Very nice pattern of win size to loss size. Sez good things.
TA forward testing sample size?
- Thread starter illstep
- Start date
Very nice pattern of win size to loss size. Sez good things.
I agree, this looks very good. I like it very much that your win rate is not very high. If you confirm that your loses are small because you take them quick (vs sitting and waiting for them to come back or working rhem), you are golden.
thanks for replying
Yes, I can confirm that I never just sit in a loser hoping it'll 'come good'. I'm strict with my stops. They are typically like 4 NQ points or so, but it varies depending on the trade setup. Always small, though.
You sound more confident that me though in my method! haha. I'm still thinking it's just a lucky spell! Gonna stay in demo until I have an even bigger sample size
Cheers
No offense but you don't seem to have a very good understanding of how statistical significance works. Trading demo is pretty much worthless for increasing statistical significance of a rules based system. Did you read my earlier post in this thread?
If you demo for 6 months (that is a lot of time to waste not making money) it is not much different from having a backtest that is 6 months longer. This will make no difference unless your original system was so curve fitted that you can tell within 6 months that those 6 months are statistically significantly worse than expected based on the rest of the backtest. In other words, trading demo only helps for proving the system doesn't work, not for proving that it does.
And that is assuming the best case scenario where the sim is coded properly with realistic execution and slippage -- many are not.
The right way to do this is to just do a proper backtest. Learn how statistical significance works, learn walk forward optimization, learn how to not curve fit. Invest in a quality dataset with at least 10 years of tick data, make sure you understand how fills are being simulated and that it is all realistic.
How can one backtest discretionary system?
Well with care and discretion, of course!
OP wasn't completely explicit but everything about this thread suggests this is not a discretionary system. Trading discretionary based on TA makes little sense.
As opposed to trading automated based on TA!?
Was a bit dramatic with that last post lol. Took a massive drawdown at one point that day where everything was losing, but managed to get it back to only a 'fairly' red day! (still my biggest red day yet, mind)
Anyway, new account highs now. Nearly made the 300 trades, too
Maybe another 300 before I feel confident though lol
Anyway, new account highs now. Nearly made the 300 trades, too
Maybe another 300 before I feel confident though lol