2

Short Neck 0.25

Neck 0.35

½ L 0.5

¾ L 0.75

1 L 1

1 ¼ L 1.25

1 ½ L 1.5

1 ¾ L 1.75

2 L 2

2 ¼ L 2.25

2 ½ L 2.5

2 ¾ L 2.75

3 L 3

Over 3 L Accordingly

[speculus]

Originally Posted by BillW

%E is a Sartin calculation representing "percent early energy" expended by a thoroughbred. It is calculated as follows:

%E = EP / (EP + LF ) where EP = early pace and LF = last fraction.

EP is the velocity in feet per second at the second call (distance in feet to 2nd call divided by time in seconds to 2nd call) and LF is in feet per second of the last fraction (dist of last fraction in feet divided by the time in seconds of the last call).

Speculus, paste your HTML table into a private message to me. I'll try to convert it to text for display.

Bill

I am very surprised. This is very strange. You are taking the RATIO of SPEED to SPEED, and calling it ENERGY? Why, may I ask?

Energy has the dimentions of ML^2T^-2 or (ML*L)/(T*T)

By no stretch of imagination can you call it percentage of energy or %E.

[chickenhead]

Originally Posted by speculus

The accuracy was + or - 1/100th of a second for upto six furlongs and 2/100th of a second for distances beyond mile, when compared to the device.

I love a bold statement, but I have to claim foul here, this is patently absurd. All horses would have to be running at the same velocity to be able to convert distance to time so accurately using the same constant, and I KNOW this is not true.

Doesn't mean I don't think your formula gives a good result, tho there are easier ways to do it if all you care about are finish times.

[DJofSD]

Don't get hung up on a label.

The very first time I heard Sartin talk about his idea and how he applied it, my training in the physical sciences started to blow a fuse.

Just think L'hospital rule and you'll be OK.

DJofSD

[speculus]

Originally Posted by chickenhead

I love a bold statement, but I have to claim foul here, this is patently absurd. All horses would have to be running at the same velocity to be able to convert distance to time so accurately using the same constant, and I KNOW this is not true.

You are wrong when you say "All horses would have to be running at the same velocity to be able to convert distance to time so accurately using the same constant".

I think you are confusing between the static and the dynamic view of the finish.

Suppose you have access to a still photograph of the finish when the winner is crossing the wire and the runner up is, say, 8 lengths behind. Now depending upon whether the runner up is slowing down or accelerating (of course with respect to the winner, make no mistake about that!), the actual verdict of lengths (that will go down into the charts or record books) between the two, would be more or less than 8 lenghts.

Why? Because the final verdict is decided by the image of the runner up on the photo finish strip, and it is nothing but the exact instance when the runner up crosses the wire.

If it is 7 lenghts, then that's what goes into the formula; similarly if it's 9 lenghts, then that's what goes into the formula.

So the finishing time, given by the formula, never errs.

[kenwoodallpromos]

"*Lengths behind is prone to human error unless
it is recorded from a frozen video frame"
Ain't it the truth!
Just like across the pond, PP's do not give times from video tapes; they want us to calculate it! You guys do not even have fractionsl times!
You can call %E % velocity or visible result of actual energy expended or anything else you want.
Can you information be used without access to full electronic timing or hand timing? Do you time off of video tapes by hand? How do you get fractional time information?
Is there any way to get workout (training) time information?
In the US decades ago the only way to get such information was to be present for workouts and races and hand time them; So those people had a big advantage.
Even though I am no good at certain math, your information is presented in a very easy-to-read table.

[BillW]

Originally Posted by speculus

Gosavi's Formula
for Accurate Timing

Attached is the file Speculus has formatted here. Slight improvement in format, some may find it easier to read. Probably most easily read in wordpad or something similar. Let me know if there is any problem (I don't run windows and I'm doing this blind).

Bill

[chickenhead]

Originally Posted by speculus

You are wrong when you say "All horses would have to be running at the same velocity to be able to convert distance to time so accurately using the same constant".

I think you are confusing between the static and the dynamic view of the finish.

Suppose you have access to a still photograph of the finish when the winner is crossing the wire and the runner up is, say, 8 lengths behind. Now depending upon whether the runner up is slowing down or accelerating (of course with respect to the winner, make no mistake about that!), the actual verdict of lengths (that will go down into the charts or record books) between the two, would be more or less than 8 lenghts.

Why? Because the final verdict is decided by the image of the runner up on the photo finish strip, and it is nothing but the exact instance when the runner up crosses the wire.

If it is 7 lenghts, then that's what goes into the formula; similarly if it's 9 lenghts, then that's what goes into the formula.

So the finishing time, given by the formula, never errs.

aha! I did assume you were talking a static view, as you were talking about lengths as if they described distance rather than time (you claimed a length was equal to 2.8 meters). If you agree that a length has no meaning with regards to a distance(at the finish), we are in agreement.

I still claim foul on your claim of accuracy, even if using charts you will get round off errors that will put you further off than you claim. I also do not believe that your constant should change based on the length of the race, the photo finish camera does not change, and therefore the time increment x number of lengths represents does not change either.

[DJofSD]

I think you are confusing between the static and the dynamic view of the finish.

Yes, the difference between assuming a constant velocity v. a model that takes into account delta velocity over delta time -- acceleration.

Of coarse, all of this tweaking would be totally unneeded if we had electronic transmitters embedded in the nose of the horse to allow tracking of all 4 vectors.

Hell, combine that with the radio identification chip and then we can go onto the next problem of getting horses weighed.

DJofSD

[speculus]

Originally Posted by kenwoodallpromos

Just like across the pond, PP's do not give times from video tapes; they want us to calculate it! You guys do not even have fractionsl times!kenwoodallpromos] You can call %E % velocity or visible result of actual energy expended or anything else you want. Can you information be used without access to full electronic timing or hand timing? Do you time off of video tapes by hand? How do you get fractional time information?

We have fractional times, but they are hand-clocked by time-keepers of various private race books/sheets, and we have a number of them at each track. I however, get my fractions from a chinese vieo software which allowes me to view the race frame by frame if I need.

Originally Posted by kenwoodallpromos

Is there any way to get workout (training) time information? In the US decades ago the only way to get such information was to be present for workouts and races and hand time them; So those people had a big advantage. Even though I am no good at certain math, your information is presented in a very easy-to-read table.

The professionals who collect data are called track reporters at our place. Unfortunately, most of them are simply duds, a very few know what to look for in a horse to spot improvement. The better ones, obviously, find employment with reputed race books/sheets, and some of them know their job, are very good at it, and at a price, "discuss" some of their so-called secret findings only with big punters who pay them handsome money for that kind of information.

Personally, I enter all the track work data in my excel file, set auto filters, and use by sorting it on horse_name when I study the card. But I don't fancy myself as a good track-work reader from recorded data.

[speculus]

Originally Posted by chickenhead

I still claim foul on your claim of accuracy, even if using charts you will get round off errors that will put you further off than you claim. I also do not believe that your constant should change based on the length of the race, the photo finish camera does not change, and therefore the time increment x number of lengths represents does not change either.

The proof of the pudding is in the eating. Why don't you apply the formula and check for yourself?

Or even better, PM me your address, and I will send you a CD of race video with time stamped onto it (in 100th of a second). Play it in slow motion or freeze it at the finish, and check the truthfulness of the formula.