Results of Backtesting>> Profitable system!! What other tests should I perform?

Discussion in 'Trading' started by Iwilldoit, Jul 12, 2016.

  1. I recently completed a manual backtest of the M6E contract, where N= 145, Win Ratio = 68% , Ending equity balance is $1675.26.

    Period tested: Nov 5 2015 - Feb 3 2016 (and plan on extending to the current date)

    One contract

    It does seem too good to be true....are there any tests I could conduct to verify how accurate the metrics are?

    I should not have to say this, but, please do not ask me for the Rules of Engagement.

    Thank You,


    @userque @Wallace @Marco Picos @RadarGuy
    @Handle123 @Redneck
    @NoDoji @Buy1Sell2 @aquarian1
    @Gueco @pak @slugar @zbestoch @dratsum
    VPhantom and userque like this.
  2. K-Pia


    Did you factor commissions ?
    It's 2 to 3 ticks commissions per trade.
    Otherwise trade 6E directly. Factor Costs & see.

    What's N ? 145 trades ?
    What if win ratio drops to 50% ?

    Well. Try to implement & see if there are bias, inconsistencies.
    Last edited: Jul 12, 2016
  3. Handle123


    So you have 3,000 sample size?

  4. Most of the trades were done in market replay....

    N means the sample size , if you know anything about statistics.

    Implement in demo?

    I designed it for M6E, b/c that is what my account can afford

    @K-Pia @Handle123
  5. TradeCat


    Testing means you know how to make money in theory. Real market conditions don't work that way.
  6. Your biggest risk at this point is that you've been lucky and think you've found the Holy Grail.
    Nov 5 2015 - Feb 3 2016 is not a long period of time and does not cover every possible market condition.
    If you think your system has legs why not forward test in demo or even better live with 1 contract.
    Since the M6E is basically EUR/USD why not trade it live in a micro forex account. The markets are correlated so your system should work in both.
    Last edited: Jul 12, 2016
    VPhantom, Xela and louis_w like this.
  7. Bullseye ;):cool: -- I agree,
    I equate testing/paper scientists....and real world traders as the ...well, real-world hero.

    Each time it's a unique cocktail of different environments and scenarios and variables in play collectively.
    No amount of backtesting can prepare you for this. You have to have it in you.
  8. You might want to see if market conditions in general were favourable during that period. If I recall, that was part of a mini-bull run era. You could try finding the worst market conditions and run it there (say Jan 2016 to April 2016). Of course, longer periods would be good too :)
  9. Chris Mac

    Chris Mac

    Bravo ! You are at stage 1. You got a plan and want to apply it. Better than 50% of traders.
    Stage 2 : test in real market conditions
    Stage 3 : it doesn't work? why?
    Stage 4 : adjustments
    Stege 5 : it doesn't work ? why ? I got a new plan, back to stage 1 or my plan works, go to stage 2

    And after weeks (proba : 1%), or months (proba : 2%), or years (proba : 3%), or never (proba : 94%) :
    Stage 6 : my process works.
    Stage 7 : yes it works but I don't care
    Stage 8 : I lose money
    Stage 9 : I am smart : go to stage 6. Or I admit my process works but my mental doesn't work.
    Stage 10 : I need to be disciplined but I don't care. Go to stage 6
    Stage 11 : I need to be disciplined and I am working on it.

    After several weeks (proba 1%), or months (2%), years (3%) or never (proba 94%).

    Stage 12: I am disciplined and I make money.

    Last edited: Jul 13, 2016
    SimpleMeLike likes this.
  10. Zzoom


    Most important stat is Profit Factor: (Total Profit)/(Total Losses)
    Second is Max Drawdown% : 100 x (Max Drawdown)/(Account Balance At Start Of Drawdown)

    Win Ratio is irrelevant, although a good psychological prop if trading "manually".
    #10     Jul 13, 2016
    Slope Trader likes this.