SUBMIT


  close
Hi everyone! After spending about a year learning algotrading (and trading overall - mostly Forex) I've came to a few conlusions that I'd like to discuss with more experienced algotraders. From what I've seen the most popular approach for TA based strategy development looks like that: get some data, throw some TA tools on it, look for patterns, optimize, backtest, repeat. With some effort it is usually possible to find a strategy that will perform quite well in backtests over some limited period of time. Out of sample testing may show that such strategy is profitable during particular market conditions. With this approach it is guarranted that the strategy will stop working once market condtions change (and it will inevitably happen). It is obviously possible to be profitable with this approach, but one would have to constantly discover new strategies and update old ones. Don't you think that instead of working on particular strategies one should be more concerned about developing a global framework for managing strategies, reoptimizing them, discovering new ones and discarding the obsolete? This may sound cliche but I think that such approach may be much more productive in the long run.