Can Performance Be Improved By Cutting Back After Large Winning Or Losing Periods?

Posted by Art thoughts on February 5, 2017 in Bids and Offers

     Can trading be improved by cutting back or stepping aside following unusually profitable periods? This is in line with the theory that any account that performs unusually well (or poorly) is sure to revert to its normal mean.

      It’s a seductive notion, but like many similar ideas that seem like they’d have to work, it doesn’t.  At least not in my experience.

     Part of the problem is timing.  As my late mathematically gifted partner used to say whenever I’d throw an idea at him, “quantify it.”  So how would you define an excessive winning streak?  Would your “step-aside” point be an average yearly profit plus x additional dollars?  Would you make the threshold more dynamic—use a percentage excess amount instead?  And if so, what dollar amount?  What percentage? 

     As always, research could give you the most optimal numbers.  It would be an especially difficult strategy to construct unless you were going to resort to a crude look-back in the trade by trade window (which would get pretty complex if you were combining strategies).   That’s been the clumsy not-so-mathematical approach to which I’ve had to resort.  Again, the theory seems so can’t-miss. 

     But no reward, and I’ve seen all the variations.  I’ve put trade lists in spreadsheets.  I stipulated no trades after a small number of losses—two, three and so on.  I’ve also used small total dollar amount losses.  I made the both numbers larger.  It’s easy to group cells together—the sum of the last 40, for example.

     The vexing reality is, when you have a system with a positive expectancy, that overall gain will come in lurches and starts.  The better systems will have minimal such bunch-ups, while systems that generate all their profit from an isolated timeframe (like a crash environment) are as good as useless.  To some extent, however, count on it—there will be strings of wins and losses that will frustrate this exercise.

     How about the flipside then?  If sporadic streaks are the nature of a trading timeframe, should we put the pedal to the metal after a defined revenue surplus?  The same erratic win-loss sequence thwarts us in this direction as well.

          There is only one definable aspect you can get your arms around in a winning trading system, and that’s the long run.  A single trade means nothing.  Mechanical traders are happy to leave that to the gut traders.  The latter group can only consider the trade in progress—do they take profit now or let it run more, do they add to it, where is the cry uncle point on losses.  The best traders among us will still be wrong roughly as often as right—they’re successful because they effectively manage their trades in total. 

     A small bunch of trades isn’t much more definable that an isolated one, which is probably a main reason the scale-back strategy doesn’t work.  Any trade may win or lose.  If your methodology is reliable, let it play out without trying to pinpoint when.