Hello,
I am testing a strategy where i lay with offset and greening in running on the horses. There are stop losses included as well. Its all fully automated via the automation tool and also linked to excel as well.
Current stakes are £5 with a maximum liability of £3.25.
i have completed exactly 200 races and i have a profit of £14.15. I believe that the strategy if proved to be sucessful could easily be scaleable by 100 x my current stake
My question is this :
Is 200 races a reasonable sample size to come to an assumption that i may have an edge. Does it need to be 1000 races ? At what point do i start to bump up the stakes ?
Any guidance on what the correct sample size should be or any thoughts or pointers would be much appreciated.
Sample Sizes
Look at the total amount staked. Look at your profit, map your cumulative profit against total amount staked. When that settles to a percentage and stays there. You probably have a big enough sample. If it's shooting around all over the place then you need it to settle to be sure you have an edge.
Thanks Peter,
i did what you recommended and the cumulative profit as a percentage of cumulative staked started off erratic but then settled to between 1.5% and 2.5% over the last 50 races.
The methodology makes perfect sense - thanks.
Feeling a bit more comfortable now that it could be a goer. Think i will bump up the stakes a bit and see how it goes.
i did what you recommended and the cumulative profit as a percentage of cumulative staked started off erratic but then settled to between 1.5% and 2.5% over the last 50 races.
The methodology makes perfect sense - thanks.
Feeling a bit more comfortable now that it could be a goer. Think i will bump up the stakes a bit and see how it goes.
The problem with spot losses in play is their reliability. Are you getting the stop loss triggered and matched every time?THENUTS wrote:Hello,
I am testing a strategy where i lay with offset and greening in running on the horses. There are stop losses included as well. Its all fully automated via the automation tool and also linked to excel as well.
Current stakes are £5 with a maximum liability of £3.25.
i have completed exactly 200 races and i have a profit of £14.15. I believe that the strategy if proved to be sucessful could easily be scaleable by 100 x my current stake
My question is this :
Is 200 races a reasonable sample size to come to an assumption that i may have an edge. Does it need to be 1000 races ? At what point do i start to bump up the stakes ?
Any guidance on what the correct sample size should be or any thoughts or pointers would be much appreciated.
Hi Steve,
You are 100% right about the reliability of the stop being triggered in play.
My strategy is lay with offset and greening. If it goes wrong i then get out if the price dips below 1.19.
Because the market reacts violently and moves incredibly quickly across a very large number of ticks when it gets to 1.19 or below the traditional method of setting a stoploss at the point of entry does not work ( i tested that and its just all over the place ). So what i do is once my lay bet is in place if it goes wrong i have a formula in the speadsheet that say if the horse i have laid price goes below 1.19 " green all " which effectively gets me out for a loss at that point but not a full loss.
it doesnt work sometimes because the price just crashes straight down to 1.04 or below but it does work a lot and i just need a split second for the match. It makes the difference between a profitablke strategy or not.
You are 100% right about the reliability of the stop being triggered in play.
My strategy is lay with offset and greening. If it goes wrong i then get out if the price dips below 1.19.
Because the market reacts violently and moves incredibly quickly across a very large number of ticks when it gets to 1.19 or below the traditional method of setting a stoploss at the point of entry does not work ( i tested that and its just all over the place ). So what i do is once my lay bet is in place if it goes wrong i have a formula in the speadsheet that say if the horse i have laid price goes below 1.19 " green all " which effectively gets me out for a loss at that point but not a full loss.
it doesnt work sometimes because the price just crashes straight down to 1.04 or below but it does work a lot and i just need a split second for the match. It makes the difference between a profitablke strategy or not.
This strategy seems quite systematic, if so it should be straight forward working things out. What is your average win and your average loss? My gut feeling is you don't have enough info, but your sample size is in "making conclusions" territory.