If you take a bunch of completely random systems, and trade them for a couple of years you'll get a distribution in profits and losses, with an average return of 0 minus transaction costs.
If you retrospectively look at the non profitable systems, and ask what the optimum stop size should have been, you'll generally find that a tight stop would have improved the performance of those systems. These systems where non profitable because they had a lot of losses, so limiting those losses makes perfect sense.
However, if you retrospectively look at the profitable systems, and determine an optimum stop size, you'll find that tight stops are detrimental because they eliminate many of the winning trades (winning trades that where achieved by random chance rather than skill)
In my experience, most professionals are pretty clueless, and they are operating in an environment that is similar to complete random chance. If they are profitable, and they analyse their trades, they find that wider stops are actually beneficial, and tight stops are detrimental. The same argument applies to profitable amateur traders, their statistics will indicate the same thing. It's a classic case of being fooled by randomness. The majority of profitable traders fall into these categories, and in an attempt to help others, they mislead possibly with the best of intentions.
Of course, with our random traders, over the longer term mean reversion will occur and those profitable traders, will encounter a rate of losses that are above average, but with larger stops. The non profitable traders will encounter an above average win rate but with tighter stops, and of course, those using tight stops will do better than those with wider stops.
Ironically, the situation inverts once you start trading with a genuine edge.
Mr Socco is 100% correct, tight stops are the way to go.
