Pick one market that has enough volatility to trade. Pick one indicator and this is crucial;one time period like 8:30 10:30 session that has volume and institutional trading. I think the biggest flaw in backtesting is to not test in reference to time frame. Take the YM or the Forex markets, there are very specific timeframes when meaningful movement takes place and the rest is meaningless chop or low volume spikes. Back testing the entire 24 hour forex session will produce a meaningless result.