Hi Lads
I testing a stratagy
in your opionion is 1.000 trades enough to base the profit and loss on ??
thanks in advance
Marc
Testing a stratagy
Marc
Depends upon the strategy.
Ive had systems of which Im confident after a couple of days horse racing and others like the greyhounds where Im still not 100 % sure after 10,000+!
Depnds upon many factors eg Strike rate coupled with ratio win/loss. If you only win £2 on a race but you could lose £2000, then you would want to test much more than 1000 times
If its the opposite, order the porche!
Regards
Peter
Depends upon the strategy.
Ive had systems of which Im confident after a couple of days horse racing and others like the greyhounds where Im still not 100 % sure after 10,000+!
Depnds upon many factors eg Strike rate coupled with ratio win/loss. If you only win £2 on a race but you could lose £2000, then you would want to test much more than 1000 times
If its the opposite, order the porche!
Regards
Peter
-
convoysur-2
- Posts: 1110
- Joined: Thu Jan 12, 2012 10:00 am
thanks for your replys...im only at 300 so ill go to 1000 and see how it is,,but looks good so far
I'd agree with what Peter said, and also suspect I have had experiences like marko236 (though probably not to that extent, i.e. very successful to start, then soon became a lot less profitable after running for a month or two).
Another thing to take into account - Is the strategy trading, or gambling. i.e. Does it depend on who wins the race?
From my experience, if it's a strategy that trades pre-play only and closes for a profit or loss before the race, then you need a lot less of a sample. I test trading strategies over a few days of races.
If you're gambling on the winner or loser of the race, then you'll need a massive sample (I aim for several thousand races, then look at the biggest drawdown and double it, and ask myself if I'm happy to withstand that kind of drawdown or not, and then set that as a maximum loss before I completely quit the strategy).
Or maybe it's somewhere in between trading and gambling, e.g. open a position and trade out in-play. Then I'd say you need a sample in between the two above. In fact anything in-play is pretty gambly because of how fast prices can move.
Like Peter says, have a look at what your maximum loss could possibly be if things went against you. If it's a gambling strategy then it's a lot more than a pure trading strategy.
Another thing to take into account - Is the strategy trading, or gambling. i.e. Does it depend on who wins the race?
From my experience, if it's a strategy that trades pre-play only and closes for a profit or loss before the race, then you need a lot less of a sample. I test trading strategies over a few days of races.
If you're gambling on the winner or loser of the race, then you'll need a massive sample (I aim for several thousand races, then look at the biggest drawdown and double it, and ask myself if I'm happy to withstand that kind of drawdown or not, and then set that as a maximum loss before I completely quit the strategy).
Or maybe it's somewhere in between trading and gambling, e.g. open a position and trade out in-play. Then I'd say you need a sample in between the two above. In fact anything in-play is pretty gambly because of how fast prices can move.
Like Peter says, have a look at what your maximum loss could possibly be if things went against you. If it's a gambling strategy then it's a lot more than a pure trading strategy.
-
convoysur-2
- Posts: 1110
- Joined: Thu Jan 12, 2012 10:00 am
thanks xitain
its a gambling stratagy,it depends on who wins the race,
Marc
its a gambling stratagy,it depends on who wins the race,
Marc