Hello, Few questions 1. Where do you find trading strategies to program and back test? 2. Do you manual backtest first the strategy for about 50 trades for understanding of the strategy or just go straight to programming? Thank you
+1 "noggin" Evaluation? By the time the idea has coalesced into long/short/when/where, it has already gone through a gazillion cycles of 'edge-shopping'. You're either doing this, or you're wasting (valuable) time.
It does. Two things to bear in mind, in this context ... (i) The more you're aware of and able to discern "the kind of systems that might have something in them and appear to be based on sound underlying principles and are potentially worth backtesting" (and that's a skill one gradually develops by understanding price action and market behavio(u)r and by gaining experience both in terms of understanding statistics/probability and of screen-time), the less need you'll have, to "find systems to backtest" because all the time you'll also gradually be acquiring the ability to develop your own, and (unlike the stuff you can dig up from other places to backtest) that - when it happens - will be something that actually suits your own style, wishes, needs, abilities, techniques, time-frames, trade-durations, risk-management parameters and all the rest of it (without all of which it will have extremely limited if any value to you anyway); (ii) The more readily backtestable (is there such a word? yes, maybe ...) a system is, i.e. the more automable (or as they perhaps say in America "automatable"?) it is, the less likely it is to have a genuine edge - I'm not just saying "that's how it tends to work out": I'm saying much more than that, I'm saying that there are directly causal and reliable reasons for that being how it works out. So it's another +1 for "noggin", from me ...
Aren't there essentially two systems only? - Trade WITH the direction that price has been travelling. Or trade AGAINST the direction that price has been travelling. Both work if you respect their inherent structural "rules" and use a risk management protocol (that might be common to both). The evidence to support one or the other is what you then look for. But I'll bet most of us have already decided which we'd like to do, and we skip the charts that don't fit our bias. Then we just need to find a good enough argument to get in at this lovely price and set a stop at that lovely price.
In discussion with other traders. All great trading strategies should come from your head. Testing what you found in the internet is a waste of time. Remember that the goal of any strategy is minimize variance of your portfolio (i.e. risk). Don't be lured by good profit and neglecting risk. Its the main reason why most traders lose.
1. Data mining based on a basket of rules 2. No. The specific strategy is only discovered during the backtest