Effect of weight on horse speed

[cosmicsports]

In my horse racing speed calculations I use the following scale for the effect of weight on horse speed:

5 furlongs = 0.02 seconds / kg
5 1/2 furlongs =0.03 seconds / kg
6 furlongs = 0.03 seconds / kg
6 1/2 furlongs = 0.04 seconds / kg
7 furlongs = 0.04 seconds / kg
8 furlongs = 0.05 seconds / kg
8 1/2 furlongs = 0.06 seconds / kg
9 furlongs = 0.06 seconds / kg
10 1/2 furlongs = 0.08 seconds / kg
12 furlongs = 0.09 seconds / kg
15 furlongs = 0.17 seconds / kg

Someone gave me those figures a long time ago and I sort of believe them ever since.
In fact it was an American trainer with whom I was chatting in a forum.

But can they verified or has someone else a more exact scale ?
What baffles me is I have a database of 6000 horses, those horses have raced from once to 200 times and I can find nothing of the sort.
I tried least squares, least squares with outliers excluded.
I tried to compare raw speeds or raw speeds plus track variant, nothing of significance comes out !
Web search does n't help either on this topic.
Can you help ? Can you perhaps suggest foreign keywords to see if there is something in the other languages ?

A certain racing magazine I have here suggests 0.20 seconds / kg for all distances.
That's gross though is n't it ?
I asked the editor and he said to me "my oh my it's only logic can't you understand ? 1 kilogram = 1 length".
I thing he 's crazy. If it was like that then one horse would finish at Picadilly circus and the next would be in Manchester.

Yet some others are saying "neglect weight - it has no effect". I think this can't be true either.

 

[cosmicsports]

Mordin says 0.2 secs per Kg per mile which sounds more reasonable than 0,2 secs per Kg for all distances.
But how can this result be derived ?

 

[Outlander]

At 0.02 secs per kg over 5 furlongs this would mean 50kg per second or 110 lb per second at a ball park average a horse is not going to run more than about 6 lengths in 1 second even on fast ground .
So 110/6 = 18.33 lbs per length.
Most handicappers would give between 3 and 3.5 lbs/length over 5 furlong , even if you stretch it most wouldn’t give more than 4lbs/length
18.33 lbs per length clearly coming from someone who believes weight has very little impact on time.
I came to the conclusion myself that weight doesn’t really have a massive effect on time so maybe this American trainer isn’t so crazy, I did a study for a few months last year and there was very little difference in the profitability of ratings that allowed for weight in a conventional manner and those that completely ignored weight given all other factors the same.
I came to the conclusion the effect of weight was not in line with conventional thinking or linear, and that as a factor class was many times more important than weight.
When doing Hong Kong ratings I have a time : weight ratio much bigger than anyone has suggested in compiling speed figures texts I have read and I still find if anything the rating have a bias to lower weighted runners.

 

[TheBluesBrother]

Most handicappers would give between 3 and 3.5 lbs/length over 5 furlong

As you know I use different lbs per length due to the different racecourse configurations, I divide the standard time for any distance into a constant of 200.

Epsom 5f - 54.00s (3.70 lbs per length)
Pontefract 5f3y - 61.50s (3.25 lbs per length)

I came to the conclusion the effect of weight was not in line with conventional thinking or linear

The old adage states that 1lb equals 1 length, is simply not true.

Years ago I put a maximum weight cap of 10-0 when calculating my speed figures, so if a horse carries 11-6 on the flat in an amateur race I calculate it as carrying only 10-0, this was bourne home to me years ago when an amateur race run on Derby Day, the winner carrying about 11-6 had a faster speed figure than the winner of the Derby, weights over 10-0 are certainly not linear.

I have read and I still find if anything the rating have a bias to lower weighted runners.

Like I have already mentioned, I use a cut off weight when rating of 8-0, it's not so bad these days due to the jockey's weight rise allowances, back in the days of the lightweight jockeys carrying 6-0 it was a nightmare, the reverse effect of the weight carrying amateur races.

Throw in the subject of WFA into calculations and you are up shit creek without a paddle, I have no doubt that Admiral Rous was a very clever man, he made an incorrect assumption that horses acquire stamina with age.

In North America, the speed boys have it easy, the majority of the racecourses are flat and uniformed.

Mike.

 

[Outlander]

got these from book Bioenergetics and Racehorse Ratings , his calculations for standard ground that is not aiding or hindering speed, I converted to the sec/kg so you could compare to the scale you have from the American trainer, Bob Wilkins who wrote the book calculated with his scale much less value on weight than I have read elsewhere on the subject but still massively more than your scale

furlongs	lbs/sec	kg/sec	sec/kg
5​	29.94​	13.58054​	0.073635​
6​	23.91​	10.84538​	0.092205​
7​	19.89​	9.021945​	0.110841​
8​	17.03​	7.724672​	0.129455​
10​	13.26​	6.01463​	0.166261​
12​	10.87​	4.930545​	0.202817​
14​	9.21​	4.177582​	0.239373​
16​	7.99​	3.6242​	0.275923​

 

[Outlander]

TheBluesBrother has a very important point the above are calculated on a standard 5F in his book he equates that to 58.68 seconds, so TBB is correct at the likes of Epsom the number would be lower sec/kg than in the above table as it takes much less time than 58.68

 

[cosmicsports]

At 0.02 secs per kg over 5 furlongs this would mean 50kg per second or 110 lb per second at a ball park average a horse is not going to run more than about 6 lengths in 1 second even on fast ground .
So 110/6 = 18.33 lbs per length.
Most handicappers would give between 3 and 3.5 lbs/length over 5 furlong , even if you stretch it most wouldn’t give more than 4lbs/length
18.33 lbs per length clearly coming from someone who believes weight has very little impact on time.
I came to the conclusion myself that weight doesn’t really have a massive effect on time so maybe this American trainer isn’t so crazy, I did a study for a few months last year and there was very little difference in the profitability of ratings that allowed for weight in a conventional manner and those that completely ignored weight given all other factors the same.
I came to the conclusion the effect of weight was not in line with conventional thinking or linear, and that as a factor class was many times more important than weight.
When doing Hong Kong ratings I have a time : weight ratio much bigger than anyone has suggested in compiling speed figures texts I have read and I still find if anything the rating have a bias to lower weighted runners.

For an additional weight of 3.5 lbs (= 3.5 x 0.454 = 1.589 kg) the horse loses a length (0.17 secs) over 5 furlongs.
That's Mordin's figure for 5 furlongs.
He says 2.4 lbs per length per mile to be precise and I take it one length is 17 seconds.

Here:

Mordin on Time

In Mordin On Time, Nick Mordin sets out his method for answering the most fundamental question facing punters in any race, namely: which is the fastest horse? He was timing the sections of races with a stop watch, estimating wind strength and direction, adjusting for movements of running rails...

books.google.gr

I have the book too, but not here with me.

But it's above what the friend from America said, 0.02 sec /kg. So he all but neglects weight at the short distance.

The question here is how to derive this.
My database contains 5809 horses.
For each horse, for each appearance I have total speed from start to finish, weight carried plus the track variant (a good estimate of
track variant).
So for 5 f some 12500 cases.

I do this:
I use as datum the 58 kg (127.8 lbs).
Then I use a formula like

corrected time = official time + k . (58 - weight)

Then I try to minimize the variance of the 12500 readings but it does n't work.
Is there something wrong with this approach ? How else ?

 

[cosmicsports]

furlongs	lbs/sec	kg/sec	sec/kg
5​	29.94​	13.58054​	0.073635​
6​	23.91​	10.84538​	0.092205​
7​	19.89​	9.021945​	0.110841​
8​	17.03​	7.724672​	0.129455​
10​	13.26​	6.01463​	0.166261​
12​	10.87​	4.930545​	0.202817​
14​	9.21​	4.177582​	0.239373​
16​	7.99​	3.6242​	0.275923​

I don't know the name of this American fellow I 'm talking about.
He was writing in a forum as Jeff if I remember rightly and he said he was a trainer.

Now those figures above look ok too (a bit below Mordin).
But how to derive ?
Also I don't think the race course standard time matters.
Why should it ? If I get tired by carrying a big gun in Greece, I get tired by the same amount in Spain.
But the type of surface maybe matters (dirt-artificial-grass).

 

[Outlander]

I don't know the name of this American fellow I 'm talking about.
He was writing in a forum as Jeff if I remember rightly and he said he was a trainer.

Now those figures above look ok too (a bit below Mordin).
But how to derive ?
Also I don't think the race course standard time matters.
Why should it ? If I get tired by carrying a big gun in Greece, I get tired by the same amount in Spain.
But the type of surface maybe matters (dirt-artificial-grass).

I have to many unknowns to help you , I don't know what units your track variant is in for a start this needs to be converted to time and added/subtracted from time to get corrected time.
Also even if you get your corrected time calculations correct this will only be of any use when comparing the times at the same distance on the same track, if you want to compare between variations of distances and tracks then having standard times for every track/distance it a vital element.

 

[cosmicsports]

I have to many unknowns to help you , I don't know what units your track variant is in for a start this needs to be converted to time and added/subtracted from time to get corrected time.
Also even if you get your corrected time calculations correct this will only be of any use when comparing the times at the same distance on the same track, if you want to compare between variations of distances and tracks then having standard times for every track/distance it a vital element.

Track variant is in seconds, like RP more or less.
There are many ways of assigning a track varian to a race - I have the best version known so far in my data.

Example over 6f in one case it says time = 1.14.06, weight = 58 kg, t.v. = 2.30, position finished 4th and further down the line winner's time = 1.12.53, winner's weight = 60 kg.
This 2.30 track variant effectively means that you add it. So in the imaginary standard surface conditions for the said race course, the time of our example horse would be 1.16.36 and the winner's time 1.14.83 (2.30 looks big - they did n't have enough money to buy sand in the summer !).

It's one track, not many tracks.
By standard times what exactly do you mean. Maybe that's something I can compute.

 

[cosmicsports]

The standard times are needed for track to track conversions I think, when there is shipping of horses from one place to another.

 

[gbettle]

A nice profitable NH Handicap weight system I've seen is from Steve Jones, where he thought a top-weight horse on a tight run track - think at the most extreme, Fakenham - could negate that weight penalty:

NH Track Matrix

 

[Chesham]

A nice profitable NH Handicap weight system I've seen is from Steve Jones, where he thought a top-weight horse on a tight run track - think at the most extreme, Fakenham - could negate that weight penalty:

View attachment 105909

NH Track Matrix

Quick way to the poor house

 

[gbettle]

Quick way to the poor house

View attachment 105910

Many thanks for your reply Chesham - I stand corrected.

It could be so bad that it maybe now a Lay system ...

 

[Outlander]

Track variant is in seconds, like RP more or less.
There are many ways of assigning a track varian to a race - I have the best version known so far in my data.

Example over 6f in one case it says time = 1.14.06, weight = 58 kg, t.v. = 2.30, position finished 4th and further down the line winner's time = 1.12.53, winner's weight = 60 kg.
This 2.30 track variant effectively means that you add it. So in the imaginary standard surface conditions for the said race course, the time of our example horse would be 1.16.36 and the winner's time 1.14.83 (2.30 looks big - they did n't have enough money to buy sand in the summer !).

It's one track, not many tracks.
By standard times what exactly do you mean. Maybe that's something I can compute.

I have never come across a track variant expressed in seconds, usually it is either seconds per furlong , seconds per mile or pounds or kg depending on what part of the process you account for it.
But taking your variant at face value the horse that finished fourth in you example is correct at 1.16.36.
The winner would be 1.14.83 less (60-58)x0.03 sec/kg as per your scale in post#1
1.14.83 is 74.83 seconds - 0.06 = 74.77 corrected time.
If your variants are accurate and your weight conversion scale is accurate you will know what horse is fastest at a particular distance.
But if you were attempting to line up horses in a future race that might have past runs over 5f , 5 1/2f and 6f. You will not be able to line them up in order of speed unless you have a standard time for each separate distance.
The good news is if you corrected times are accurate you can easily find a standard time for each distance , just line up your times from fastest to slowest and take the 15th percentile of each list of times at each distance.
Once you have these standard times work out how much your corrected time is faster or slower than the standard and diivide by the distance to get a seconds per furlongs time and this can be used to compare directly all times at different distances.

 

[davejb]

Applying deep thought to the issue, although my wife thinks I just look constipated, I've concluded that weight has less effect than we tend to allow for, and more to the point its effect is all but unpredictable because horses have jockeys.

My opinion is that the same horse will produce very different results from running over a given distance carrying a given weight entirely due to the fact there is a chap on his back pulling him back, scrubbing him along, or banging on him with a whip. We know, or seem to think at least, that a horse that is maybe going at a trip they only just can get will have much more chance of doing so if either it gets good cover in the pack until a couple of furlongs from the line or if the only runner to go forward to set the pace does so at a leisurely pace - reducing our (let's say) 10f race to what is actually a 6f jog followed by a 2f gallop that ends with a 2f sprint. The tactics of the jockey and the tactics of all the other jockeys (intentional or not, I am sure sometimes that somebody finds himself leading the race when they'd much rather be hiding midfield) affect when a horse has to go at or close to his maximum speed.

My opinion, as I say, but it seems to me that what we really have to play with are a range of factors - the first is speed: how fast can the horse run? What is its maximum speed? How fast can it run, for how far, if there's little weight on board? Next we have weight: If we know a horse can actually run at x mph, what is the maximum weight it can carry before that top speed becomes impossible to achieve, or the time it can maintain that speed is severely degraded? Third we have energy: To me this is what it really boils down to, how much energy does the horse have and what fraction of it can it release in a race? If a horse has say 1000 units of race energy (just invented) that it can expend in a race, and it runs at a steady pace that doesn't really push it for say 8f before going as fast as it can for the final 2f of our 10f race, then maybe it uses 600 units in the first 8f and has 400 still in the tank for the final quarter mile - but if he's been leading the race and trying to run the finish out of the rest of the field then he's going to use up 800 units by the 2f marker and only have 200 to finish the race on. We talk about 'the tank running dry' and similar when some runner has little to give at the finish, which is when somebody taking life easy in the pack sweeps past having conserved energy better.

Weight will have an energy cost, but I think it is quite likely that if a horse is given a ride where it stays in the leading half dozen but with a little cover, maybe sitting in 4th or 5th, then the jockey can be said to be managing its energy quite well. In this situation I suspect the effect of weight is minimised - the horse is not being stretched, so it can cope with the weight it carries. For the final 2f or so it has a reserve of energy, and is better able to run nearer to its top speed whilst bearing significant weight. A horse that has not had an energy conserving ride, that has had to put more effort into that first mile, has also found the weight burden taxing its energy reserve - it has less in the tank for the final 2f, and may struggle to maintain speed let alone increase it, it may also find the weight very taxing. We seem to have general agreement that weight has little effect over 5f - and that the effect increases with distance - which I think is logical, provided we aren't talking ludicrous amounts (like shoving a 20 stone jockey up) it seems to me that the average sprinter can lump almost anything for a few furlongs, by which point they know the finish is just ahead.

Depending on energy management, in my view, weight will have a range of effects from (possibly) very little to significant. The energy management is down to the jockey, but also depends on the rest of the runners allowing it, and of course a good few horses simply don't perform unless they run the type of race they seem to prefer - some always front run, many a race is hampered by there being no 'natural' front runners to set the pace, it's a minefield!

It's almost certain, I think, that every horse has a different weight carrying ability, which ranges from 'doesn't bother me much' to 'blood and sand, have you put an elephant on my back?'

I don't think it's possible to actually put figures on this accurately - all we can do is adopt our best guess and figure that we probably can't improve much on it.

Like O Outlander I see class having as much of a bearing - not always, or life would be too easy - but a horse can drop in class and we see it as a chance to win, rather than notice the inevitable weight increase involved.

Dabe

 

[cosmicsports]

I have never come across a track variant expressed in seconds, usually it is either seconds per furlong , seconds per mile or pounds or kg depending on what part of the process you account for it.
But taking your variant at face value the horse that finished fourth in you example is correct at 1.16.36.
The winner would be 1.14.83 less (60-58)x0.03 sec/kg as per your scale in post#1
1.14.83 is 74.83 seconds - 0.06 = 74.77 corrected time.
If your variants are accurate and your weight conversion scale is accurate you will know what horse is fastest at a particular distance.
But if you were attempting to line up horses in a future race that might have past runs over 5f , 5 1/2f and 6f. You will not be able to line them up in order of speed unless you have a standard time for each separate distance.
The good news is if you corrected times are accurate you can easily find a standard time for each distance , just line up your times from fastest to slowest and take the 15th percentile of each list of times at each distance.
Once you have these standard times work out how much your corrected time is faster or slower than the standard and diivide by the distance to get a seconds per furlongs time and this can be used to compare directly all times at different distances.

As we talk about the 6f races, our fastest ever was Lieiz, a French raider, 1.10.09 in October 2004.
It says kg = 51.5, track variant = 0.40.
The slowest ... must be some Thames barge somewhere in the data sheet. About 1.18.00 probably.

So you want me to sort all the 6f times or just the winners times ?
There are many but It's not difficult using the Shell-Metzner sorting algorithm.
Then what ?
Take the 15th percentile out of how many equipopulated percentiles ?
Or you mean 15% down the ladder ?
And why the 15th ?

Anyway suppose that comes out as 1.14.00.
Then in one case suppose a horse's time was 1.13.50, track variant corrected.
So for this case the figure is 0.5 / 6 =1 / 12.
For all cases suppose the mean is again 1 / 12.
How is that related to kilos ?
It looks to me like it is related to mixed abilities of the runners.

 

[TheBluesBrother]

Take the 15th percentile out of how many equipopulated percentiles ?

Here is a tutorial I put up regarding standard times using the 15th percentile etc.

Calculating standard times using Z scores

Calculating standard times is an arduous task, due to the excessive rail movements on British racecourses, plus various other practices, for example the official race distances in Ireland! Using Excel, I am going to work out the standard time for the 5f (1000m) at Deauville using the fastest 25...

www.theukbettingforum.co.uk

Mike.

 

[Chesham]

@davejb

 

[Outlander]

As we talk about the 6f races, our fastest ever was Lieiz, a French raider, 1.10.09 in October 2004.
It says kg = 51.5, track variant = 0.40.
The slowest ... must be some Thames barge somewhere in the data sheet. About 1.18.00 probably.

So you want me to sort all the 6f times or just the winners times ?
There are many but It's not difficult using the Shell-Metzner sorting algorithm.
Then what ?
Take the 15th percentile out of how many equipopulated percentiles ?
Or you mean 15% down the ladder ?
And why the 15th ?

Anyway suppose that comes out as 1.14.00.
Then in one case suppose a horse's time was 1.13.50, track variant corrected.
So for this case the figure is 0.5 / 6 =1 / 12.
For all cases suppose the mean is again 1 / 12.
How is that related to kilos ?
It looks to me like it is related to mixed abilities of the runners.

Yes use all times not just winners(remember to use the corrected times adjusted for the weight and the track variant as already discussed how to do this)
15th percentile down the ladder
why 15th …just experience tells me this is as good as any….no correct answer….the median I believe would include too many slow times, 15th is a sensible representation and cuts out ant outliers at the top

I thought you just wanted to line up the times, if you are attempting to create usable ratings to adjust for future races, then it’s very difficult to give full advice…I don’t have the data you have.
Im guessing we are dealing with a country I know nothing of with just one racetrack, I don’t know where you obtain your track variant, the track variant could be useless. Creating an accurate track variant is essential.
In a days results if the track variant is 2,30 seconds is this the same variant for 6 furlong races and 7 furlong races, if so that makes no sense.
Does your data contain official ratings for the horses or any indication of class of runners in each race.
When you find out that a horse has run a corrected time of 1.13.5 and your standard is 1.14
You know this horse is 1/12 sec/furlong fast. You need to multiply this by a constant that will give your horse a rating that will then line up with kg so you can adjust for future weights in future races.
Again there is no correct answer…for U.K. I ended up with a constant of 126 to convert to a pounds friendly scale.
I would suggest you use 58 as the constant.
so in our example the horse is 1/12 sec/fur fast so 1/12 x 58 so the horse is 4.83 kg fast
So we could call this horse 104.83 speed rating
Do this for every horse, then adjust for kg carried in future races
The track variant will be vital, it will be a case of if your variant is accurate your ratings will be accurate.