Ralph Vince Test Metrics - Discussion

[Roscoe]

Is anyone interested in discussing the test metrics that Mr Vince describes in his book "The Mathematics of Money Management", specifically the metrics in Chapter 1 "The Emprical Techniques"?

[sluggo]

I don't really see very many "test metrics" in that chapter. What I think I see is

• The runs test for serial dependency of trade outcomes
• Pearson's coefficient of serial correlation of trade outcomes
• The Geometric Mean of trade outcomes
• The Optimal f for a set of trade outcomes

Not covered in that chapter are some aspects of real-world trading that, in my opinion, are important

1. Trades last longer than 1 day and equity fluctuations occur all throughout a trade, not just at the ending day when you know the trade outcome.
2. More than one trade is active at any given moment.
3. Marked-to-market accounting means that the profits (or losses) of a trade which is still underway, can increase or decrease the amount of money available to fund a new entry signal. This turbo boost effect is invisible if you only look at trade outcomes.
4. Humans are emotional beings whose behavior is influenced by pain and/or euphoria. There really is a maximum drawdown beyond which, real humans will quit trading. They are defeated, disgusted, and miserable. There really is a maximum profit beyond which, real humans will take some (or all) chips off the table and not continue blindly parlaying their optimal bets into the trillions of Euros.

[Roscoe]

Hi sluggo.

I would like to focus on the following metrics: Geometric Mean in it's unweighted form (ie. without the inclusion of f in the HPR formula) and in it's weighted form (which does include f) as well as optimal f and Estimated TWR.