Speed Ratings – 4

In my final post for today...

I am going to look at how the Average Race Times vary for different types of races.

I'm coming into this blind as I really don't have any pre-conceived ideas.

I'm sorry for the inundation of emails today but I felt that it was easier to split the logic up into small chunks to make it easier for me to handle and to allow you to respond to individual "concepts" in a manageable way for me.

So, lets see what the numbers say...

Race Type	Races	% of all Races	ASF	Variance
Unknown	453	13.44%	12.71676408	1.23% faster
Amateur Race	17	0.5%	13.42718137	4.29% slower
Apprentice	93	2.76%	13.1077474	1.81% slower
Auction Stakes	4	0.12%	12.56392857	2.42% faster
Claiming	132	3.92%	12.93814258	0.49% slower
Classified Stakes	24	0.71%	13.0724709	1.53% slower
Conditions Race	33	0.98%	12.47466198	3.11% faster
Handicap	1661	49.29%	12.89773023	0.17% slower
Maiden	675	20.03%	12.87396174	0.01% faster
Novice	12	0.36%	12.46003571	3.23% faster
Nursery	168	4.99%	12.86146119	0.11% faster
Selling	98	2.91%	13.01006692	1.05% slower
Total	3370	100%	12.8752988	0%

ASF stands for Average Seconds per Furlong.

Variance is the difference in the ASF and the ASF overall.

What stands out for me is the following...

1). 74.30% of all races are races type where the ASF is within 0.2% difference of the Total ASF.

These races are Handicaps, Maidens and Nursery races.

That 0.2% difference equals 0.257505976 seconds a furlong difference; and remember that a furlong is 220 yards.

To be honest, I'm not sure how significant that is.

2). The really wild differences in ASF come from statistically insignificant races (Amateur and Novice races.)

Conclusion

My inclination is to not include Race Type as a filter (it can always be added in later) as it would then overly complicate the generation of the Horse Speed Ratings.

My thinking is as follows...

If a horse starts off running in an Apprentice race and then moves on to a Maiden before running in a Seller, we''l have horse speed figures for the first two race types but, by the time we get to rating the horse for the Seller, all we can say is that the horse has run in two 6 furlong races at Lingfield (over Standard going) and on both occasions, finished with an ASF of X percent of the ASF of those races.

Sometimes you can build too much complexity into a system.

As it is, we'll end up with Average Seconds per Furlong Standard Speeds for the following combinations...

Course,
Distance,
Handicap (or not),
Month of the Year and
Class.

I will be able to display that data for every AW race.

For each horse I'll be able to display its Average Speed per Furlong over its last 5 races (and with a plus or minus sign to show the difference between its lifetime ASF. In other words, is it improving or not.)

Then we'll be able to look at how that relates to the actual results.

For example, we'll be able to see the correlation between the horse with the highest ASF of the race and the actual winner of the race.

We'll also be able to look at the performance of any horses running with an ASF higher than the ASF of the race.

And, finally, we'll be able to see how that ties in with the R4 form based rating.

And that could be a very useful added angle.

And, this time next year, Rodney...

We'll all be millionaires!

Well, I can dream!

Sometime tomorrow (if all goes well), there'll be a summary post (and, who knows, maybe the actual Standard Times.)

As always, your thoughts and comments are welcome.

My kindest regards