View Full Version : Stats Approach
gunny72
17th February 2012, 05:10 PM
I like to apply stats to my ratings. I am looking at ways of correcting for the number of starts a horse has had be it too small or too large.
I would like to get the mean and standard deviation of number of starts for a larger sample than I have at present if someone can help.
I have found this for a recent 7 weeks of racing in Brisbane and found the mean to be 21 and the standard deviation to be 14.
This is just a ball park stat because of my small sample size but it suggests that some adjustment should be made for horses that have had fewer than 7 starts (mean-one stdev) and for those with more than 35 starts (mean+one stdev).
I would appreciate any ideas and better data on this.
gunny72
Chrome Prince
17th February 2012, 06:03 PM
Gunny,
I assume you are looking at number of wins or win percentage to add to your ratings.
I have found that a multiplier factor works best.
For example 1 win from only 1 start you might multply by .80 but 5 wins from 5 starts you might multiply by 1.50.
This is just an example, you could get more accurate multipliers by looking at impact values.
gunny72
17th February 2012, 08:55 PM
I want to give a pro-rata rating to horses that fall outside the mean+/-one sd range. e.g. my ratings of all horses rated (not just winners) have a mean of 50 so if a horse has 5 starts and a rating of 60 then that horse would have a final rating of 5/7*60 + 2/7*50 = 57, using my above results.
woof43
17th February 2012, 09:16 PM
Try using a rolling avg based on starts to find the number of starts that are required when your sample reaches the normal lifetime stdev.
also you may want to try folding the data in a way that your example horse who has had only 3 starts and you need a 12 races before you have confidence.
This 3 start animal may have rated 45 most recent start, 48 , 50
you might feel the most recent is more important so you
would create a synthetic stdev
45,45,45,45,45, 5 times
48,48,48, 3 times
50,50 2 times
and so on an you move to a better stdev.
or if you have a good database
you select all runners that have exactly this type of avg rating for 3 starts, then look at that group and see how they performed their next start and make that type of adjustment to todays expected rating
Raven
17th February 2012, 11:07 PM
Gunny, whats your email
gunny72
19th February 2012, 11:44 AM
check my profile
gunny
lomaca
19th February 2012, 11:52 AM
check my profile
gunnyWe are not allowed to do that on this forum, and thank God (and management) for that too.
gunny72
19th February 2012, 05:54 PM
Agree, I a.m reluctant to reveal my email in a forum too
lomaca
19th February 2012, 06:29 PM
Agree, I a.m reluctant to reveal my email in a forum toogunny72 there is a very easy way out that we all use, open a free account like hotmail or similar and give that one out.
Once you are happy with the other forumite doing the right thing, you can both agree to give the proper email address to each other privately, or just use the free one.
Anytime you are unhappy with the correspondence you can just ditch the free address. You can also put anyone on the ignore list as well.
Open a free account and let people help you, you won't regret it I assure you, some really nice people out there.
Good luck
vBulletin v3.0.3, Copyright ©2000-2025, Jelsoft Enterprises Ltd.