First back testing...

[schizo]

I'm not the guy you're replying to but I found back tests utterly useless since I'm not purely systematic. So for mildly discretionary (I can't figure out how to code/train the entry / exit patterns) experimenting with statistical hypotheses is more fruitful. Generally, is looking at the past helpful? Absolutely. But I put way more value on forward testing with either paper trading or live trading than I used to.

Whether you're discretionary or mechanical, it all boils down to one thing: PRICE PATTERN.

Find the profitable pattern and stick to it, come rain or shine. However, in order to find a pattern, you need the help of a backtest.

 

[nooby_mcnoob]

Whether you're discretionary or mechanical, it all boils down to one thing: PRICE PATTERN.

Find the profitable pattern and stick to it, come rain or shine. However, in order to find a pattern, you need the help of a backtest.

I would agree with this wholeheartedly. I just wouldn't necessarily make it an automated backtest. And I find my price patterns (usually) by experimenting with the data.

 

[schizo]

I would agree with this wholeheartedly. I just wouldn't necessarily make it an automated backtest. And I find my price patterns (usually) by experimenting with the data.

Price pattern need not be limited to chart patterns. It could be a pattern of price flow, eg. DOM. But you need a price pattern nonetheless to become a profitable trader. And I want to stress that it must be "consistently applied".