Hi Guys..sorry to have been quiet.......
Have been working on handicaps averages and improvements
but, I havnt been idle!
I ran a data export of all runners from 2002...........shut the old computer up for a day or so!
I then imported the data into excel and did a bit of manipulation........warning if you try this the file generated is too large for excel!
The data snippet above shows teh results for Pivotal in each code......handicaps are unadjusted....
Available are the average and max of the handicap and the standard deviation.......70% of all results will fall inside 1sd and 95% within 2sd.
WHilst there is more work to be done I think that this data can be used to identify horses on teh improvement trail........take Entitles in the 2.00 at Doncaster.......a Pivotal........running off 89 today. Pivotal data shows that ave is 72 with sd 16......meaning that it is running at the top of the its average rating and there is only about a 14% chance of being better.......so it can be discounted.........
Obviously this needs more work but it might help to pick a winner or two in tight handicaps!
John
3 comments:
Can you explain how to use this data for the Doncaster race, ive understood a bit of what you saying, but im still in the dark?
Hi John - you make my head hurt!
Pivitol is a high class sire for fast ground animals and taking literally the ratings for Entitled show that with her conditions she can better the current rating when she gets decent ground -- if there is any merit in the Ceiling level for progeny surely the ground is the key element?
Early days.......can easily add in average going....
Rade, what I am looking for are improving horses, the idea being that if you have several horses rated 67, 72, 65, 72, say they are all similar weights.....if the first three are at the top of the offspring ratings there is less chance that they will improve........if the last 72 is at average but still has a possibility of 10lb further improvement then it would be the one to back. Knowing how much a horse is going to improve is impossible to know what I am trying to develop is a way of reducing the guesswork.....
I'm happy for any ideas of how we can develop this data....I do at least need to be able test the theory!
John
Post a Comment