Speed Handicapping

[mr forecast]

actuary had a blinding day with his figures at wolverhampton

with a total of 22 horse which if you would of bet all of them e/w would of returned just over 84 points….with a 44 point outlay

my speed figures (top rated produced 4 winners totaling 50 points with a 16 point …..8 bets e/w outlay ….so 34 points

one forecast were the 1st and 2nd rated came 1st and 2nd paid 20.93, but been a while since a forecast

[Gerald]

I accidentally came across one of Actuary’s videos on Youtube when I was searching for race videos. Couldn’t help noticing his revolutionary new speed handicapping method is full of 99%s and 98%s. Was wondering whether it was similar to this method of Carroll’s that I outlined:-

I don’t know whether this thread has been prompted by Carroll’s book with the same title.

I left all my books behind in Harlesden, so this might be wrong in places.

First of all you have to get your par or standard times. Carroll bets on Quarter Horses, in Texas and New Mexico, I think. He got his par times by looking at the world record for various distances, and then drawing a line where the world records ought to be for the odd ones out.

He equates this time with 100%.

I think he doesn’t bother with going allowances, which is fair enough in New Mexico.

The time each horse runs is given as a percentage.

So if the world record is 60 seconds, and a horse runs in 61 seconds, this is a rating of 100*60/61 = 98.361

(Personally, my head would get done in by all the 9s and 8s that appear in the ratings of horses, and I would set my pars to 111 instead of 100, in order to reduce the number of them.)

So say you have your set of par times. Now the thing is, for the past few years we have known the times taken by the beaten horses, but at the expense of knowing what the actual distance gap was. This actually is a boon to speed handicappers.

On the Flat (Turf) on Good ground it is 6 lengths per second, according to BHA diktat. So if a horse is beaten by a length, it crossed the line 0.167 seconds after the winner, and this is without knowing how long a length is, and with a length being a different distance according to the speed that the horses are travelling at. So, the actual speed that the horses cross the line gets factored out as well. All we have is 0.167 seconds, and that is all we actually need.

Let us say that the pars for the two distances of 5f and 10f under consideration are 58 seconds, and 121 seconds. Also assume that the winners actually won in par. The horses beaten a length must have run in 58.167 and 121.167 seconds. The two winners both achieve a rating of 100. The horses beaten a length have ratings of 100*58/58.167 and 100*121/121.167, which are 99.713 and 99.862 respectively.

[mr forecast]

i think actuary is spot on by not giving the speed ratng of a horse, just its percentage

several horses can have the same rating but different percentages

one question i would like to ask, should your pars be produced from record times or the racing post standard times,

some course records are over three years old so i am very wary of this,obviuoly the two throw up completely different figures

i personaly believe that actuarys site is the best aw speed related website i have found, and its still free

formbet was up there but charges now

[Gerald]

I’ll actually go and visit Actuary’s website over the weekend, for the first time. (Glad I wrote that and didn’t have to say it!)

Yes Mr Forecast, it would probably be a bad idea to use World records. Carroll was just using interpolation at shorter distances, for races run on a straight track (presumably). Once one is racing around variously shaped bends and uphill and downdale using the World Record method is out of the window.

It is very difficult to produce standard or par times for turf courses, as they necessarily only have a few races each year. I think Actuary concentrates on the All-Weather, which makes it easier (to produce the figures). He’s operating in a very competitive market though, as the AW is where most of the time merchants operate, like yourself .

[mr forecast]

i do struggle to understand how you get a figure of above 100% if you are dividing by 100?

my top three rated for the first at southwell were before the race

russian invader 99.91 98.05 98.05
stagecoach amber 99.25 97.08
Merrion Tiger 99.72

and the figures for todays race

Russian Invader 1m6f 188.76 2 0.33 189.09 99.82
Stagecoach Emerald 0 b 7 09/03/09 52 67 73 1m6f 188.76 2.25 0.38 189.14 99.8
Merrion Tiger 28J 4 09/01/09 50 53 72 1m6f 188.76 0 0
188.76 100

this just using the race times and are not the course record vs finishing time figures

[Gerald]

I won’t assume too much what Actuary is doing, because I don’t have the time to work it out – just to say that it is easy to get figures greater than 100% if the par time is some kind of median figure.

[mr forecast]

time consuming is an understatement,

[mr forecast]

i have been trying to improve on the ratio of 0.167 secs per length

i have asked loads of people, beyer, actuary,bhb and they all have a different answer

i am just testing this but if you take the course finishing time and divide this by the set of numbers at the end of each courses distance it gives you a slightly better figure( the red figure)

considering the factor of

furlong times 220 divide by 0.9144 divide by 3.5

220yards to a furlong
0.9144 meters to a yard
3.5m a length (open to conflict)

lingfield
5f 58.29 343.7[/color:27pzf82n]
6f 69.99 412.44
7f 83.68 481.18
1m 96.38 549.93
1m2f 121.63 687.41
1m3f 135.1 756.15
1m4f 148.1 824.89
1m5f 161.08 893.63
2m 200.07 1099.86

JUMP RECORDS
2m 212.7 1099.86
2m2f 243.7 1237.34
2m4f 275.3 1374.82
2m6f 299.7 1512.31
3m 326.7 1649.79

wolverhampton
5f 61.4 343.7
5f20y 61.13 349.95
5f216y 72.61 411.19
6f 73.8 412.44
7f 86.8 481.18
7f32y 86.65 491.18
1m1f 109.5 618.67
1m100y 108.2 581.17
1m141y 108.08 593.98
1m1f79y 120.02 643.35
1m1f103y 120.76 650.85
1m2f 123.8 687.41
1m4f 154.1 824.89
1m4f50y 154.75 840.51
1m5f194y 179.85 954.25
1m6f166y 191.4 998.62
2m46y 216.6 1114.23
2m119y 215.85 1137.04

southwell
5f 57.14 343.7
6f 73.5 412.44
7f 86.8 481.18
1m 98 549.93
1m2f 119.75 687.41
1m3f 141.5 756.15
1m4f 153.9 824.89
1m5f 172.5 893.63
1m6f 181.6 962.37
2m 217.6 1099.86
2m2f 245.1 1237.34

kempton
5f 58.96 343.7
6f 71.31 412.44
7f 84.73 481.18
1m 96.82 549.93
1m1f 110.1 618.67
1m2f 123.77 687.41
1m3f 137.74 756.15
1m4f 150.48 824.89
2m 205.83 1099.86

[TheBluesBrother]

Interesting stuff:

Everytime I try something like this I end up with a headache

[mr forecast]

the other problem i am finding is that you end up with a list of horses with all ratings of 100 as they won, now when i look at the speed database i have no way of knowing by how many lengths each horse won by

so again i am testing this method,

i have two ratings per horse the first rating goes from 100 and descends depending on the lengths beaten in a perticular race, then the main figure which is the winning time vs course record

i dont touch the course time and course record figure but with the other figure i give the winner a minus figure in lengths so when you look at the figure it shows a percentage above 100 which indicates the lengths the horse won by, sounds complicated but if i have a race with four recent winners in it the figures give you a better indication of ability, the difference being winning by a nose or 20 lengths

Batchworth Blaise 101.8 99.26[/color:110ggdf3]
Saute 101.03 98.76[/color:110ggdf3]
Inflammable 101.17 96.67[/color:110ggdf3]
Shadows Lengthen 101[/color:110ggdf3].17 98.71[/color:110ggdf3]

it sounds a weird way of doing it but makes sence to me,
the idea is that the first figure should show above 100 if the horse has stormed home, and the second figure should only be above 100 if the horse has beaten the course record time

this will need to be kept seperate than the normal figures database as putting aminus lengths beaten will give a false course winning time, so i keep two databases going whilst i test these mad ideas

[TheBluesBrother]

I tackled the problem by developing class pars for every race grade.
Looking at a whole years results from every racecourse, I came up with these pars which were developed using pace + class for each grade; took about 2 months to complete.
(C7) 52, (C6) 60, (C5) 68, (C4) 80, (C3) 92, (C2) 96 (C1) 100, (G3) 106, (G2) 112, (G1) 120.

Taking the first horse you had as an example, this how I way up its chances.

Batchworth Blaise 09-May-09 Lin 7GF C5Hc 3K 49 110.23%
Batchworth Blaise 10-Jan-09 Kem 7St C6Hc 2K 48 107.95%
Batchworth Blaise 25-May-09 Goo 8GF C6Hc 3K 47 106.82%

Batchworth Blaise top three class ratings are 49, 48 and 47 and this shows me that the horse’s true ability is Class 7 (see class pars above).
Looking at his OR % against his top class figure of 49 shows 110.23%, which indicates that he is well handicapped against his official BHA Official Rating.
So the horse is approx 5lb well in on his best form which equates to 2.5 lengths over 7 furlongs.

5f = 3.5lb length
6f = 3.0
7f = 2.5
8f = 2.0 etc.

Saute 05-Dec-09 Sth 14St C6Hc 1K 72 124.14%
Saute 11-Nov-09 Sth 14St C6Hc 2K 62 106.90%
Saute 26-Oct-09 Kem 12St C63yHc 2K 60 102.59%

Salute best rating was 72; I would ignore this, as his next best were 62 and 60, so take 62 as the standard, his true class would then be Class 6.
He is 4lb well in which equates to 3.2 lengths (1.25lb per length over 14 Furlongs)

Inflammable 05-Dec-09 Wol 12St C5Md 2K 60 91.54%
Inflammable 21-Oct-09 Sth 11St C53yHc 3K 52 79.23%
Inflammable 24-Nov-09 Sth 12St C5Md 2K 50 76.92%

Inflammable best class ratings is 60 and looking at his OR % of 91.54% he not well handicapped (Needs to be 100% +).
If Inflammable was to run in a (C7) race I would be interested.

Shadows Lengthen 08-Dec-09 Sth 11St C6Hc 2K 78 100.65%
Shadows Lengthen 05-Dec-09 Sth 11St C5Hc 2K 72 93.51%
Shadows Lengthen 17-Nov-09 Sth 12St C6Hc 2K 69 89.61%

Looking at the form shows Shadows Lengthen to be a class 4 horse (i.e. 78) and the handicapper is not showing any mercy by lumping pounds on his back.

Finally looking at Batchworth Blaise best rating of 49 and OR % of 110.23% it is possible for him to run off
a rating of 54 if producing his best form i.e.((49 x 110.23%)/100)=54.01)…just food for thought.

[mr forecast]

its good to see the different angels everyone takes

i have been watching yr work for a while blues, always updating the new figures, but never had time to endulge more

try not to get to side tracked ,doing just the speed figures is enought for now

ill keep watching thou

[Author]

Posts