Compiling Standard Times

[The Hare]

Hello

After my original post about the worth of speed figures i decided to get around to compiling some Standard Times.

There are so many ways they can be compiled which i feel is the major problem with speed figure calculation. So using the excellent work of "The Blues Brothers" (Mile), I obtained his latest update of Racing Post Standard Times. (Thanks for your efforts Mike!).

Although i am really more interested in Jump racing that looks hard to decipher at the moment due to all the distance changes!! That can wait for another day. So British Flat Racing was the start.

I've subscribed to Raceform Interactive to have easy access to the data in the required format. After a couple of months of sorting the data and determining a method i am finally able to calculate some Standard Times. It's not an easy task for sure!

The problem then is ending up with some wildly different figures to those of the Racing Post. I don't have access to Timeform so can't compare with those.

I suspect everyone has slightly different methods for compiling Standard Times so that will explain some of the differences but there are some really wild differences.

To obtain the standards my feeling was i needed to adjust the actual winning time based on a number of factors as below:

1. The actual winning time
2. The weight carried by the jockey (i included jockey allowance)
3. The ability of the horse - I used the Raceform Rating for the winner (or Racing Post)
4. I have used the last 5 results for each track.
5. By using RR/RP ratings the WFA scale shouldn't affect the final times as it has been factored in.
6. The number of runners in a race has been adjusted for though i haven't don this on a race to race basis which perhaps i should.
7. the most obvious affect on the time is the hardest to adjust for!! That is of course the going...i have made no correction for this as really until calculating my figures there is nothing to compare with - I suspect this will have a major problem on the final times until i adjust for this.

Whilst all this is good there are of course problems with this method even though it is almost certainly more accurate than the RP method? I will continue

 

[The Hare]

So then the problems from this.

I decided that after sorting the adjusted finishing times i would then use the upper quartile as the standard time.

So if say there are 80 races at the trip then the 20th fastest was my chosen Standard Time which i then amended based on the average number of runners in those top 20 races. This is an arbitrary selection which as a standard method seems reasonable. I believe most methods use say the 20 fastest times at the trip which falls down if there are many more races at one trip than another.

So the major problem with this and all Standard Time calculations is the lack of data. Especially over longer trips there are so few races in a 5 year period that the calculation is based on very flimsy evidence. However the times produced at these longer trips do appear reasonable in most cases.

My main reason for doing the work will be to use for Irish and National Hunt Racing as there seems to be no reliable figures produced for these races, Ireland i suspect due to the "casual" outlook on timing from the courses there. However i know from my previous efforts 20 years or so ago that NH figures were very useful.

So after compiling a few Standards for Flat Racing in the UK there appear to be some major differences on some Standard Times compared to the RP figures - woth variations of close on 50lbs which is an enormous error. Perhpas Mike would like to check these out?

Ascot 2m Standard on the RP looks too low, whilst the 1m(Rnd) and 10f look too high - on my figures a swing of 30lb or so between the ratings these would produce.
Ayr 9f time on the RP looks too tough - maybe a second too low?
Carlisle has my biggest anomaly. The Racing Post Standard for the 7f200y trip looks way too tough - best part of 3 seconds too low? Checked the figures and they seem correct.
Catterick - The 7 and 14f times appear a little tough
Chepstow - I imagine the RP standard on Mike's figures is incorrect? Surely should be 207.0 rather than 177.0? Even then the Standard used by the RP iwould be real tough as also appears to be the 2m2f though i have very little data for that.

Can post more later if any interest.

 

[TheBluesBrother]

So the major problem with this and all Standard Time calculations is the lack of data. Especially over longer trips there are so few races in a 5 year period that the calculation is based on very flimsy evidence. However the times produced at these longer trips do appear reasonable in most cases.

I solved this problem about lack of data over longer trips by using linear regression.
When I was working on the standard times for Auteuil (FR) , once a year they run a 3m6f chase so I use a standard time calculated using linear regression.

Chepstow - I imagine the RP standard on Mike's figures is incorrect? Surely should be 207.0 rather than 177.0? Even then the Standard used by the RP iwould be real tough as also appears to be the 2m2f though i have very little data for that.

You were right about the standard for 2m49y at Chepstow being wrong, I have just updated the Chepstow standard times and changed the standard time to 208.0s, I am forever changing and updating the list on a daily basis.

Carlisle has my biggest anomaly. The Racing Post Standard for the 7f200y trip looks way too tough - best part of 3 seconds too low? Checked the figures and they seem correct.

Again you are correct, I just updated the standard times for Carlisle and have the new standard time for the 7f200y trip at 98.50s
These standard time changes don't affect my speed figures, as I do not refer to the standard times list unless I am working on the French speed figures.

Ascot 2m Standard on the RP looks too low, whilst the 1m(Rnd) and 10f look too high - on my figures a swing of 30lb or so between the ratings these would produce.

Reference the 2m standard, linear regression says it should be 204.5s, so I made a note of it.
Looking at the standard for the 1m2f trip, I have to admit it does look lightly high, I calculated it should be 123.10s and noted it on the list.

My main reason for doing the work will be to use for Irish and National Hunt Racing as there seems to be no reliable figures produced for these races, Ireland i suspect due to the "casual" outlook on timing from the courses there. However i know from my previous efforts 20 years or so ago that NH figures were very useful.

You have a very serious problem when it come to compiling standard times for Irish racing, take yesterday's meeting at Sligo, I calculated the 2m hurdle races were run over 1m5f, and most of the time you can take the advertised distances at racecourses like Tramore with a big pinch of salt.

Just recently they re-measured the race distances at Galway and found the 2m1f(C) trip was actually being run over 2m2f.

Mike.

 

[seb3939]

Hello how do you calculate the rating on the race on which you based thank you