Back Testing
How many 'back testing' you do before going 'live' with a strategy. I have some sort of strategy and analysed around 100 matches trough Fracsoft /Betfair historical data/Practice mode, but is it enough? Or do you go 'live' with lower stakes/liability? Any tips?
imo It doesn't really matter if you have back tested it over a 1,0000000 markets,
it still gives no guarantees it will work going forward.
Edges come and go all the time it's very much a moving target im afraid.
the best way is just to go for it with a small bank and if it works it will quickly grow and if not you've not lost much and can go on to try the next idea
.
it still gives no guarantees it will work going forward.
Edges come and go all the time it's very much a moving target im afraid.
the best way is just to go for it with a small bank and if it works it will quickly grow and if not you've not lost much and can go on to try the next idea
.
When the variability of your returns don't vary much in % terms then you probably have a bit enough sample. But back testing doesn't really tell you much because there is a tendency to see patterns in historical data that don't really exist.
If you do go forward with a strategy then set up a control account to measure relative performance.
If you do go forward with a strategy then set up a control account to measure relative performance.
Nice Blog, thank you, consider me enlightened.
When you said set up a control account, I thought you ment use a seperate account for that strategy so as you could seperate it from your main account and this would make it easy to review. I was not sure if Betfair would allow you to do this, it would be nice if they would.
When you said set up a control account, I thought you ment use a seperate account for that strategy so as you could seperate it from your main account and this would make it easy to review. I was not sure if Betfair would allow you to do this, it would be nice if they would.