Detailed Horse Racing and Football Stats

May 22, 2014 -- 9:27AM, kenilworth wrote:

So we just back G.Lee's mountsand get a 9.6% return?

I think you are prob just trolling but..

OP asks for free stats resources and gives the example of G LEE in Class 5 races

I draw his attention to Flatstats that can provide stats for G LEE in class 5 very quickly and free

By coincidence  G LEE class 5 shows a historic profit so I remark on this and use the accepted internet method of denoting that a remark is not serious by ending it with smileys

Then you get all uppity

Fair summary?

May 23, 2014 -- 10:04AM, donny osmond wrote:

there are lots of articles about the bible code, its about secret messages hidden in the biblethen you can take harry potter books and find similar patternsyou need to know if your stats are a random pattern or a meaningful discoverygood luck

Agree

Two factors that 'distort' stats is the timeline and context of each piece of data.

For time-line, you need to decide how far to go back and the debate could go on for days.  My personal view comes from trading on the stock market and I fiddled with standard formulas to create my own perspective and moving averages.

Context is, again, up to individual interpretations.

Two basic examples from sports that I developed with someone:

stats for baseball: he ignores extra innings and bases it on the nine innings.  We batted it back and forth for a couple of seasons and found the approach.  Thus, in our ratings, teams can 'draw' in baseball.  As well, each pitcher is given an 'index' so a player getting 4 hits at 4 at bats against a player with a lower index means less than, say, 3 hits from 5 at bats against a better pitcher.

We adapted the same index for tennis which worked well up to last year but is going through a flux.  We're watching it without tweeking since, overall, if it ain't broke, don't fix it.

For horse racing, there was an example yesterday:

Alderbrook Lad was running with Joe Colliver's as jockey at Wetherby.  The pundit on Willie Hill had napped him and a listener emailed

"how can you back a horse when the jockey is 1 for 65 at the track."

This is a perfect example of when stats need to be put in perspective and all angles need to be weighed out.

If I feel a horse will run well under the conditions, the two steps for me with a jockey are if they have experience at the track and have they ridden the horse before (though the list of jockey's who can 'first time win on a horse' exists).

I had looked at Joe Colliver the last time Alderbrook Lad ran in April at the track.  At that time, Joe Colliver was 1 for 64 at Wetherby.

But look through the rides:

Joe Colliver's rides at the track have shown promise with several big priced horses making it into the top 3, often ahead of jockey's on better horses.  In Feb, he got a 94/1 shot into the frame in a ten runner race with no fallers (important, as well compared to stats that show a jockey winning when five of the ten runners pull up or fall *context).

So despite only one win, he knew the track and could get bigger priced horses around, especially in ground with give. As well, for 65 rides, only a few 'fell' and UR's. 
So, yes, one winner in 65 (May 11) but some fantastic riding.

Before April, how had Joe Colliver done on the horse: he has ridden Alderbrook Lad five times before April and produced two wins and two seconds.

So now we have a jockey who knows the track and, in my view, his experience coupled with a strong horse was worth backing.

That day when I backed the horse EW (5/1) and they fell around two or three out when leading though were under pressure. 

But the 'research' was done and I put it in the notebook which alerted me on Wednesday night.

Stats on the surface are, in my opinion, a bookies best friend.  But context of stats can tip the scales in the punter's favour and can be an important key to finding value when the price out because of their perceptions.