|
PureBytes Links
Trading Reference Links
|
On Thu, 20 Dec 2001 13:25:27 -0500
"rascal2" <rascal2@xxxxxxxxx> wrote:
rascal2> I trade the S&P futures contract. I understand the pitfalls of
rascal2> overoptimization and always optimize any systems I develop on
several
rascal2> expired contracts before trying them on out-of-range data.
rascal2>
rascal2> However, I have a friend who used a totally different approach and
with good
rascal2> success.
<<....SNIP....>>
rascal2> I know that theoretically this should not work, but in fact it
seems to work
rascal2> pretty well. His system consistently turns a better profit with
smaller
rascal2> drawdowns that it would when optimized over a longer period and
then run
rascal2> against the same data sets.
rascal2>
rascal2> I'd be interested in any thoughts anyone might have on this
approach.
rascal2>
I recently licensed the trading system development package and add-ons that
Ward Systems Group sells. The explicit recommendation that WSG makes
concerning the use of its products is that the Training/Optimization period
length be kept low. One of the "add-ons" that it sells works to compute the
"optimal" training period length and the optimal "retraining" interval.
Out-of-Sample trading system testing is performed automatically with a few
list-box choices and check box choices. One can optimize on Max Net Profit,
Max Profit Factor, Max Sharpe Ratio, Max Profit/Max Drawdown, etc., etc.
It has taken me several weeks of intensive use of the product to understand
it and it's potential capabilities because it works differently than TS.
So, I'm not in a place yet to make a definitive judgement about the
optimization question. I'm sure the answer to the question partly depends
on the tool one is using. At this point, though, I can say this: I've used
a slightly modified out-of-the-box strategy from WSG to train and
out-of-sample test a system that consistently has a 2+ profit factor both
in-sample and out-of-sample against several different futures instruments
using the adaptive optimization period length add-on.
|