• 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

Arks Speed Figures

I will use the historical data I have to generate the Standard Times for each AW course.
I will then rate each run at each course since 1st Jan 2007 and will then see if the figures obatined can be used as a guide to todays races.

I don't use the Class Variance table so its not relevant but to answer your question the table is used for the race you are analysing. So if it was a class 3 race you would use the 5.5 figure.
 
Will keep checking in to see how your progressing. With the AW getting into full swing it will certainly keep you busy!
 
a lot more in depth than how i work out mine ark,but mine underperformed badly last month for some reason,and cant seem to find out why as they have done ok in past,will look more closely at yours over weekend when off work and get back to you
 
Thanks mate, sometimes I can get a bit too clever for my own good and appreciate another pair of eyes giving it the once over.
 
Hello Ark

Mordin et al are doing the same as you, they don't actually convert times to 1 mile they convert the difference between their standard and the winning time to 1 mile so a variance at 1 mile can be calculated, then apply it at per mile, so like yourself when they make a going allowance of 1 it'll be 2 over 2 miles.

Converting to 1 mile does bring a slight inaccuracy into the calculation of the extremes that is accentuated in extremes of going, not really a major issue on the AW though, plus I tend to concentrate on races up to 1m2f which is only 3f and 2f either side of a mile so it works Ok for me.

Originally I started making figures as outlined in Mordin On Time, writing the winners figure in the Raceform results section. It led me to learn about Access and a bit of VBA and how to obtain the data from RI and then the Sportsbook till I learned to scrape it from the web. I did use Mordins standards for a couple of years till I decided to revise things. I did my own standard times based on a 0-80 horse and set the figure so a theoretical wonder horse would break 100 (none have, Frankel has come the closest). I revise my standards using my figures not the winning times, it's all about them being balanced, if at a course 6f winners at all Grades are generally earning 2 pts above winners at different distances for the same grades I'll adjust the standard time to bring them into line.

My database includes adjustments as in the class adjustment table above, it's not that important if they are inaccurate, they just help with the first "guess" at what the variance is, after the results are imported a form shows the difference between the winners time and expected with the class and age adjustments, I then guess (I've become quite a good guesser over the years) at a going allowance (I’ve reversed the plus and minus as well, prefer a + meaning it’s speeding things up) and generate the winners figs. Then I compare the generated winners figs with the figs for the card and adjust the variance as necessary. I have compiled thousands of meetings and the 3.7 secs between a class 7 and class 1 isn't a bad estimate, certainly not for a first guess at a variance.

The AW (polytrack) generally allows poor horses to run closer time wise to their betters than turf. The difference between fit and healthy horses top speeds aren’t that great, the differences in ability is down to how quickly it can achieve it, how long it can maintain it and how slowly it can decelerate. The polytrack helps horses maintain their speed for longer and decelerate slower, it also probably hinders the better horses acceleration. The difference doesn’t seem to be proportional although we do get very few good class 1 horses running on it. A class 7 horse is probably a little closer to a class 1 on the polytrack than on turf.

A lot of people seem to get caught up in the mechanics of trying to create some sort of perfect and accurate mathematical formula to compile figures, arguing about what races to use for a variance and such. A speedfigure is simply the difference of the actual time verses an expected (standard) time. The best way to improve them is using the figures produced to improve themselves, both your standards and variances. The process of compiling them is more the balancing of an equation than solving one.


ATB
Steve
 
Hi Steve,

Nice first post.
Can I ask what strike rate you would have blindly backing your top rated at eachh race?

All the Best
AR
 
It’s not something I keep a record of anymore. The flat finishes tomorrow I’ll do some analysis on it next week.

I did keep track a number of years ago and the highest figure lto regularly had around a 22% strike rate, 2nd lto aprox 15 % and 3rd 12%.
Be interesting to see the strike rate now.

I don’t view them as ratings though, I would only use the last time out figure as a “top” rated for Maidens and Handicaps and as they are the races that interest me that’s the way they’re presented.

I’ll post something later explaining why I think they are useful where I use them.
 
In nearly every sporting event including horseracing there are 2 fundamentals – ability and fitness (couldn’t argue with anyone who thinks human intervention is one, at either end of the scale, brilliance to ineptitude dicisions made by humans will regularly trump the fundamentals of ability and fitness).

In horseracing distance, going draw etc are details that have more or less relevance depending on circumstance.

Ability and Fitness are both made up of a mental and physical aspect.

The mental aspect of ability is what is referred to as class, class is the ability to perform to your ability, desire and belief are probably the main components.
In some circles it’s believed a look from 1 horse to another in the stalls will mean the other will be reluctant to pass it in a race (an angry 9st on its back may persuade it otherwise). Horses being herd animals have a ranking, the foal of the number 1 mare will be treated with more respect than the foal of a lesser mare, the pair will grow up “believing” one is superior to the other. As a punter desire is probably the better characteristic to seek.

The main aspects of mental fitness are probably confidence and focus, although in humans belief can be trained and improved. Horses doing what they have done before will be generally more relaxed and confident in their present situation.

In an above post I mention setting the level of my figs so a wonder horse would break 100 and in many years this has remained unachieved. Horses seldom, if ever, get the optimal conditions to run the fastest time they are capable of. I find winners generally run about 10 pts below what they should be capable of in theory, as it tends to be across the board it’s not a major issue if you want to use figs as ratings. It does though for me indicate a problem with using them as a measurement of ability, if a horse gets a very good set of conditions it’s likely to record a figure that is probably going to be unrepeatable.

For me the strength of speed figures is in their indicating fitness, an unfit horse is unable to record a fast figure.

I also mention above I concentrate on handicaps and maidens, maidens who are fast don’t tend to stay maidens for long. In Handicaps the official handicappers do an excellent job of controlling the ability fundamental for us, especially since the handicap ranges have been compressed. Horses with recent good figs should do OK then, as they’ve shown a level of fitness and the handicapper does well making sure they don’t have to compete with horses of much better ability.

As race distances increase so does the number of falsely run races and ergo the number of false recent figs, a low figure doesn’t necessarily indicate unfitness.

My preference is to concentrate on 1m2f and less. At the extremes of my range of distances I prefer to raise the minimum class of race I’ll consider, a lot of luck is involved in 5f races and poorer horses struggle to overcome bad luck than do better ones. If you are going to use speed figures as a component of studying longer distance races I’d suggest ignoring the lower class races.

On the AW I’ll tend to consider all age handicaps up to 1m2f min 0-60 (0-70 for 5f and 1m2f) on turf I prefer to raise the min class of race to 70

For my figs the AW season starts on the 1st Nov including figs from 1st Oct. So far this season there have been 20 handicaps that fit the above criteria

1st
Ling
2.20 LOCKANTANKS L
3.50 HAWAANA (LAYALI DUBAI w13/8)
4.20 ARTFUL LADY 3rd 11/1
Kem
7.10 VOLCANIC WIND L
7.40 CAPONE W11/2

2nd
Wol
4.45 GOLCONDE L
7.15 OUTLAW TORN w17/2
7.45 LOYAL N TRUSTED L

5th
Wol
1.45 WAABEL 3rd 10/1
2.15 ATHWAAB L
3.55 MAGGIE PINK L (ARAGORN ROUGE w11/8)
4.55 PARTNER L

6th
Sou
1.40 KEY AMBITION (FAYR FALL w5/1)
Wol
7.00 SPANISH PLUME (SILVERWARE w10/1)

7th
kem
4.55 NIGHT AND DANCE w10/1
7.25 RUSTIC DEACON L

8th
Ling
1.40 MEN DON'T CRY L
2.40 DELFT 2nd11/2 (SCHOOLMASTER w5/1)
3.10 ARCTIC LYNX (MUHDIQ w4/1)
4.10 SONKO w16/1


2nd highest last time out winners in brackets

10 of the 20 handicaps have been won by horses from the top 2 LTO figs
5 of the above races have been below 0-70’s with the only winner being 2nd rated ARAGORN ROUGE

A small sample I know but evidence that it may have merit
Now if I could only figure out those Erin Numbers
 
I have had a look at how mine compare to yours Steve

1st
Ling
2.20 LOCKANTANKS L (HAAMAAT W 6/1)
3.50 LAYALI DUBAI w13/8 KELPIE BLITZ 3RD 16/1 EQUAL TOP
4.20 JACKIE LOVE L
Kem
7.10 VOLCANIC WIND L
7.40 LAST SOVERIGN L

2nd
Wol
4.45 WHO'S SHIRL L
7.15 FISH FOR COMPLIMENTS 2ND 12/1
7.45 MISTRESS SHY L (SPACE WAR W 9/2)

5th
Wol
1.45 POWERFUL PIERRE L
2.15 ATHWAAB L
3.55 ARAGORN ROUGE w11/8
4.55 PEA SHOOTER L

6th
Sou
1.40 SEHNSUCHT (FAYR FALL w5/1)
Wol
7.00 MY SINGLE MALT

7th
kem
4.55 NIGHT AND DANCE w10/1
7.25 NET wHIZZ L

8th
Ling
1.40 ISINGY RED L
2.40 DELFT 2nd 11/2
3.10 ARCTIC LYNX L
4.10 SONKO W 16/1

Not quite as impressive as your 10 winners in the top 2 but did nail four winners at 16/1, 10/1, 11/8 & 13/8 for the top rated
 
There is a little error in my above results, CAPONE on the 1st probably shouldn’t be included, it is ranked 1st by virtue of being highest rated from last season. As a “rating” it’s still top but as an example of my point the race is irrelevant.

Both applicable handicaps at Wolvs Friday meeting were won by the second ranked LTO at 11/1 and 6/1 both races highlight something some might find useful (I find it very) but I’ll save it for a later post. Anybody reading this and wondering what ratings/figures we are talking about can find them at:


Code: 
www.speedfigs.com


The next thing I’d do Ark is build a list of negatives to apply to your top rated. I have many, none of which on their own are enough to eliminate a horse.

Let’s use a few simple ones on your above top rated

1) Relevant rating must be within 30 days – if we are using them to represent fitness we want something recent.
2) Draw – for simplicity sake here we’ll use must be single figure
3) Relevant rating from a handicap (Listed an Group would be OK)
4) Distance beaten must be less than 6 lengths

Eliminating your top rated who fail on 2 of the above you’d be left with

LAYALI DUBAI w13/8
JACKIE LOVE L +30 days
WHO'S SHIRL L
FISH FOR COMPLIMENTS 2ND 12/1 beaten +6
MISTRESS SHY L beaten +6
ATHWAAB L Drawn 11
PEA SHOOTER L +30 days
SEHNSUCHT L Drawn 12
NIGHT AND DANCE w10/1 beaten +6
DELFT 2nd 11/2
ARCTIC LYNX L
SONKO W 16/1 seller

A 25% strike rate and a LSP of +18.63 (at industry prices)

It wouldn’t take a lot to find reasons to eliminate WHO'S SHIRL and MISTRESS SHY and I’d want some strong positive for ATHWAAB and SEHNSUCHT to overcome their relevant draws (which is by far the biggest negative of the 4 for me)
 
Good advice, thanks Horseplayer.
What is the third row of info on your ratings as your help only describes the first two?

AR
 
Ark, I’d like to tell you but you know how it is, I’d then have to kill you :D

They’re what I say I’ll save for a later post above. They are all to do with the OR.

I’ll post something on them tomorrow
 
On hovering over a figure you get a pop up who’s third line reads something like

70 - 70.27/0 - 63.4/5

70 – is the official handicap rating the horse ran off

70.27/0 – is the average of all the runners OH’s followed by how many not included in the calculation
Handicaps should always be 0, a 10 runner maiden with only 3 horses who’ve earned a rating will be 7

63.4/5 – is the average of the highest winning ratings of horses (over the last 3 – 4 seasons) in the race followed after the / by the number of winners


The above shows a horse who ran off 70 in a handicap with an average of OR of 70.27 so wasn’t affecting the average too much. The fact the average of the highest winning ratings for the horses in the race was 63.4 suggests it wasn’t a particularly strong 70.27 race

The example is of Mcmonagle from the 4.20 at Wolves on Friday, the 3 highest LTO figs look like

FigHorseOHAveORAve Win OR
80MCMONAGLE7070.2763.4/5
78DIMAN WATERS7076.2577.56/9
78ANOTHER CITIZEN6768.571/2

This is a 0-70 so even without knowing the average OR of this race the 76.25 average of DIMAN WATERS is obviously big drop in class, the winners average of 77.56 suggest there is nothing false about the 76 average. His own 70 fig is not inflating the average either. Importantly he performed well in the race only beaten 3l.
ANOTHER CITIZEN’s average of 68.5 is franked by the 71 winners average (although only 2 winners)
MCMONAGLE’s line suggests the 80 speed figure is a little high, enough so I’ll have a look and see if I’ve made a mistake.

DIMAN WATERS probably got a little bit of luck you need in these apprentice races (wasn’t the best of efforts by ANOTHER CITIZEN’s young jockey) but there is plenty to suggest he was going to run his race.

The 6.20 at the same meeting looks like

FigHorseOHAveORAve Win OR
76ELEGANT MUSE5957.9260.4/5
76FROGNAL6269.275.25/4
73PICCOLO EXPRESS5956.8358/10

FROGNAL is very similar to DIMAN WATERS above albeit a lower level as this is a 0-62, you do need to be a bit more cautious the closer you get to the bottom of the barrel
 
I think Fulham uses something similar in his VDW thinking for rating the class of a race, although I may be wrong
 
Let’s hope it helps him make a selection rather than a arriving at a no bet.

Think a few of the VDWers are using the average OR as a class of race. It’d be nice to be able to represent the class of race with just 1 number but you don’t get a view on the competitiveness of it, all 70’s won’t be the same. Maybe having a number for class with a 1 to 9 after the decimal point representing competitiveness might be a way to go.

I’ve actually just been reading a bit of The Golden Years this afternoon (might have to start taking an interest in these twig hoppers). I just can’t take it serious anymore, even if VDW does go on to provide some useful food for thought, when you read it knowing G Hall is VDW the start of it is just too comical.
 
Horseplayer,

Have to say I have found your posts and site very interesting.

Unfortunately in your posts you have concentrated on races I don't work on, A/W, sprints, and maidens.

I use a combination of class and speed, and can understand your thinking but I do wonder what your thoughts are on the following.

I can see if a horse has run recently within your time limit it stands a good chance of being fit, but how many horse turn out in the top class races within that sort of time span? In these races would you just dismiss the horse?

Another factor is while I agree the longer the races the less chance of it being true run, but when it is, isn't it worth taking the form as worthwhile?

Have to say I have just as good results with the jumpers as on the flat.

Ark,

I'm sure Fulham has stated he uses average OR's in many of his posts.

Be Lucky
 
63.4/5 – is the average of the highest winning ratings of horses (over the last 3 – 4 seasons) in the race followed after the / by the number of winners

while i understand the 63.4 i dont see where u get the fig 5 from or am i being thick
 
Hello Mtoto

I can see if a horse has run recently within your time limit it stands a good chance of being fit, but how many horse turn out in the top class races within that sort of time span? In these races would you just dismiss the horse?

30 days was an arbitrary suggestion. I have a general number I use which is based on previous selections, it isn’t a rule just a guideline. Days since a run is a detail and depends on circumstance, under certain circumstances, in the races I use, 21 days would be a negative for me but in a Group horse those same circumstances wouldn’t be negative at all. I do agree as you suggest it’s something that can be stretched the further up the class scale you go. It’s not something I would just dismiss the horse for.

Another factor is while I agree the longer the races the less chance of it being true run, but when it is, isn't it worth taking the form as worthwhile?

It probably is, the problem for me is there is more likely to be horses with lower figs a lot better than the figs showing. If I was interested in the longer races I’d up my minimum class to 0-80 maybe 0-75 on the AW, I’d also prefer a back up fig whilst at the shorter distances I may take a chance on just the 1.


The main reason Group races don’t interest me is I don’t do any foreign figures. If I were looking at them I’d change the tools or alter the settings I use on them. The figures I’d be more interested in would be the Highest rather than LTO. In effect it’s a switch from the fitness fundamental to the ability fundamental and you’re trusting the trainer to deal with the fitness.

My original interest in racing was the NH, nowadays even Cheltenham and the National hold little allure for me. Being a long time since I’ve applied any sort of analysis to it I’ve got no idea what works or doesn’t, I suspect fitness and ability are still paramount though.

ATB
Steve
 
Hi ArkRoyal ArkRoyal, I found this post interesting,and would like to know how you compile your standard times?,I did read that the Beyer figures were the last three quickest times averaged for each distance, do you use a similar method?
 
You must log in or register to reply here.
Back
Top