[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Will "trade" my great tradingsystem



PureBytes Links

Trading Reference Links

Robert,

> Most importantly notice that the system switches "on a dime", but look at
> the relation between losses and gains!
> Now almost all of these losse were ALL on the same day, I also have the
> system which simply quits for the day after 4 or 5 losses (most of you
> "matematics" will probably like that one better because of the improved
> ratios).
>
> Probably there could be some sort of filter applied.
> I'll see if I can repost results with such a filter (like I mentioned the
> system is now completely "non-optimized" and it truely is !).

Thanks for taking me behind the numbers.  I'm not a big fan of perf stats by
themselves.  They aren't wrong but they often don't show what's going on.

Seems to me that 26 trades in a day is overtrading. Probably a good idea to
stop trading after a certain number of losses since the system doesn't know
when to quit. The 26 losses total 16025 while the two following gains total
15175.  That's a ton of commission and work.  I guess there are lots of ways
around the problem:  Trade an equity curve where the system shuts down once
the cumulative losses since the last winner exceed some % of equity at the
last winner; try some kind of trade filter;  look at the distribution of
your losing streaks, pick the mean after excluding extremes and use the mean
as a threshold to stop trading until the system generates a winning trade.
I lean to the equity curve because that's tied directly to the bottom line.
The mean is just a statistic.  The filter may work today and stop working
tomorrow.

Thanks for the insight.  And I would never get as far as kicking my cat. (g)

Regards,
Mike