I stripped this of everything that does not appear to be necessary (no seasonal stuff, no order cumulation, position size always = 1), and the results still look almost to good to be true. Backtest done in tick-by-tick mode.
defparam flatafter = 210000
defparam flatbefore = 090000
cx1 = cci[11]<80
cx2 = cci[21]>-90
cx3 = Momentum[6](close)>0
cx6 = TR(close) >25
cx8 = EaseOfMovement[14]<95
cx9 = SmoothedStochastic[14,3](close)<95
cx10 = PriceOscillator[5,25](close)<0.5
cx11 = LinearRegressionSlope[10](close)>-14
cx12 = AccumDistr(close)>-0.1
cx13 = Chandle [20](close)>-51
cx15 = DPO[21](close)>-50
cx16 = DPO[21](close)<50
IF cx1 and cx8 and cx9 and cx10 and cx12 and cx15 and close < ExponentialAverage[8](close) and (not shortonmarket) Then
sellshort 1 CONTRACTS AT MARKET
ENDIF
IF cx2 and cx3 and cx6 and cx11 and cx13 and cx16 and close > ExponentialAverage[8](close) and (not longonmarket) Then
buy 1 CONTRACT AT MARKET
ENDIF
Set Stop %Loss 1.7
Set Target %profit 1.7
So, probably it is just a matter of overfitting, because so many different parameters and conditions are used. Stop loss and target profit are both 1.7 % (percent !), so I don’t see a problem here. Short average duration of positions may be a problem, but I am not sure,
[attachment file=59286]
Stop loss and target profit are both 1.7 % (percent !)
Aaaah! Well spotted. I need new glasses as I didn’t see the %! Still how often does the market move 1.7% in 7 seconds? Aaaaah! Just spotted my second error – it is a German screen and an s in German is an h in English!
Stop loss and target profit are both 1.7 % (percent !)
Aaaah! Well spotted. I need new glasses as I didn’t see the %! Still how often does the market move 1.7% in 7 seconds?
This is on a 4 hour chart, and stops or targets are rarely reached.
I’ve just tested it (after putting my glasses on) and it does seem incredible – but so many different indicators for longs and so many totally different indicators for shorts rings alarm bells for curve fitting and also the quarterly graph is a lovely bell curve which also rings alarm bells. I’ll shove it on demo as a ‘fun to watch’ I think.
[attachment file=59295]
but so many different indicators for longs and so many totally different indicators for shorts rings alarm bells
I guess that’s the point.
I’ll shove it on demo as a ‘fun to watch’ I think.
No I won’t as the DPO indicator is not allowed in automatic trading.
No I won’t as the DPO indicator is not allowed in automatic trading.
Ahh, so this is another point. DPO is like zig zag, it repaints the past, and cannot be used for trading therefore. So forget about the strategy.
Detrend Price Oscillator
Here you have the comparison with and without DPO. Upper curve is without DPO, everything else the same.
[attachment file=59301]
English speaking here please!
DPO use future datas, that’s why the equity curve looks so good. We all want to know what the future will be like.