Trading with Stops

This morning's rally hit my model's stop loss which is set 1.25% from the previous close. I ignored it, which turned out to not be a good move. Most writers on trading state that risk control including use of stops is more important to trading success than finding good trades. For example, notorious hedge fund LTCM had no problem finding good trades, but their risk control was way out. Same with Victor Niederhoffer's fund that blew up a little before then. Less spectacularly you can make profits and then give them away again. Which is what I was doing till recently.

But setting stops is very tricky. If they are too close to the current price they will get hit all the time unnecessarily. Too far away and you lose too much money before the stop is triggered. For the moment I am sticking with 1.25% but adding two rules:

1. If the price comes back to the stop I should re-enter the trade. The 1.25% stop means that only powerful moves will trigger the stop. If those reverse sufficiently to return to the stop then the reversal is probably serious. This rule would have added a little less than 4% to returns in 2003 the one year I have tested fully.

2. If the next day's forecast suggests a trade in the opposite direction to the current trade then you should reverse to the opposite position at the stop. This adds more than 30% of compounded returns in 2003. I know it is hard to believe that - but it is true - it lets you get in on the beginnings of powerful new moves.

This chart shows the model's performance to date relative to the NASDAQ 100 index, assuming $1000 was invested in both at the close of trade on 29 June, and assuming no commissions, no spreads, and no leverage. Given the high rate of return a little leverage could offset the trading costs anyway: