• 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

db standard times and ratings

  • Thread starter Deleted member 7962
  • Start date
Update: So I've taken some time just going over the Chelmsford figures I have created but one thing that keeps bugging me is the going allowances and how best use this. I take the winners on the card and use a method where firstly the going allowance is set as 0 and I create a base speed rating based on that which is then compared to the Horses OR, this then gives me a difference for each winner, I.e. below;

Screenshot_20210604-115315_Office.jpg

So this is a made up 8 race card. You can clearly see horse C and horse E have ran far better than anything else and suggesting the track is speeding up the horses. Now usually I would take the 3 best times in a 8 runner card but with the 3rd best appearing at -2 (horse G) it doesn't seem right to include this to calculate a Going Allowance for the meeting as the other 2 are far better/quicker. Would others use the 3 best times regardless? I guess with more experience and alot more time on all this I would just be able to look and go that's what I'm going to use but still learning the trade. Any thoughts/help welcome.
 
db1 said:
Update: So I've taken some time just going over the Chelmsford figures I have created but one thing that keeps bugging me is the going allowances and how best use this. I take the winners on the card and use a method where firstly the going allowance is set as 0 and I create a base speed rating based on that which is then compared to the Horses OR, this then gives me a difference for each winner, I.e. below;

View attachment 101445

So this is a made up 8 race card. You can clearly see horse C and horse E have ran far better than anything else and suggesting the track is speeding up the horses. Now usually I would take the 3 best times in a 8 runner card but with the 3rd best appearing at -2 (horse G) it doesn't seem right to include this to calculate a Going Allowance for the meeting as the other 2 are far better/quicker. Would others use the 3 best times regardless? I guess with more experience and alot more time on all this I would just be able to look and go that's what I'm going to use but still learning the trade. Any thoughts/help welcome.
Click to expand...
It’s so difficult using OR , horse C and E might have won by many lengths and due a big rise in the OR, if that was the case you could argue F and G give a truer reflection of going. This is why I prefer to use performance figures given the choice like RPR
 
Outlander said:
It’s so difficult using OR , horse C and E might have won by many lengths and due a big rise in the OR, if that was the case you could argue F and G give a truer reflection of going. This is why I prefer to use performance figures given the choice like RPR
Click to expand...
I guess OR are my choice as its easily accessible to me via results from HRB. I could use performance figures like the RPR or BHA but would be more manual input I suppose unless I can search for the formula of some sort to convert. Either that or could remove the outliers? Fastest/slowest to get more of a truer number? It's something I need to look at more I think before moving on.
 
db1 said:
I guess OR are my choice as its easily accessible to me via results from HRB. I could use performance figures like the RPR or BHA but would be more manual input I suppose unless I can search for the formula of some sort to convert. Either that or could remove the outliers? Fastest/slowest to get more of a truer number? It's something I need to look at more I think before moving on.
Click to expand...
Yes that is the problem like you say I use HRB and the OR is the available rating to anchor your speed ratings, but it is flawed in a lot of ways, from all the figures you could possibly use the winners OR is the most volatile and one that is going to be the most out of date because the handicapper puts up the winners OR, lots of times more than double figures, you also have the problem of winners who don’t have an OR yet.
I guess that one of the reasons it’s so difficult to get your ratings on an optimum scale.
 
This is the beauty of speed ratings. We have a load of numbers that are generally easy to find if you look in the right place but its the conclusions we then go on to reach.

For me, I did think about RPRs but I found from the scrapers as I don't subscribe to the RP, you don't get them for a few days afterward and it's put me off a bit. I also prefer one source of data to work from, so went for the ORs.

The difference with my approach is I'm using every runner to form the allowance, and adopting a mid point. I'm also creating an OR for each horse without one, using a formula generated from my own HRB ratings. It isn't ideal but I feel like its better than just taking the average for those that do have one, and its not complete guesswork.

It isn't perfect though. This is a race I rated at Southwell in April 2013, and most of these were debutants.

1622834712537.png
The 57/38 is an estimate of how much more these horses are capable of. My HRB formula gave this a base rating of 78, and none of these horses really got anywhere near that in the end. Split Rock himself was rated 78 by the handicapper when given his mark, but it was always beyond him, and Kitty Brown proved to be a chronic horse, on its final run as a 2yo it was beaten out of sight in a seller off a mark of 43.

In some ways its actually not so bad to give a more generous base rating, if I'd made the rating lower then these horses might have run closer to what was "expected" and it would have inflated their performances. As its turned out, none of these impressed on the clock.

Another from much more recently, this one was a 2yo novice race and all horses were debuting. I gave it a base rating of 72 to work from and a couple of these performances were pretty good for 2yos in March. Thunder Love got a TS of 68 for this performance and followed up last month too.

1622835398184.png

Later on the same card, a novice race where there was little previous to go off and no ORs. The base rating for this one was 64. Pretty impressive performance from the winner. Bonneval was assigned a mark of 76 subsequent to this and was pipped by a head. Mount Marcy hasn't been seen since but it will be interesting to see where it goes next.

1622835661149.png

1622835816710.png
 
I'm going to try play around with the RPR and see what happens. I just fear there will be alot more manual process to it. If I had any idea how to create a potential OR through HRB ratings ibhave access to then I would start there but that's a different minefield I've yet to explore. The speed ratings I have so far I'm okay with for now but always looking to improve if possible. I just need a consistent way to use the going allowances really.
 
Just a thought O Outlander AustinDillon75 AustinDillon75 if replacing the OR with performance ratings of some kind from RPR or Timeform etc as I don't include WFA with my ratings sheets and I think they do when rating, would this potentially skew my figures when looking at going allowances? Still playing around with a couple ideas.
 
db1 said:
Just a thought O Outlander AustinDillon75 AustinDillon75 if replacing the OR with performance ratings of some kind from RPR or Timeform etc as I don't include WFA with my ratings sheets and I think they do when rating, would this potentially skew my figures when looking at going allowances? Still playing around with a couple ideas.
Click to expand...
The OR figures include WFA also, WFA just another issue when compiling speed figures
 
@db1 I have my own set of HRB master ratings, from the last 10 performances of a horse, the jockey/trainer and stallion rating, I convert each to an OR. So even in a race of 8 debutants, I effectively get 24 ORs, taking the middle one, and then subtracting a certain number of lbs off for age to get a "par" figure.

So on my scale, a jockey rating of say 2.5 gives an OR of 80.
A trainer rating of 4 might offer an OR of 102.
A stallion rating of 2.5 might lead to an OR of 95.

I just take the middle one of these three, then the same for each horse, so I will then end up taking the middle figure of all the runners, if that makes sense. Unless you have a scraper and a full sub to RP then it won't shorten the process, but the flip side is taking longer to get the figures they way you want could prove very rewarding. I've still to put my actual figures into practice just yet.
 
Update: Rightly or wrongly I've proceeded with using just the OR for the time being. I've changed a couple of formula everso slightly in my sheets and added a couple extra columns to help rank performances against OR on the day of the meeting. I've redone the Chelmsford Times. My top 20 or so previously hasn't changed too much but here's the new sample. Screenshot_20210608-141922_Office.jpg
 
Top performers for Wolverhampton, 1/6/20 up until their most recent meeting on 8/6/21. That's 2 tracks completed for the last 12months racing. I can now keep on top of those for future meetings. Not quite sure yet on how to proceed with the ratings just yet, nor do I know how accurate these currently are. Only time will tell. Need to complete the other tracks first.

Screenshot_20210610-092101_Office.jpg
 
These are my top 20 for Chelmsford City over the same period. Again we have a lot in common with these figures, at least in terms of ranking, what's noticeable is that there's 7lb seperating my top 20 here (only 4lb for Wolverhampton). So long as the courses are properly calibrated I don't think its a problem, its about getting the horses somewhere in the right order for me.

1623318325264.png
 
the BHA results page gives time beaten as well as distance, you may find it handy.
last nights 6:30 at kempton.
6:30PM
THE UNIBET EXTRA PLACE OFFERS EVERY DAY HANDICAP STAKES (CLASS 4) (Jockey Club Grassroots Sprint Series Qualifier)
6f
Dream Composer (FR)
Laura Pearson
+
POS
NAME
TRAINER/OWNER
DISTANCE
SP
1st
Dream Composer (FR)
Laura Pearson
Current handicap mark: T:82
James Evans
Peter Clarke Racing Partners

1m 12.99s
15/2

2nd
Night Narcissus (IRE)
Adam Kirby
Current handicap mark: T:80
Clive Cox
AlMohamediya Racing
½ Length
1m 13.08s (+0.09s)
3/1

3rd
Ahlawi (GB)
James Doyle
Current handicap mark: T:78
Hugo Palmer
Dr Ali Ridha
¾ Length
1m 13.2s (+0.12s)
11/4

4th
Enduring (GB)
Charles Bishop
Current handicap mark: T:79
Eve Johnson Houghton
Mr Marc Middleton-Heath
Neck
1m 13.25s (+0.05s)
14/1
 
db1 said:
This is my same view on it. Actual ratings will differ but if I can get some sort of similar order for say the top 10/20 I'd be happy enough.
Click to expand...
I've got to the stage where I could spend endless hours tweaking my figures but equally I've found tweaks aren't making them obviously better or worse. There aren't too many performances where the OR seems absurdly different to the speed figure and in my other thread I've started posting a few PDF examples.

I don't think there's another AW now until Monday so a bit of time to catch up and post a few examples of how I am using my figures, throwing out bad runs and not trusting those that seem too good to be real either.

If they don't offer any value I start looking for patterns to see if there's any real glitches, but I've posted 22 May Wolverhampton sheets and from 5 races posted there were 3 top rated winners at 16/1, 15/2 and 11/4. I'll seperately put sheets up for all races since then until Monday.
 
AustinDillon75 said:
These are my top 20 for Chelmsford City over the same period. Again we have a lot in common with these figures, at least in terms of ranking, what's noticeable is that there's 7lb seperating my top 20 here (only 4lb for Wolverhampton).
Click to expand...
I just took a look at my speed figures for the AW racecourses 2020/2021, for the top twenty I had Wolverhampton 9lbs, Chelmsford 10lb, Kempton 11lbs,
Lingfield 8lbs, Newcastle 11lbs, Dundalk 11lbs and Southwell 12lbs.

To check yourself, just use the filters.

FLAT2020_2024.xlsx

drive.google.com drive.google.com

Mike.
 
Last edited:
AustinDillon75 said:
I've got to the stage where I could spend endless hours tweaking my figures but equally I've found tweaks aren't making them obviously better or worse. There aren't too many performances where the OR seems absurdly different to the speed figure and in my other thread I've started posting a few PDF examples.

I don't think there's another AW now until Monday so a bit of time to catch up and post a few examples of how I am using my figures, throwing out bad runs and not trusting those that seem too good to be real either.

If they don't offer any value I start looking for patterns to see if there's any real glitches, but I've posted 22 May Wolverhampton sheets and from 5 races posted there were 3 top rated winners at 16/1, 15/2 and 11/4. I'll seperately put sheets up for all races since then until Monday.
Click to expand...
You've got some excellent work going on in your own thread and an excellent start so far with the ratings with those winners. I'll be keen to see how they progress. Some great stuff there.

I'm happy enough with where the basis of my ratings are at the moment. The Rank v OR column I have helps easily identify which runners have performed close to or better than their ability, assuming this means those races were closer to truly run? It something I'm still looking at though.
 
db1 said:
You've got some excellent work going on in your own thread and an excellent start so far with the ratings with those winners. I'll be keen to see how they progress. Some great stuff there.

I'm happy enough with where the basis of my ratings are at the moment. The Rank v OR column I have helps easily identify which runners have performed close to or better than their ability, assuming this means those races were closer to truly run? It something I'm still looking at though.
Click to expand...
Its a curates egg of sorts for me really, if all the races are run close to OR it begs the question of just using the ORs and being done with it, what you probably want is a fair few races close to OR and then a good reason explaining those that aren't. That's what I reckon will highlight either falsely run races, races too quick to be credible (see Primo Oumo in my thread) and those where a performer has really stood out against a host of others running close to what might have been expected. I've worked to not just calibrate courses but distances at each course and I'm sure it would help.
 
TheBluesBrother said:
I just took a look at my speed figures for the AW racecourses 2020/2021, for the top twenty I had Wolverhampton 9lbs, Chelmsford 10lb, Kempton 11lbs,
Lingfield 8lbs, Newcastle 11lbs, Dundalk 11lbs and Southwell 12lbs.

To check yourself, just use the filters.

FLAT2020_2024.xlsx

drive.google.com drive.google.com

Mike.
Click to expand...

The issue for me with that top 20 is that a couple of good performances or high ORs throw it all out. I have wanted to compare the variants in performance with ORs, but if Enable runs at Kempton its got about 15lb over everything else so its not really workable.

Instead, I've compared my figures with ORs from the 21st ranked downwards. This is how it figures by track. So for example, there is a 2.9lb difference between the 21-40th best performances on my ratings at Chelmsford, compared with a 3lb difference in ORs between those.

The second set of boxes is the 25th percentile performances across each course. So if you were to say the 25% best at Chelmsford was 68.9, but the 25% OR is 76, it doesn't compare so well with Kempton, which has a 25% best of 64.9, but an OR of 78.

1623399667626.png

So course by course,
Chelmsford 76 - 68.9 = 7.1
Dundalk 73 - 65.3 = 7.7
Lingfield 74 - 67.9 = 6.1
Kempton 78 - 64.9 = 13.1
Newcastle 75 - 66.7 = 8.3
Southwell 69 - 62.7 = 6.3
Wolverhampton 72 - 63.1 = 8.9

This exercise seems to suggest that perhaps I have Kempton too slow, maybe to the tune of 6lb.

At the 75th percentile, we get this:

Chelmsford 53 - 45.3 = 7.7
Dundalk 55 - 45.2 = 9.8
Lingfield 56 - 47.2 = 8.8
Kempton 58 - 41.8 = 16.2
Newcastle 57 - 44.2 = 12.8
Southwell 51 - 36.2= 14.8
Wolverhampton 55 - 41.9= 13.1

There's definite evidence there that I could be raising all the Kempton performances by 6lb to bring them into alignment with other courses, equally, you could aim for a 10lb difference between all courses, so I could add 3lb to the performances at Kempton, but reduce the other courses by between 2-3lb.

I think I will re-rate my tracks based around this. To be honest, it doesn't bother me if the variants between top 20, top 50 whatever aren't bang on, as it starts to get into the realms of semantics, but there is some clear evidence that I may have got Kempton wrong.

EDIT - there is a far higher percentage of 2yo at Kempton than all the other courses which is clearly partial explanation, but a quick re-run applying to all horses 4yo and over suggests its still wrong.

New corrections are needed as follows:
Chelmsford - reduce by 0.7lb
Dundalk - reduce by 2.8 lb
Lingfield - reduce by 3.5 lb
Kempton - improve by 2.8 lb
Newcastle - reduce by 1.8 lb
Southwell - reduce by 3 lb
Wolverhampton - reduce by 0.9 lb.

So Kempton performances are under rated compared to Lingfield by 6.3 lb.
 
Last edited:
You must log in or register to reply here.
Back
Top