A custom criteria for MC optimization, the expectancy:

 

Code

if ( StrategyPerformance.TotalTrades.Count == 0 || StrategyPerformance.LosingTrades.Count== 0 ) {   return 0; } return  (StrategyPerformance.PercentProfitable*StrategyPerformance.AvgWinningTrade - StrategyPerformance.WinningTrades*StrategyPerformance.AvgLosingTrade);

It might be good to be able to optimize a strategy parameters with a Sharpe ratio, as this ratio is already known by MC, but I haven't already found if it's possible or not...

I have not done any custom criteria, would you mind explaining how to use that in MC?

Check screenshot below.



There is 18 standard criterias (Net profit, %Profitable, ...), but you can also define yours.

Awesome, thx Sam - I needed this

Mike

Hmm, Sam I had problems with your code.

I define expectancy like this:

((avgWinningDollar * pctWinners) - (avgLosingDollar * pctLosers)) / avgLosingDollar

[img]https://nexusfi.com/v/cuhyfh.png[/img]

A real world example:

avgWinningDollar = 405.88
pctWinners = 43.59%
avgLosingDollar = 209.39
pctLosers = 56.41%

The result should be:
(405.88 * 0.4359) - (209.39 * 0.5641) / 209.39

Which gives us:
0.28, a very low and poor expectancy (1.0 or above is desired).

Since my code is dramatically different than yours, I'm including it here. Thanks for forcing me to take the time to do it - needed to be done!

 

Code

// Expectancy if ( StrategyPerformance.TotalTrades.Count == 0 || StrategyPerformance.LosingTrades.Count== 0 ) {   return 0; } // (avgWinDollar * pctWin + avgLoseDollar * pctLose) / Abs(avgLoseDollar) return ((StrategyPerformance.AvgWinningTrade * (StrategyPerformance.PercentProfitable / 100)) + (StrategyPerformance.AvgLosingTrade * (1 - (StrategyPerformance.PercentProfitable / 100)))) / -StrategyPerformance.AvgLosingTrade;;

Mike

Hmm, here is another update. Seems I needed to check for some more division by zero.

 

Code

// Expectancy if ( StrategyPerformance.TotalTrades.Count == 0 || StrategyPerformance.LosingTrades.Count== 0 || StrategyPerformance.PercentProfitable == 0 || StrategyPerformance.PercentProfitable == 100 || StrategyPerformance.AvgWinningTrade == 0 || StrategyPerformance.AvgLosingTrade == 0 ) {   return 0; } // ((avgWinDollar * pctWin) + (avgLoseDollar * pctLose) / Abs(avgLoseDollar) return ((StrategyPerformance.AvgWinningTrade * (StrategyPerformance.PercentProfitable / 100)) + (StrategyPerformance.AvgLosingTrade * (1 - (StrategyPerformance.PercentProfitable / 100)))) / -StrategyPerformance.AvgLosingTrade;

Mike

Hmm, for me it's:
Expectancy = (Probability of Win * Average Win) - (Probability of Loss * Average Loss)
, and the higher the better.
But I'll check this, I'm maybe wrong.

There is a post on this here.

sam028 

Hmm, for me it's: Expectancy = (Probability of Win * Average Win) - (Probability of Loss * Average Loss) , and the higher the better. But I'll check this, I'm maybe wrong. There is a post on this here.

Hi Sam,

That post says:

Quoting 

Expectancy is your profit percentage per win multiplied by your win rate minus your loss percentage per loss multiplied by your loss rate.

Hmm. I got mine from Elliott Wave's version in NinjaTrader.

We do agree, higher the better. An expectancy of less than 1.0 means you can expect it to lose money

I wonder how it will compare to each other in actual optimization testing?

Mike

More from user Elliott Wave here:
Optimizer Type: Max. Expectancy - NinjaTrader Support Forum

Let me know your thoughts!

Mike