- This topic has 109 replies, 27 voices, and was last updated 11 years, 6 months ago by
.
-
Posts
-
I just spent several hours putting together a list of Racing Post standard times for all British racecourses, might be of use to somebody (Excel.xls).
<!– m –>http://tinyurl.com/69fol76<!– m –>
Using a simple formula I have added a LBS per length column to the Racing Post standard times Excel.xls sheet.
*The standard time for Haydock Park’s 1m 2f distance is correct, before anybody says it isn’t
I have just received a PDF file from Dave Edwards "Top Speed" listing all the standards times for the British/Irish National hunt.
He hasn’t finished the Flat yet
- Total Posts 17716
Blues brother, very interesting indeed.
By standard time you mean track variant = 0, I presume.A horse that interests me right now is Shearman (Elusive City – Champion Tipster).
Raced at W’ton AW (won) with 1.14.26 slow by 1.26 (that makes it 1.13.00 ?)The standard W’ton time for 6f is also 1.13.00 in your chart.
If with the same weight he goes on to Ascot where you say standard time is 1.13.50 then the W’ton time is equivalent to (73.5/73) x 73 = 73.5 = 1.13.00.Correct ?
Shearman seems to have flopped at Ascot but that does n’t matter. Maybe it was an unlucky ride.
To work out standard times for a new racecourse we have to compare horses that ran on both. Correct ?
I can’t judge the accuracy of your results and the lbs per length (kg per sec in my units) but one thing you need for sure is conversion formulas from one distance to another, to make sense of it all.
If you need any help pls write.I have done all this work before.
Now with Shearman doing 1.13.00 (= 1.14.26 – 1.26) I can’t really tell what this means. I have n’t looked at the card featuring Shearman in all detail but it seems the press thinks as firm favourite Sharp Eclipse (Exceed and Excel – Helen Sharp) and Shearman is an also run.
@Froddo
I will explain what I am attempting to achieve with my lbs. per length calculations.
I am a big admirer of Dave Edwards “Top Speed” but I noticed a flaw in the way the Racing Post used their lbs. per length calculations.When they rate a 5f 2yo race at Epsom they use 3.54 lbs. per length and .16 seconds equals one length and they use the 3.54 lbs. as a constant for all 5f races.
OK you are thinking what is wrong with that?
Take the 2yo 5f furlong track at Epsom it is run downhill so how can you have the same calculation 3.54 lbs. for the 2yo 5f race at Pontefract which is run uphill.
RP Epsom 5f = 3.54 lbs.
RP Pontefract = 3.54 lbs.I would use Epsom 5f = 3.61 lbs.
I would use Pontefract = 3.23 lbs.Every course must have its own set of lbs. per figures, which is why when looking at the last 3 top speed figures of any horse you see no consistency as they are using the same constant for the race distances instead of the individual track configurations.
I was looking at a race at Wolverhampton recently run over 5f 20yds and the race time was 1.69 slow and Top Speed gave the winner a rating of
61
, remember they use 3.54 lbs. per length at the Racing post and .16 sec per length.
So using my calculator and my lbs. per length for Wolverhampton’s 5f 20yds which is 3.32 lbs. I roughly arrive at a figure for the winner of63.8
and that was using their .16 sec per length calculation and if I used my .20 sec per length calculation the figure comes out at
70.9
Do track variants matter? I think so…
- Total Posts 17716
You have many race courses.
We have only one.
I had to do track to track conversion only once, in 2004 because of the new race course.Regarding weight I use the following standard figures:
5f – 0.03 kg/sec
6f – 0.05 kg/sec
7f – 0.06 kg/sec
8f – 0.06 kg/sec
9f – 0.08 kg/sec
12f – 0.08 kg/sec
15f – 0.20 kg/secThat means if for ex. a horse raced 6f at 1.13.00 with 58 kg and is racing again 6f with 60 kg, the 1.13.00 becomes 1.13.10.
I showed them to an American trainer once who more or less agreed.
I also have a periodical here which says 0.20 kg/sec for all distances. That looks awfully wrong to me.One question is how does one derive those.
I don’t like using the least squares method.
Maybe weighted least squares.
My 2005 statistics based on least squares were crappy (I had to make statistics on that year because of the new race course).
Then in 2007 I used some other method based on probability theory with better results.You say the 0.06 kg/sec if uphill becomes 0.08 and if downhill perhaps 0.05 ?
I don’t know what you can do to overcome the limitations of leat squares but you have to work out distance to distance conversion formulas.
Why this ?
If your horse has made a debut over 8f and now runs 7f what are you going to say ? Nothing ?
If your hourse runs 7f and the last outings are 8f-5f-5f-7f you go to the -4 race to get clue ? Bad idea I should think.Perhaps you can help me make a UK to Greece chart.
I used to have the beginnings of a table somewhere but I lost it.
It’s not diffcult though to reconstruct one.see this:
http://en.wikipedia.org/wiki/Linear_regression
I was reading an old post of Simon Rowlands, when he mentioned that the BHA official WFA chart was a bit out, so I asked Dave Edwards "Top Speed" about this an he emailed me the Racing Post’s WFA chart.
Simon was right, it’s miles out, just compare the two…dont get me wrong i love the racing post but their standard times are just too standard. If you want to getinto standard times you need to make a whole set of figs on your own only to realise what little use figs are on the turf and uk in general. However whilst building my figures i understood how figs are achieved. And that was more important that having a set of perfect figs/times to hand. sorry.
Hi Thebluesbrother
Unless I am misunderstanding you i think your statement below is incorrect, due to you thinking about the time and lengths beaten equation the wrong way round.
"Take the 2yo 5f furlong track at Epsom it is run downhill so how can you have the same calculation 3.54 lbs. for the 2yo 5f race at Pontefract which is run uphill."
In actual fact the Official Lengths beaten that you will see in the result of a race has nothing whatsoever to do with distance. It is in actual fact just a representation of time e.g. On good or better ground on the flat for every second a horse beats another this is represented in the result by 6 lengths and for soft ground this represents 5 lengths.
So in your example a horse beaten a length at Pontefract on good ground finished .16 seconds behind the winner. A horse beaten a length on good ground at Epsom also finished .16 seconds behind the winner. Therefore it is perfectly reasonable to value that length the same in your speed figure calcs.
Of course in reality the horse beaten a length at Pontefract would have been closer to the winner when the winner past the post, than would be the case for the Epsom horse however this is irrelevant for what you are trying to do.
I hope this makes sense!!
@ Alan1
I was just reading an article written by B.J.Healey a professional handicapper in 1989 and edited by Tony Peach.
When I got to his pounds per length tables he had already produced a similar format to what I had been working on.
An Example from his table:
Epsom 5f = 4.04
Pontrefact 5f = 3.62Variant = 0.42 lbs.
The Racing Post says:
Epsom 5f = 3.54
Pontrefact 5f = 3.54Variant = 0 lbs.
I say:
Epsom 5f = 3.61
Pontrefact 5f = 3.23Variant = 0.38 lbs. (Very similar to B.J.Healey)
In his article he states that he spent a long time trying to come up with the holy grail formula to work out each track’s variants and achieved this by accident when doodling one day.
I had already worked this out by myself, it took me 5 minutes to come up with a simple formula.What I am trying to achieve is to get tighter speed figures for every race run on different tracks, i.e. "Go Maggie Go" which was the first of my trials, her last two ratings came out at 64 & 62 so I can see a true reflection of a horse’s ability no matter what track they run on.
I had better go and finish off the Irish tracks
Of course in reality the horse beaten a length at Pontefract would have been closer to the winner when the winner past the post, than would be the case for the Epsom horse
Is this really the case?
Just finished the RP Standard Times for all the Irish courses on both the flat and Jumps.
Now I know why I have never seen a list of standard times for the Irish flat, as they keep on moving the starts i.e. Wexford 1m100y & 1m110y etc.
I had to pick through my excel data list of 150,000 races to check the distances and dates.
I realise I may be opening a can of worms here but can anyone shed light on how Dave Edwards arrives at his going correction?
The reason I ask is that when looking at the times of winners and the seconds per furlong outside/inside the RP standard, I then find it difficult to work out how Edwards gets to his going correction figure for the meeting?
Thanks in anticipation.
An update on my project.
Now that I have finished the standard times for all Irish and English racecourses and applied the lbs. per length to every race distance at every racecourse, I have started to apply the speed ratings.
I can rate an eight race card at Wolverhampton AW in 15 minutes which isn’t too bad.
I am trying to achieve tighter speed figures for horses across all distances and tracks so that I can assess a horses true ability.
An example:
I have four speed figures for "Captain Dimitros" trained by David Evans and they are:50,51,59 and 54
and if you compare them to "Top Speed" figures you get
47,47,61 and 53
.
So what is Captain Dimitros ability, I say
54
.
So far so good
I have my ratings set out on an excel sheet, I have separated them into three groups i.e. 2yo’s, 3yo’s and 3yo+.
The ratings will cover every flat race run in England and Ireland and will be updated at least twice a week.(Excel.xls)
Of course in reality the horse beaten a length at Pontefract would have been closer to the winner when the winner past the post, than would be the case for the Epsom horse
Is this really the case?
Yes it should be, at least based on conventional thinking.
Firstly I should say that I don’t compile ratings of any kind and I am far from an expert in speed ratings and I would appreciate if one of the forum speed ratings experts could either confirm or condemn what I have written below
My understanding of the theory however is as follows:
Suppose Horse A is beaten by 1 second at Epsom and Horse B is beaten by 1 second at Pontefract. Both horses will be given the distance beaten as 6 lengths because 1 second = 6 lengths.
However Horse A (Epsom) should actually have finished very slightly further away from the winner than Horse B (Pontefract) because the race at Epsom will have been run at a faster pace because it is a downhill track.
This raises 3 other questions IMO
1. Is a horse beaten by 1 sec on a downhill track better than a horse beaten by 1 sec on an uphill track?
Most people would say that the horse beaten on the uphill track is slightly better.
2. Is the difference in track already built into the differing standard times?
Again most people would say yes
3. Is the difference in the standard times sufficient to ignore the differing track undulations i.e. – is the standard time accurate enough that you can ignore the track variances when compiling speed figures?
Nobody can answer this!
Depending on how the actual speed figures are compiled, there may be a danger that you overrate the Pontefract horse because you are double-counting the distance beaten of the Epsom horse by counting it in both the standard time and then again in the lengths beaten.
- You must be logged in to reply to this topic.