|
PureBytes Links
Trading Reference Links
|
Ross,
>AM> Well, you probably know my answer, if you read my article at
>AM> http://unicorn.us.com/trading/expectancy.html
>
>Thanks for the article Alex, I appreciate your insight.
>
>I am currently coding your version of Expectancy - I have been using
>the Ryan Jones version up to now. Am I correct in assuming that your
>version is more conservative than Jones' in that it disregards the
>outliers? Sounds sensible to me.
I didn't know Ryan Jones had an expectancy formula. I tended to
ignore anything he wrote after learning about his fixed ratio
position sizing method, which I regard as a debacle, a ridiculously
dangerous way to trade that preys upon the newbie psychological need
to be right.
Before calculating expectancy, I first toss out the largest win as
an outlier, reducing the total number of wins by one. If it's an
outlier, then discarding it makes the expectancy more representative
of the system's expected performance. If it isn't an outlier, then
discarding it won't matter anyway.
Granted, if you have *two* outlier winning trades in your data, then
this crude technique won't work.
What Van K. Tharp does also discards outliers: he has you separate
the range of wins and losses into bins, then you make a histogram
of winning trades and losing trades, and you pick the bin that
has the most winning trades and the bin with the most losing
trades. Basically the distribution of wins and losses conform
to bell-shaped curves, and you pick the peak of the curve. This
discards outliers, but is not practical to implement mathematically,
and some subjective judgment is needed to determine the appropriate
bin size.
Mathematically, the mean win and the mean loss correspond roughly to
the peaks of the bell curves, but a calculation of the mean can get
distorted by outlier data. So I just throw out the largest win. In
practice this seems to work pretty well. If one has enough data,
the median value might be more representative than the mean.
>>> Opportunity * Expectancy > 2.0 Opportunity = Trades per Year
>
>AM> Huh? An opportunity factor of 2 and an expectancy of 0.6? That's
>AM> only 3 trades per year! You want to maximize both.
>
>No, not just Opportunity - in Tharps terms this is called 'Expectancy
>score' and the calc is [(trades per year) * Expectancy].
That's what I meant. You said opportunity*expectancy>2.0 as a
requirement. If expectancy is 0.6, and the expectancy score is 2.0,
then the opportunities are 3 trades per year.
-Alex
|