[masterK]

Thanks for all the replies guys, some interesting food for thought.

Has anybody any experience using the Retirement plan?
http://www.grandstand.com.au/master.htm … ement.html

It seems to satisfy a lot of my requirements in that stake increase with growing banks, yet there is an element of bank protection.

[masterK]

Yes, I always bet 2% of the bank. What happens after say a run of 20 losers is that the bank can get down quite a bit, when the winners come the stakes are low (as betting with a reduced bank) that it takes the bank quite a while to recover. In my case I had a very bad losing run early on so it took the bank quite a while to recover.

[masterK]

I guess you mean Alfie Flits, was 4th in the Aintree bumper before winning a couple of listed races on the flat

[masterK]

Sounds like an interesting program if it can scrape data – how easy was that to develop Masterk?

It’s not too bad, most of the data is fairly consistent, I parse the html as text, however some races can have some odd formatting and require a bit of messing around. The worst thing is that they keep changing the format of the pages every month or so, meaning I have to keep changing my code to get it to work.

[masterK]

Where did you get the racing data from?

Initially I used these guys:
http://www.racingsense.com/download.htm

But for the last year or so I’ve written a program to take the results down from the Racing Post website, there’s a bit of manual processing to it too.

[masterK]

After spending 3 years on and off working on a piece of software I’m still non the wiser if prediction software can be profitable. My software is a combination of form study and predictive. The form study part has always been very useful, I still can’t believe that the Racing Post don’t provide the ability to filter a horses form based on going, distance, left/right handed, jockey etc. Another thing I have automated is for any race to find all previous meetings of the horses, I can then filter out meetings that were under vastly different going conditions, distances or were just too long ago to be relevant. I’ve found these form study aids very useful over the years. I can also compare 2 horses to find any other 3rd horse they’ve both raced against (very useful when trying to assess juvenile/novice races).

On the prediction side of things over 11000 races my software shows a LSP of -400. The summer before last it excelled on the flat making a 500% profit in 4 months, I thought I was onto a good thing then last summer using exactly the same system there was a 20% loss to the starting bank.

[Author]

Posts