• Hi Guest Just in case you were not aware I wanted to highlight that you can now get a 20% discount on Inform Racing.
    Simply enter the coupon code ukbettingform when subscribing here.
    We have a lot of members who are existing users of Inform Racing so help is always available if needed.
    Best Wishes
    AR
  • Hi Guest Just in case you were not aware I wanted to highlight that you can now get a free 7 day trial of Horseracebase here.
    We have a lot of members who are existing users of Horseracebase so help is always available if needed, as well as dedicated section of the fourm here.
    Best Wishes
    AR

Approaching speed figures

I've now got a few standard times to have an experiment with, and using those I've collated in my thread from the weekend on that subject, I've now decided to have a go at creating some ratings from them.

To explain, the standard times are for all weather races from 2019, and I've now tried some ratings for the first couple of days of 2020. The steps I've used are these and the horse in question was Cat Royale who won the first at Chelmsford City on 2 January 2020. It looks more convoluted than it actually is, if you have the data.

1619555726730.png

I have taken the view that in creating standard times that I got some decent ones when removing outliers and going to the 6th percentile. In the above example, I've aligned all performances to what a 100 horse would be expected to achieve, and in the above example, the theoretical 100 horse might well have won the race in 123.38 (8) as opposed to the standard of 121.32, Cat Royale therefore ran just under 2 seconds above the standard for 10 furlongs, and his going adjust was 0.21 (9).

After removing all outliers, the 6th percentile figure going adjust for the meeting was 0.14, so I've adopted this as an acceptable going allowance for the meeting. Should be noted, the last two races did not have any horses with ORs, so none of those runners are factored into the going allowance at this stage, I could use a formula from HRB ratings to guess their likely ORs, or take the resulting RPRs and convert them to OR, but I'm reasonably OK with the idea that most horses running here had an OR that allows a fairly accurate impression of the going allowance.

Below were the figures I got for the winning horses. The idea would be to have a rolling 12 month standard time for each track, so a race on 1 April would be based on standards applicable to the 12 months before, and so on. There's no obvious suggestion that the track got especially faster or slower as the meeting went on, based on the last column, but if you feel there's a clear trend I guess you could always alter your going allowance to suit.

1619555998214.png

Brigand won the second race with a decent figure, but I've noted that Agent of Fortune was 4th with a SPD on my scale of 62 which matched his OR, he then went on to win 3 of his next 4 races. Again interested in views from members of the forum.
 
I also did Lingfield on 2/1/20 and I found the track to ride evenly with no going allowance at all. For Southwell on 1/1/20 it seemed to be fast by 0.02 seconds per furlong. So the Chelmsford track above looks quite slow in comparison on the 2nd.
 
Very impressive, my old maths teacher would be proud of you showing your workings. Sadly I don't have enough knowledge to make a helpful comment, other than this amount of effort deserves to succeed and I hope it does for you. Will follow and looking forward to when you get some ratings to use in practice.
 
There's an error in step 9 of my formula. I did the initial one for a 7f race at Southwell but amended for the 10f so it should be 2200/ 220 rather than 1773 / 220 but the actual calculation itself is correct.
 
Hi AustinDillon75 AustinDillon75 , it looks like you have been very busy and the best of look going forward.
If you get the time can i ask you what you think are the merits of both RANENSCAR & CHARLIE D based on your numbers for thursday 7.25 south.
My choice is for CHARLIE D (9/2 atm) who has only had one go here and won 14f, Ranenscar is 3 - 3 here so i can see why he is fav 9/4.

No prob if you're too busy.
 
Hi AustinDillon75 AustinDillon75 , it looks like you have been very busy and the best of look going forward.
If you get the time can i ask you what you think are the merits of both RANENSCAR & CHARLIE D based on your numbers for thursday 7.25 south.
My choice is for CHARLIE D (9/2 atm) who has only had one go here and won 14f, Ranenscar is 3 - 3 here so i can see why he is fav 9/4.

No prob if you're too busy.
Thanks Tacker, yes working and trying to squeeze a bit of this in when time allows.

What I'm doing is taking 2019 standard times and working through from 1 Jan 2020, to get accustomed to how to work figures out and seeing how my numbers play later in practice. I don't have any figures for today at all, sorry. No doubt others can give you a far more informed view.
 
@ AustinDillon75 AustinDillon75. Excellent work but looking @ TheBluesBrother TheBluesBrother ratings for the same meeting there is clearly a problem.
Not the standard times, they will throw up slightly different ratings, but if you look at the last 3 races over 1 mile Mike's ratings are very different from your own. I don't have any ratings myself for UK races but from Mike's yours should be similar with 6.30, 7.30 and 7.00 in that order. These do not match your own and.

I suspect this has something to do with your adjusting the time of the winner as if a horse of 100 rating was carrying 9-00. this method is used in the calculation of standards as a way of putting all the times on a basis they are for a horse of similar ability. I have not known this be used in the calculation of the Going Allowance. Usually the difference in the winning time and standard time is adjusted to a rating in pounds and compared with the expected rating of the winner and then compared to the other races at the meeting. The adjustment would be using your calculation items 2 and 3.

Apologies if I am incorrect but the experts should be able to help you on this.

Great work though Austin.
 
The Hare The Hare yes I think the reason TheBluesBrother TheBluesBrother posted those ratings was to issue a challenge, i.e., "here are my times try and work out for yourself why they are different."

So that is exactly what I am intending. I wouldn't believe I could ever be as accurate as some of the experts on here but if any comments get rid of utter howlers then this is exactly what the thread is for - its a process aiming to work at being more accurate.

As you say looking at those three mile races at Chelmsford may prove instructive. I have added 12lb WFA to the performances of Summit Reach and Al Maysan, whereas I don't think TheBluesBrother TheBluesBrother uses WFA. If I removed the WFA we'd have the following:

1619677296904.png

As you can see, that changes things a fair bit. However, there are fundamental differences in some of the others, notably Southwell where he rates Xian Express at 37 to my 47, which I think is possibly due to us having completely different standard times, and for the first race, he had Jazz Hands at 28 to my 56. I didn't think it was unrealistic to believe a 61 rated horse might have achieved the latter.

That said I've spotted some formula errors in the sheets which could be throwing out a couple of others so the Southwell ones might end up being corrected with the Chelmsford ones above going wrong.

I think The Hare The Hare might be suggesting is to take the times on face value, then if its a 60 horse project what the time of the 60 horse might have been and work off that to produce going allowance. I will have a play with the spreadsheet on that basis and recalculate those meetings.

Great advice again <thumbs up>.
 
I think @
The Hare
The Hare might be suggesting is to take the times on face value, then if its a 60 horse project what the time of the 60 horse might have been and work off that to produce going allowance. I will have a play with the spreadsheet on that basis and recalculate those meetings.

@ AustinDillon75 AustinDillon75 that's correct, don't know how these will affect your figures but may be the reason. As for the WFA, never thought of that! There is nothing wrong with adding in WFA, that is a personal choice and explains the difference in those 1m races.

 
I don't think @
TheBluesBrother
TheBluesBrother uses WFA. If I removed the WFA we'd have the following:
It is not recommended to use WFA when compiling speed figures, why corrupt your good work, Admiral Rous was no doubt a very clever man but he made the incorrect assumption that horses obtain stamina with age, the RP uses a complete different WFA table to the BHA.

I developed my method of speed figures and the lbs per length, which are calculated from the standard times for every racecourse and distance, it took me a couple of years to understand going allowances which is the critical concept of speed figure compiling.

I have every meeting I have worked on in a Rar file from 2018, to unpack use the free utility 7-zip if you do not have Winrar, Winzip.


My speed figure excel sheets.

Mike.
 

Attachments

  • WFA.pdf
    29.9 KB · Views: 17
TheBluesBrother TheBluesBrother thanks I will have a close look at all that when I have a little more time and point noted re: WFA. I am happy to experiment with it on the side, and see what differences it could make.

The Hare The Hare so I redid that Southwell meeting with the help of each winner having an OR, derived their expected time compared to the standard, I then adjusted for weight but all we are doing is robbing Peter to pay Paul, so if you look at the first race, Jazz Hands ran the mile in a time of 102.09, but as a 61 horse carrying 8st 9lb I would have expected him to run it in 101.68,

100 - 61 + (126-5) = 34 / 2.03 = 16.71 lengths
16.71 / 5 (Southwell lengths per second = 3.35 (rounded)

98.33 + 3.35 = 101.68

So he runs it in 102.09

102.09 - 101.68 = 0.47 / 8 = 0.05.

The fact the last three races at that meeting were faster compared with expectation these calculations means a 6th percentile for the meeting as a whole of -0.02. They have run well when you consider each were carrying decent weight in their races, Oblate was carrying 12lb more than Jazz Hands, and in my own mind its not unreasonable to believe his rating should be 15 better when he's ran the race just 0.39 seconds quicker, yet TheBluesBrother TheBluesBrother has him +26 better?


1619689655063.png

The other step here is compare standard to actual giving a SPD performance, simply derived from actual - standard time divided by lb/length. The weight adjustment is what it carried under 126 and then the speed adjusted. If compared to the winning ORs look at the differences in those times. Xian Express is 0.46/second slow on the above chart, but -41 in the chart below, and orderwise Raydiance looks the next slowest in both charts below (0.27 vs -22). The fastest was Oblate (-0.16 vs +13).

1619690548651.png

How the hell do you pick off a going allowance from all that? Did something happen here after race 4?

Also apologies if this should really go in the established Speed Ratings thread but I didn't feel I could justify cluttering up all the discussion in there so thought better to plough my own path here seperately.
 
@
TheBluesBrother
TheBluesBrother thanks I will have a close look at all that when I have a little more time and point noted re: WFA. I am happy to experiment with it on the side, and see what differences it could make.
Just compare the RP WFA for a 2yo in March against the original BHB WFA table and note the difference, then you will understand why using WFA is a no-no while compiling speed figures.

Mike.
 
@AustinDillon75
Hello
When i compiled speed figures for the UK, I would use the second method in your post - adjusting the time difference between the winner and standard into pounds. So your second table above would appear correct. I don't understand the first table you have, seems a complicated method if you want to rate the horses in pounds.

So the difference you come across here is the dreaded WFA! Also if there is a jockey allowance.
Firstly WFA, @TheBluesBrother doesn't use WFA, which i agree with but it is a personal choice. The problem this has on your ratings is that you need to subtract the WFA from the OR where it applies. In this case Xian Express as a 3yo will have WFA included in the OR and quite likely some of the 4yo also as it is early in the year. For instance in the 3.05 race the 4yo appear to have a 2lb WFA. If you are not including the WFA in your ratings then you need to subtract these from the OR or you will be on a diffeent scale. The table for WFA Mike has attached above has 3lbs for WFA in January, so may be slightly different but who knows? It is different wherever you look! The BHA figures are what you need.

Then there is the Jockeys allowance. I note in the 1.55 a 3lb claiming jockey rode the winner. So you appear not to be including that, again that is a matter of choice and discovery! The problem again is that the OR will include these 3lbs in that rating. So you will need again to deduct this amount from the OR if you think the Jockey allowance has any bearing on the performance which it should. The problem here though is it will take a lot of work finding all the jockey allowances. Though I think HRB does have them, the Racing Post details are harder to find.

Great effort though. As for the Going Allowance, doing it the 2nd way gives you a Going Allowance in terms of pounds. You could just adjust the scale @TheBluesBrotheruses into pounds and this will allow you to determine the type of going it is.

Good Luck
Kevin
 
@AustinDillon75
Hello
When i compiled speed figures for the UK, I would use the second method in your post - adjusting the time difference between the winner and standard into pounds. So your second table above would appear correct. I don't understand the first table you have, seems a complicated method if you want to rate the horses in pounds.

So the difference you come across here is the dreaded WFA! Also if there is a jockey allowance.
Firstly WFA, @TheBluesBrother doesn't use WFA, which i agree with but it is a personal choice. The problem this has on your ratings is that you need to subtract the WFA from the OR where it applies. In this case Xian Express as a 3yo will have WFA included in the OR and quite likely some of the 4yo also as it is early in the year. For instance in the 3.05 race the 4yo appear to have a 2lb WFA. If you are not including the WFA in your ratings then you need to subtract these from the OR or you will be on a diffeent scale. The table for WFA Mike has attached above has 3lbs for WFA in January, so may be slightly different but who knows? It is different wherever you look! The BHA figures are what you need.

Then there is the Jockeys allowance. I note in the 1.55 a 3lb claiming jockey rode the winner. So you appear not to be including that, again that is a matter of choice and discovery! The problem again is that the OR will include these 3lbs in that rating. So you will need again to deduct this amount from the OR if you think the Jockey allowance has any bearing on the performance which it should. The problem here though is it will take a lot of work finding all the jockey allowances. Though I think HRB does have them, the Racing Post details are harder to find.

Great effort though. As for the Going Allowance, doing it the 2nd way gives you a Going Allowance in terms of pounds. You could just adjust the scale @TheBluesBrotheruses into pounds and this will allow you to determine the type of going it is.

Good Luck
Kevin
Thanks, that is great stuff and very kind encouragement.

I do need to get my head around it a bit. I looked at TheBluesBrother TheBluesBrother sheets yesterday and I must admit, after an hour of completely drowning in all those figures (for which I am grateful) I decided to strip it back a bit, not exactly back to the drawing board, but picking up a very different pen.

My impression of The Hare The Hare comments and observation of how others work is that simplest is best and complex maths can sometimes steer in the wrong direction. You can try and be perfect but its clear that doing so can risk sending you further down blind alleys.

So I saw TheBluesBrother TheBluesBrother had run an example few years run an example using Dawn Approach. What I've done is used that to try and get a figure for Red Cymbal winning the first on 3/1/20 at Dundalk. Using Mike's methods and figures rather than mine, I worked out how he got his figure of 29 (I got 28, which might be down to a rounding issue).

I found without his going allowance of 0.17 Red Cymbal would have actually run to 43 and as such the track was perceived to be speeding that particular runner by 15lb.

In other words for Red Cymbal I had the horse running 2.17 seconds over my standard (vs. 2.44), but then divided that by 0.16, and as I have the same 3.47 per length, the pre going adjustment figure on my running amounted to 42, and I think that makes more sense that the figure I previously produced. I can now experiment from there are learn methods for going adjustments etc.

However, one question arises, I note that the corrected going allowance is (at least in that example) divided in the examples by 0.2/length. Its no different to actually multiply it by 5.

So what I would ask is, whether there is a good reason for certain UK All Weather tracks not to multiply it by 6, or divide by 0.16666/length, when the lengths per second on BHA sheets suggest that exact figure? To organise my own standard times I followed the BHA sheet which I think davejb davejb put up but I just used 5 for Southwell and felt from looking at Irish Racing times that Dundalk was flipping between 5.5 and 6 lengths per second so for my sins I settled on 5.75.
 
So what I would ask is, whether there is a good reason for certain UK All Weather tracks not to multiply it by 6, or divide by 0.16666/length
If you are looking at my work, first of all, I do not use any BHA adjustments, I use my own lbs per length figures.
I work to 0.2s equals I length (10ft horse) if you want to use 0.166s (8ft horse) that would be fine, as you rightly said instead of dividing by 0.2s you could multiply by 5.
At the bottom of the rar file you will find a template it has all the formulas I would use, all you have to do is enter the data, furlongs, lbs per length etc.

Mike.
 
If you are looking at my work, first of all, I do not use any BHA adjustments, I use my own lbs per length figures.
I work to 0.2s equals I length (10ft horse) if you want to use 0.166s (8ft horse) that would be fine, as you rightly said instead of dividing by 0.2s you could multiply by 5.
At the bottom of the rar file you will find a template it has all the formulas I would use, all you have to do is enter the data, furlongs, lbs per length etc.

Mike.
Noted. I am looking at your work but wouldn't be repeating it, because we will have our own standard times, allowances, beaten distances and seconds/length etc. That said I am pretty confident my next proper stab at this is going to feel far closer to where I want to be.

I do have a pretty testing standard time for Southwell 5f, it seems about a second slower than yours, but we'll see what we get as time goes on. Maybe experience of the figures it derives will confirm a need for adjustment later.
 
I do have a pretty testing standard time for Southwell 5f, it seems about a second slower than yours, but we'll see what we get as time goes on. Maybe experience of the figures it derives will confirm a need for adjustment later.

Southwell is changing surface very soon as was mentioned elsewhere recently.

 
I do have a pretty testing standard time for Southwell 5f, it seems about a second slower than yours
I wouldn't bother about the standard times for Southwell AW yesterday I do believe was the last fixture on the fibresand.
The main problem at Southwell is the extremes in the going, just recently the surface has been riding standard to slow -0.25s/f when it has been raining and the surface has been packed down, you end up with surface speeds of +0.40s/f (standard to fast), that's a 65lb per furlong difference.
 
Back
Top