profitseer's time tested e mini system

I'm not talking about what the system "would" or "should" do. I was just posting the results of my backtesting, because I thought some hard facts might be of interest here in this discussion. Here are some results:

test period: 6/13 - 8/2
ES02U 1-min: Net Profit -1493.60 $
ES02U 5-min: Net Profit 12781.89 $
ES02U 30-min: Net Profit 4635.60 $
YM02U 1-min: Net Profit 9456.89 $

If you're interested I can post the script, and you can simply try and see for yourself.

 

Just out of curiosity, can you add a trend detection script? Like...a 20 EMA on a daily chart? If you tested assuming holding overnight...that probably wouldn't work for too many people. This is interesting stuff. Thanks for the systems test.

 

if it is in easylanguage... you can post it. I will run it on european futures.

alain

 

'got you willibald.
OK, i was replying in the spirit of the thread (since it is about profitseer's system).
However, your contribution is more interesting in my opinion because you actually provide the COMPLETE system and test. profitseer is evasive about several aspects which imho will not work based on my experience.

send the code over !
tntneo

 

Here it comes:

var Start, Bar, NumberOfTicks, StopPriceSeries: integer;
var Tick, StopPrice: float;

EnableNotes(false);
StopPriceSeries := CreateSeries;
Start := 20;
NumberOfTicks := 10;
Tick := 0.25;
BuyAtMarket(Start, 'First Trade');
for Bar := Start to BarCount - 1 do begin
if PositionLong(LastPosition) then
begin
StopPrice := PositionEntryPrice(LastPosition) - (NumberOfTicks * Tick);
@StopPriceSeries[Bar] := StopPrice;
Dec(NumberOfTicks);
if SellAtStop(Bar, StopPrice, LastPosition, '') then NumberOfTicks := 10;
ShortAtStop(Bar, StopPrice, '');
end
else
begin
StopPrice := PositionEntryPrice(LastPosition) + (NumberOfTicks * Tick);
@StopPriceSeries[Bar] := StopPrice;
Dec(NumberOfTicks);
if CoverAtStop(Bar, StopPrice, LastPosition, '') then NumberOfTicks := 10;
BuyAtStop(Bar, StopPrice, '');
end;
end;

PlotSeries(StopPriceSeries, 0, #Red, #Dots);


It is written in WealthScript - sorry, I don't work with TradeStation. The program you need to run the script can be downloaded here:

http://www.wealth-lab.com/cgi-bin/WealthLab.DLL/home

You can't use the web version because intraday futures data are not available on the web site. But the demo version (which doesn't alllow saving of chartscripts) should do the job.

 

thanks willibald.
the system seems to work. I tried ES.
but it is very sensitive to the candle size (it's a remark, not criticism).
the win ratio is 20% and risk reward is 2.35 (this is the effective ratio after actual trades, meaning you win $2.35 for each 1$ you lose)
on 67 trades.

it takes longs and shorts alternatively. it is a stop and reverse system. it's intriguing to look at..
would need to try it on longer periods of time for more accurate backtest.

some churning at times. today for instance (aug 12) it's constantly losing, trade after trade. the market is trendless and chance did not pick the right direction first.

anyway, intriguing idea. simple is good.
tntneo

 

on the "philosophy" front...the old saying that its simple but not easy comes from the fact that experienced profitable discretionary-ish traders work with a complex model of the market in their heads but only require simple triggers to let them know whether their model is, in their view, "correct" during the life of a trade. this is their edge.

their embedded experience lets them use simple systems/tools that objectively backtest horribly on continuous sets of data but that they know when NOT to use or when to use with lesser or greater size.

thus, don't listen to anyone who tells you some simple indicator works for them - if they really are following it systematically with no other adjustments/money management rules etc. then they will eventually get found out by the market - if they don't you couldn't replicate what they do anymore than you can think their thoughts...



Q1

 

Willibald,

On your test of the 5 minute bars, are you moving the stop up one tick every 5 minutes (as opposed to each minute as in profitseer's original rule)? Are you using the same 10 tick stop or does it require more room on the 5min? Does the system just "stop and reverse" after the original entry? Do you have any performance statistics like average trade or drawdowns? I can code this in TS to get those if you don't have them.

Its interesting that the test results were not good on the one minute bars since that is the original timeframe given.

Thanks for posting your test results.


EDIT: Oops, some of those questions were answered by tntneo and Willibald in their last posts while I was typing this...thanks guys

 

In all time frames the stops are moved 1 tick with each new bar. The script and my testing is just a first try on this. There is of course much room for optimization.

Statistics for ES02U, 5-min:

All Trades Long Trades Short Trades Buy && Hold
Net Profit 12 781.89 $ 2 639.60 $ 10 142.30 $ -7 727.70 $
Profit per Bar 4.47 $ 2.33 $ 5.89 $ -2.69 $
All Trades 152 76 76 1
Avg Profit/Loss 84.09 $ 34.73 $ 133.45 $ -7 727.70 $
Avg Profit/Loss % 0.18% 0.08% 0.28% -15.17%
Avg Bars Held 18.80 14.93 22.67 2 877.00
Winning Trades 33 (21.71%) 12 (15.79%) 21 (27.63%) 0
Gross Profit 26 037.00 $ 9 897.70 $ 16 139.30 $ 0.00 $
Avg Profit 789.00 $ 824.81 $ 768.54 $ 0.00 $
Avg Profit % 1.69% 1.80% 1.62% 0.00%
Avg Bars Held 73.76 77.33 71.71 0.00
Max Consecutive 4 2 3 0
Losing Trades 119 (78.29%) 64 (84.21%) 55 (72.37%) 1
Gross Loss -13 255.11 $ -7 258.10 $ -5 997.00 $ -7 727.70 $
Avg Loss -111.39 $ -113.41 $ -109.04 $ -7 727.70 $
Avg Loss % -0.24% -0.24% -0.24% -15.17%
Avg Bars Held 3.56 3.23 3.95 2 877.00
Max Consecutive 13 19 9 1
Max DrawDown -5 042.70 $ -4 146.10 $ -3 090.10 $ -13 462.50 $
Recovery Factor 2.53 0.64 3.28 0.57
Profit Factor 1.96 1.36 2.69 N/A
Payoff Ratio 7.07 7.49 6.87 N/A
Standard Error 1 194.72 $ 1 041.92 $ 881.73 $ 1 514.99 $
Risk Reward Ratio 0.87 0.08 1.09 -0.58
Sharpe Ratio of Trades 4.73 2.62 6.18 N/A

 

Thanks Willibald