Risk/reward. Loss 1: Profit 2 or more?
Maybe he does, maybe he doesn't but it's his strat. Talking about an indicator strategy doesn't necessarily convey that particular trader's aspect or knowledge of the market they trade, they might not even know themselves why they took a trade in a high probability area. It is never as simple as, "oh the MACD's overbought and we hit the top bollinger band, so I'm selling".
Don't get me wrong, a lot of people use bollinger bands but the market doesn't just bounce off a 20:2 bolly band because statistics say it should 95% of the time - it's all in context. Largely, instruments react to price and news...surprisingly!
If you really want to model it why not go with the trend instead? Now...how do you model a trend in an EA? You could use multi timeframes, you could use ADX, you could use angle of MAs, you could use fibs - that can all be programmed but it's not easy.
Trading in from the bolly bands is a mean reversion of sorts but you need to know when not to get in - what filter could you use for that (rhetorical question)?
With bollinger bands try creating an indicator that places a tick mark or cross every time the 3 standard deviation is crossed and only trade those back inside, higher percentage trade but doesn't happen very often. Then try if the trend is up, only trade bollinger band bounces when the bottom 20:2 SD is hit - ie trend is up but the price is very oversold so buy it cheap.
You could go on forever, but this is what you need to backtest if you are going down that route.
