Peter in his writing is offering his general analysis, thoughts and opinion about the markets.  But he never says "go short tomorrow" or "go long tomorrow".  Because he's not trading forecasts and opinions, even his own.   Meanwhile his systems are not forecasting anything and not thinking anything -- they are just going about the business of executing algorithms.  If what peter forecasts actually occurs, his systems will have him on the right side of the trade most of the time -- better than 90%.  If his forecast is wrong, his systems will still be on the right aside of the trade most of the time, even though his forecast will be wrong.  If he never wrote anything contrary to what the systems are doing he would never be able to write a forecast or opinion, but could only say where his systems are right at the moment.  And also, some will be long, some will be short at the same time depending on time-frame.