|
PureBytes Links
Trading Reference Links
|
Owen wrote:
<How well do MAE and MFE work in real time? They've always seemed to me
like the kind of thing that would seem wonderful on historical data, but
then blow up in your face the moment you tried to use them--over-optimized
by definition.>
Owen,
I have been actively using the concepts of MAE and MFE for a while now. I
guess the danger -as always- lies with the person who works with them. One
must not allow oneself to use them blindly- if the concentration of trades
which might suggest a stop level for your position is not large enough then
don't use it. In my experience it is very important that your trades show
some distinctive patterns, i.e. trades which ended up as winners versus
trades which ended up as losers should be graphically distinctively
different. You should be able to see this clearly on a MAE chart. These two
patterns should not blend too much into one another. If they are not
clear-cut you might indeed run into the problem of over-optimization. Watch
that the results are not too different while changing the levels a little to
the left/ a little to the right. I suggest you do not attempt to fiddle
around with it in order to squeeze out an additional profit and to re-run
your analysis every quarter or so to check if the analysis is still valid,
or rather if your typical trade still looks similar than in your test
period.
In my experience both concepts work really well. But I certainly agree that
they will not work like magic for every system. I love to use them for short
term trades (up to 10 days), but have found that they are invaluable for
long term trades as well, certainly on the MFE. One other suggestion which
helps me to analyse the 'look' of a typical trade: Export the P/L of all
closed trades into Excel and create a bin chart. I do this every few months
and compare the bins. You should have a similar pattern to your 'original'.
Kind regards
Markus
|