Hi Darky,

No we don't mind, ask all you like, hopefully your questions and answers and comments in this thread will help others.

I noted that your runs of outs corresponded to Paul's 'prediction'. It isn't that hard to predict when you are working with maths. A certain strike rate will produce certain numbers in terms of runs of outs and other probabilities. These figures, no matter how unattractive, are mathematical law and those that think they don't apply to them are in for a shock.

As far as the bets sample size of your method, the greater the number the more likely you are to know its actual performance level. To give example look at our daily Recent winning highlights. Our IR rating has an over all strike rate of about 26.9% over 50,000 races (top pick), however looking at the beginning of this month we had some low days (Moee pointed out one of them) and the last few days they been in the high 30s and 40+ today. What Im saying is that the method will even out at the 26.9% given a large enough sample because that is its underlying performance point. There is no question in my mind about that and I have the confidence that the rating will do just that, and it does every time. In fact if we have a few bad days you can bet that the next few will be well above average it has to happen in order to level out at our performance point.

What you have to do is establish what your underlying performance point is, and the only way to do that is with a large enough sample. The confidence that this will give you will be invaluable when you are looking at an extended run of outs and wonder if something is broken or what the problem is. We all go through those emotions. The tendency is to tweak! Add a few filters, or more rules and change the mechanics, there by changing everything including your confidence level in the method and the risk that you are willing to endure. Usually this is followed by a further few tweaking sessions before the whole system is thrown out and a new grail is sought. That's the way most punters work.

With a large enough sample you will have the confidence to say "well that is within normal and while Im not excited about it it will return to its underlying performance eventually".

Kind regards

__________________
Michal - Ratings2Win Pty Ltd
R2W Axis - Axis is Australia's leading horse racing software and database;
with sophisticated form analysis tools and accurate horse performance ratings for TAB meetings.
http://www.ratings2win.com.au/