- This topic has 101 replies, 8 voices, and was last updated 20 years, 10 months ago by
.
- Posts
Wallace,<br>In answer to one of your questions, I have outlined below a fairly basic method of compiling speed figures. It isn’t simple(no method is), but it could be adapted by anyone to suit their individual preferences. I am hoping it will draw some comment from other compilers.
It is based on weight affecting a horse’s performance, which most people accept, but not everyone is in agreement with how this is quantified.
The basic method that I would to use to compile speed ratings is as follows.<br>Flat Racing
1.Identify the standard (on a scale of 0-140) of each race at a particular meeting. This can be done by taking the RPR top-rating and subtracting 14- the resulting figure is the standard of the race adjusted to 9st. (The RPR ratings are adjusted to 10st ). <br>This is the EXPECTED standard of form and time for a race run at a fair pace on good ground.
<br>For 2yo and 3yo races, an adjustment must be made for weight for age using the agreed scale, preferably the one recommended by the Racing Post and published in the paper from time to time. The adjustment involves reducing the standard by the weight for age allowance to recognise the fact that an immature horse will not be able to run as fast as a mature horse even though they have the same RPR or Topspeed rating.<br>The RPR and Topspeed ratings as they appear in the Racing Post are already adjusted upwards for weight for age in races for 2yo and 3yo only. In all aged races including handicaps, the WFA allowances (BHB or official version) are included in the allotted weights.
2. Note the ACTUAL time of each race and whether or not it is above or below the standard time. The standard time has been calculated from a large sample of races and is defined as:
‘the time expected from a mature horse (4.5 year old), rated 100, carrying 9st on good going’.
A mature horse rated 100 is a top handicapper or listed class.
The ACTUAL race time can be converted into a rating using the following table of distance/time/weight, used in conjunction with the weight for age allowance.
Distance(f) 1 second = (lbs) <br> 5 22 <br> 618.4<br> 715.8<br> 8 13.7<br> 912.3<br> 1010.9<br> 1110.1<br> 12 9.2<br> 138.5<br> 147.8<br> 157.3<br> 166.9<br> 176.5<br> 18 6.1
For example, at 5f, 1.5 seconds slower than standard time is (22 x 1.5 = 33lbs) below the standard of 100, i.e. 67.
Provided a race is run at a fair pace, the difference between the EXPECTED time and the ACTUAL time is due to the going unless there are any other factors which are not apparent. These might include the effect of the wind, incorrect distances due to unannounced track changes or faulty timing(unlikely).
The average difference or variation for races reportedly run at a fair pace is the going correction which needs to be applied in order to calculate speed ratings. The going correction might be different on the straight part of the course to the round course and might also change during the racing if there is rain, a drying wind or the ground cuts up due to usage.
The GOING CORRECTION is expressed as a figure in lbs which recognises that horses’ times, on average, are being affected by the going.<br>If the average variation is negative, it means horses are being slowed and if it is positive it means they are going quicker than the standard.<br>A negative variation requires a POSITIVE correction; a positive variation warrants a NEGATIVE correction. This often causes confusion, but I prefer to express the going correction as negative when it is slowing horses and positive when it is making them go faster than standard, even though the adjustment to the basic race rating needs to be in the opposite direction.
3.The going correction is then applied to the ACTUAL race times which were earlier converted into ratings using the table of distance/time/weight. For example, if the going correction is -10, then 10lbs needs to be added to the rating for each race to obtain a rating which has been adjusted for the going.
4. The rating for the WINNER needs to be adjusted for weight carried (above or below 9st), WFA and any rider’s allowance.<br>Rating from (3) above<br>ADD weight carried above 9st<br>ADD back any WFA allowance<br>ADD back any rider’s allowance<br>DEDUCT weight carried below 9st
5. For The rest of the runners:<br>Rating from (3) above<br>ADD weight carried above 9st<br>ADD back any WFA allowance<br>ADD back any rider’s allowance<br>DEDUCT weight carried below 9st<br>DEDUCT distance beaten using the following table:<br>At 5f 1 length = 3.6lbs<br> 6f 1 length = 3.0lbs<br> 7f 1 length = 2.6lbs<br> 8f 1length = 2.3 lbs<br> 9/10f 1 length = 2.0 lbs<br> 11/12f 1 length = 1.6 lbs<br> 13/14f 1 length = 1.4 lbs<br> 15/16f 1 length = 1.2 lbs<br> 17/18f 1 length = 1.1 lbs
The calculations can be somewhat cumbersome, but a lot of the sub-rotines can be programmed on to a spread sheet.
As an example, I am doing Sandown on Saturday. You will probably need a copy of the Racing Post to follow my workings.
Aye!
Logic, rather than statistics, or intuition, or hard work, or whatever, seems to be in desperately short supply in many horseracing circles.
Logic
You’ve been watching too much Star Trek
Live long and prosper :biggrin:
(Edited by empty wallet at 10:53 pm on July 1, 2005)
Sandown Today.(See earlier post)<br>Expected Ratings:
2.05 113
2.35 105
3.15 124 small field, pace?
3.45 115 pace?
4.20 93 (110-7wfa = 103, reduced to 93 because the best RPR was from last season)
4.50 90 (95 -5 wfa)
5.25 74 (85 – 11 wfa)
How can 1 sec be equivalent to 22 lb in time terms and equivalent to 18 lb (3.6*5 lengths) in form terms when they are describing exactly the same phenomenon? This is an anomaly that crops up surprisingly often in this field.
Why would you treat the poundage allowance for distance/time beaten for 5f run in 55 sec the same as 5f run in 65 sec? Most form handicappers would discriminate between the two, and most form handicappers would be right.
By producing expected ratings I take it that you are recognising the significance of ability in these calculations. It is surely much better to work out an ability rating after the event with the various results at your disposal by tried-and-tested handicapping processes which allow for margins between horses, field size and so on than go for the half measure of producing some sort of inaccurate par figure.
Form and time analysis are interlinked. Each discipline benefits from the other.
Other than that, it all looks like good stuff.
Thanks, Prufrock.
Really, what I’m trying to do is help people to get started doing their own ratings by giving them one possible approach. Most forum regulars know this area is a minefield as many previous threads will testify.
The table I am working from which was published a while ago in the RP has 1 second equivalent to 6.2 lengths at 5f and 5.6 secs at the other extreme of 20f.
It is hard to argue against the other points raised. The expected ‘form’ value could just as easily be a one arrived at after the result is known. I will take this on board.
Speed ratings worked out for Sandown’s card from Saturday July 2nd.
2.05 EXPECTED 113 ACTUAL 83<br>5f(22.0) slow 0.77 = -17, 100-17=83
2.35 EXP 105 ACT 108<br>8f(13.7) fast 0.06 = +8, 100+8 =103
3.15 EXP 124 ACT 89<br>10f(10.9) slow 1.00 = -11, 100-11=89
3.45 EXP 115 ACT 111<br>16f(6.8)  fast 1.64* =+11, 100+11=111<br>*RP reports as fast by 0.04, which is incorrect according to RP Standard Time.<br>Edit:  RP have changed the standard time to 3 min 33.40, according to website, but it was 3min 35.00 in the paper. Annoying, but it makes a difference to all calculations. I will revise asap. One of the complications of producing speed ratings, but I thought Shabernak’s rating stood out. Now I know why!
4.20 EXP 93 ACT 79<br>8f(13.7) slow 1.59 = -21, 100-21=79<br>WFA 3yo, 7
4.50 EXP 90 ACT 57<br>7f(15.8) slow 2.73= -43,100-43=57<br>WFA 3yo, 5
5.25 EXP 74 ACT 50<br>10f(10.9) slow4.59= -50,100-50=50<br>WFA 3yo, 11
The 3.15 and 4.50 were reportedly not run at a fair pace and are ignored for ratings purposes.
Going Correction = SUM(ACT – EXP)/N<br>where N is the number of races taken into account
             =[(-30) +(+3)+(-4)+(-14)+(-24)]/5<br>              =-14<br>This means that, on average, horses were being slowed by 14lbs due to the going.
Ratings for the winners
      ACT  CORR  Weight  WFA  Rider  RATING<br>Resplendent Glory   <br>2.05    83   +14   -2      +3    0     98<br>Ace Of Hearts<br>2.35   108   +14   -5      0    0     117<br>Oratorio<br>3.15   89    +14   -4     +11   0     110<br>Shabernak<br>3.45   111   +14   -2      0    0     123<br>Bon Nuit<br>4.20    79   +14   +3     +7   0      103<br>Walkonthewildside<br>4.50    57   +14   +4     +5   0      80<br>Group Captain<br>5.25    50   +14   -2      +11  0      73
Ratings for all other runners can be worked out as shown in the earlier post.
These Speed ratings are based on RPR’s and should correlate closely with Topspeed ratings, although Topspeed is usually a few pounds below RPRs. It will be interesting to see how these winners are rated by RPR/Topspeed. I will update as soon as these are available, usually about 3 days.
People will put their own interpretations on the ratings for Saturday, but they need to be looked at carefully in the context of what happened in the various races.
The outstanding rating was for Shabernak(123). Although the overall form with Midas Way looks suspect, the pace looked good and the winner was going very well 2f out and ran on strongly.
The 2.35 was very solid form and the first four are all very decent handicappers.<br>Edit:<br>Speed ratings worked out for Sandown’s card from Saturday July 2nd.
2.05 EXPECTED 113 ACTUAL 83 <br>5f(22.0) slow 0.77 = -17, 100-17=83
2.35 EXP 105 ACT 108 <br>8f(13.7) fast 0.06 = +8, 100+8 =103
3.15 EXP 124 ACT 89 <br>10f(10.9) slow 1.00 = -11, 100-11=89
3.45 EXP 115 ACT 100 <br>16f(6.8) fast 0.04 = 0, 100+11=111 <br>
4.20 EXP 93 ACT 79 <br>8f(13.7) slow 1.59 = -21, 100-21=79 <br>WFA 3yo, 7
4.50 EXP 90 ACT 57 <br>7f(15.8) slow 2.73= -43,100-43=57 <br>WFA 3yo, 5
5.25 EXP 74 ACT 50 <br>10f(10.9) slow4.59= -50,100-50=50 <br>WFA 3yo, 11
The 3.15 and 4.50 were reportedly not run at a fair pace and are ignored for ratings purposes.
Going Correction = SUM(ACT – EXP)/N <br>where N is the number of races taken into account
=[(-30) +(+3)+(-15)+(-14)+(-24)]/5 <br> =-17 <br>This means that, on average, horses were being slowed by 14lbs due to the going.
Ratings for the winners
ACT CORR Weight WFA Rider RATING <br>Resplendent Glory <br>2.05 83 +17 -2 +3 0 101 <br>Ace Of Hearts <br>2.35 108 +17 -5 0 0 120 <br>Oratorio <br>3.15 89 +17 -4 +11 0 113 <br>Shabernak <br>3.45 111 +17 -2 0 0 115 <br>Bon Nuit <br>4.20 79 +17 +3 +7 0 106 <br>Walkonthewildside <br>4.50 57 +17 +4 +5 0 83 <br>Group Captain <br>5.25 50 +17 -2 +11 0 76
Ratings for all other runners can be worked out as shown in the earlier post.
These Speed ratings are based on RPR’s and should correlate closely with Topspeed ratings, although Topspeed is usually a few pounds below RPRs. It will be interesting to see how these winners are rated by RPR/Topspeed. I will update as soon as these are available, usually about 3 days.
People will put their own interpretations on the ratings for Saturday, but they need to be looked at carefully in the context of what happened in the various races.
The outstanding rating was for Shabernak(115). Although the overall form with Midas Way looks suspect, the pace looked good and the winner was going very well 2f out and ran on strongly.
The 2.35 was very solid form and the first four are all very decent handicappers, although on reflection 120 looks quite a bit high.
<br>
(Edited by Artemis at 7:22 pm on July 3, 2005)<br>
(Edited by Artemis at 5:49 pm on July 4, 2005)
Interesting exercise.
It’s hard to compare figures because I don’t know what scale you use, although I presume it is the 0-140 scale most handicapping methods use.
I think I am wrong in using the straight course time as an indicator of the going across the whole course. If I ignore it, the going correction comes down to (- 12) which pulls all the ratings down by 5lbs.
This puts all my ratings about 5 or 6lbs above Topspeed.
Your ratings are fairly close to Topspeed with the notable exception of Walkonthewildside who Topspeed has on 73 and I have on 83(or 78, if you use the revised going correction).
I think Prufrock is right when he advises the use of the actual result and other relevant information before calculating the speed figures. Certainly a couple of my ‘Expected’ figures, particularly those relating to non-handicaps were, on reflection, too high and this distorts the final ratings.
The only trouble is it makes it hard work because you have to rate the form using traditional methods(which takes time) before you tackle the relative times to see if the form is enhanced/ endorsed/devalued by the clock.
This is why I take the lazy route and use the RP ratings for form/speed. But at least I’m very familiar with how these ratings are arrived at.
Thanks for co-operating in this little exercise.
The only time I would return a figure for one race on the entirely separate straight course is if I had sectionals to guide me. The going allowance is unlikely to be the same on the straight and the round course in such circumstances, but sectionals at least allow you to make an educated guess as to whether the one race on the straight course "should" have resulted in a good time or not, and an educated guess is probably better than nothing in such an instance.
(Edited by Prufrock at 7:47 pm on July 4, 2005)
A bit of a puzzle about the RP standard for the Sandown 7f 16yds.
It’s certainly a fairly stiff 7f, but not as stiff, according to RP standard times, as a few other tracks with a similar configuration such as Carlisle, Hamilton and Ascot. Even the LH track at Bath with its uphill straight is slower according to RP.
If we could ask a few experts to rate the abovementioned  tracks in order of how stiff they perceived them to be(at 7/8f round course), I wonder what order they would put them in?
The RP order in terms of stiffness based on RP standard times is:<br>Stiffest   Ascot<br>        Hamilton<br>        Carlisle<br>        Bath (LH)<br>        Sandown<br>        Beverley
There may be other factors at work apart from the  uphill finishes at these courses such as tightness of bends which might slow horses down.
Maybe the RP has got it wrong.<br>
(Edited by Artemis at 1:36 pm on July 6, 2005)
- Total Posts 17716
Although I wouldn’t consider myself an expert, I would suggest it is pretty pointless to consider the standard times of a course like Ascot, relative to such as Carlisle & Bath, unless one takes into account the calibre of horses that race over the respective tracks and distances, thus creating those standards.<br> Similarly, from a purely time point of view, the round 1m at Ascot would appear to be a stiffer test than the straight mile, an arguable point considering the former, by dint of the sharp bend into the straight, is possibly less of a test than the all-out gallop of the latter.
(Edited by reet hard at 10:21 pm on July 5, 2005)
AS EC states, the standard times are(or should be) unaffected in any way by the class of horses running at the course.
For example, group races don’t take place at Catterick, but the 5f standard is not that much different than that of today’s July course at Newmarket. The idea is that a mature horse rated 100, carrying 9st, should run to the standard time at ANY course.
Median times and average times for various courses are proportional to the standard of races at the course and should not be confused with standard times.
- You must be logged in to reply to this topic.