Comparing testing results

Back to topics list To post a new topic, please log in or register
avatar
7
windypuff 2015.05.17 13:09
 

Hi there,

 I am trying to compare a couple of EAs based on results in the strategy tester. I attach the results (test64 versus test92). Note that test92 is broken into two tests, but when you combine them you get the same timeframe as test64.

In summary, test64 generates more profit over the (26 month) timeframe (£87K versus £74K) but with a lower profit factor (1.38 versus 1.59). Maximal drawdown percentages are the same (19%) but test64 has a much higher relative drawdown percentage (43% versus 25%). Test64 also suffers from a huge maximal consecutive loss (almost £20K, versus £5K for test92). And yet it still generates (a lot) more profit over the same timeframe.

I'm struggling to compare the test results to decide which is best. Maximal percentage is the same and yet that massive consecutive loss is a concern.

All thoughts gratefully appreciated.

Thanks,

Andrew 

 test64

test92a 

test92b 

 

avatar
37
waterhorse 2015.05.18 07:24 #
 

1) Do you notice the "Mismatched Charts Errors" in Red? You will need to fix this first.


2) Please check your objectives in trading. If you are talking about easy money of 87K or 74K gain, then you will need to tap your head three times. If it is too good to be true, then it probably is not real.


3) Instead of the initial deposit of $10,000, why don't you buy a lotto ticket for $5? You will have a better chance at the $1,000,000 grand prize instead of these EAs. 

 

avatar
7
windypuff 2015.05.18 10:11 #
 

Lol. Er - thanks waterhorse. Sort of.

 I agree that a large number of mismatched chart errors is bad. However, since I have imported all the history data from my broker I'm not entirely sure how to fix. And a few thousand chart errors (given I'm running tick by tick) over a 26 month period doesn't seem too bad. For me, it shouldn't greatly affect the result. True?

I'm not sure about "easy money" - my perception is that many people spend years working on EAs (and then constantly tweaking once they have something that generates - modest - profit since the markets are constantly changing). So not that easy. And most fail. The £87K gain is just a number based on a certain number of lots - it's irrelevant. The only important aspect is the basis for comparison against other EAs with the same parameters (lots, spread, timeframe etc.).

 

 

 

Back to topics list  

To add comments, please log in or register