Assessing trade quality
3 components of trade quality
based on a lecture given by Martin Rosenthal in FL's Corner voice chat on 1/4/06
1. mathematical quality

Average profit per unit of risk can be calculated from backtest statistics or from realtime results as follows:

Quality Index:
(P - L) / N
where P = total profit, L = total loss, N = number of trades

The resulting number is also called "expectancy." The higher the numer the better the mathematical quality of the trade. A positive number means an expectancy of gain over the course of many trades. A negative number means expect this method to lose money. It is the average result per trade including winners and losers.

Before a method is traded realtime the only data available will be from backtests, so you will only have a "theoretical quality index." After you've done the trade many times realtime you will be able to calculate the "realtime quality index." This is much more significant because it's easy to design a backtest that looks great on paper but is impossible to do in realtime trading. The bigger the series of trades used to calculate the quality index, the more accurate the result (assuming that the data is good). NOTE: In order to mean much the quality index must be based on a series of trades done according to the same rule.

(higher quality index) + (more risk units in the account)  = (less risk of ruin)
examples of risk of ruin (calculated using binomial distribution)::
Quality Index = 0.5 with 10 risk units = .003 risk of ruin (3 in 1,000)
Quality Index = 0.5 with 20 risk units = .0001 risk of ruin (one in 10,000)


2. psychological quality

This part is subjective:

  • your own ability to take the trade being analyzed
  • your own ability to keep trading through the level of drawdowns associated with the trade

NOTE: A method with lower % wins is harder psychologically and generally has bigger drawdowns, but over a large series of trades may have a much higher expectancy per trade.

3. frequency of occurence

Generally trades with a higher quality index do not occur as frequently as lower quality trades.

Frequency can be increased by:

  • adding more markets
  • adding more timeframes within the same market