The perfect mechanical trading system. - page 8

New comment
 
Xeon, thank you. I thought so :)

Gentlemen, to the question of optimisation and fitting and the influence of dealing centres on this.

Here are the results of a run with the same parameters at my place (Teletrader Demo server).

 
dmitriy писал (а):
Xeon, thank you. I thought so :)

Gentlemen, to the question of optimisation and fitting and the influence of dealing centres on this.

Here are the results of a run with the same parameters at my place (Teletrader Demo server).



Can we see a more extended report?
 
dmitry, note that TralingStop=5 here. Of course, with such a close stop the effect of noise will be very high. By the way, my Liteforex does not allow such close stops at all. My stop orders are at least 10 pips away from the market. As for the ticks issue in general, most likely, the ticks will have to be written into a file and the optimizer will be fed with them. Yes, I do not like it very much myself, but the task is to beat a very concrete broker in a very concrete and small time period. I think it is better to optimize by the exact tick chart.
 
Optimisation Report
MACD Sample2


Passage Profit Total trades Profitability Expected payoff Drawdown $ Profit %
1 28187.15 2393 11.34 11.78 136.00 0.66
2 20507.20 1722 8.32 11.91 135.00 0.75
3 14958.37 1264 6.91 11.83 157.73 1.21
4 11036.64 868 5.60 12.72 170.73 1.30
5 8446.64 654 4.77 12.92 167.73 1.31
6 7188.28 532 4.22 13.51 177.98 1.32
7 6123.41 485 3.91 12.63 187.98 1.30
8 4972.16 419 3.18 11.87 226.98 1.67
9 4524.16 382 3.13 11.84 231.98 1.73
10 4071.41 358 2.92 11.37 226.98 1.74
 
eugenk1:
dmitry, note that TralingStop=5 here. Of course, with such a close stop the effect of noise will be very high. By the way, my Liteforex does not allow such close stops at all. My stop orders are at least 10 pips away from the market. As for the ticks issue in general, most likely, the ticks will have to be written into a file and the optimizer will be fed with them. Yes, I do not like it very much myself, but the task is to beat a very concrete broker in a very concrete and small time period. I think it is better to optimize by the exact tick chart.
I have been writing ticks for a long time, but I haven't learned how to feed the optimizer with them yet. Maybe there's some tried-and-true methodology? "Throw me a link, please" (my favourite quote)

xeon:

Can we see a more extended report?
I pinned it twice and it didn't stick :?
I'm going to try again. It didn't work. :?
 
eugenk1
TralingStop=5


TralingStop optimisation from 5 to 50
 
xeon:
eugenk1
TralingStop=5


Optimize TralingStop from 5 to 50
Trailing stop = 5 pips - this is a real pip and we don't consider it :)
 

I gave optimization options for TralingStop with parameters from 5 to 50 points in increments of 5

here is a graph with TralingStop = 30 with the same other parameters



And ultimately this is not the point, I just wanted to show that by optimizing for historical data you can achieve quite acceptable results. The other matter is that an EA optimized in such a way (according to unverified assumptions) will work (for example, on a demo account), but not for a long time.
Maybe someone will want to check this "hypothesis".
 
If you optimise over a shorter period, for example a month, you can achieve even better results ...
 
xeon, sorry, I've been a bit distracted by my main work. Optimisation should be done in a day, not a month. A week at most. The algorithm is as follows: ticks accumulate. At about 2 a.m. (or on Saturday) the optimizer switches on and recalculates parameters for the next trading period. I will try to check your parameters on my quotations (liteforex).
123456789101112
New comment
Reason: