Is this system a fluke?

[wildshoetwt]

I'm not yet willing to give the details on it. Let's just say it's two weighted moving averages, simple crossing strategy, that's it, no risk management or stop loss, no filtering, no oscillators, no nada.

My problem is that I'm still not 100% sure if these results can be trusted. Is it possible that WHEN the moving averages have crossed, that 2, 3, 4 bars have already been printed. I assume that 1 bar AFTER the crossover has to print to bring the average across, but is it possible that with a weighted average it takes 2 or more bars to "bring the average across"? If this is true, then the data I am using to enter and exit trades is off..

If this is true, how then do I test this strategy with the proper entry/exit data?


Here are the details from thinkORswim's crappy backtesting system, using 2 contracts on the Russell 2000 mini futures over 180 days (the contract I obtained the most outsized returns from while backtesting)...equity curve for the Long strategy data at the bottom

Strategy: Long
Symbol: /TF
Work Time: 2010-12-22 - 2011-09-08
Max drawdown: $1,300
Max gain: 6,300.00
Total P/L: 110,420.00
Total 238 order(s)

Strategy: Short
Symbol: /TF
Work Time: 2010-12-22 - 2011-09-08
Max drawdown: $1,500
Max gain: 22,600.00
Total P/L: 118,960.00
Total 209 order(s)

 

[intradaybill]

Backtest for 10 years. 180 days is a joke.

 

[DDI]

Also try it on other instruments and different time frames.

A robust systems especially when based on tech analysis, must be consistent in pretty much any market in any time frame.

I have a system that gives similar equity curves on eurusd and gbpusd 15 min TFs, but when I tested it on 60 min or 10 min Tfs a complete disaster!!

 

[wildshoetwt]

long entry/exit criteria tested only:

/TF 2 contracts, daily, 20 year backtest
total P/L = $2,100,000

/ES 2 contracts, daily, 20 year backtest
total P/L $1,200,000

AAPL 100shares, daily, 20 year backtest
total P/L $120,000

EURUSD, 500,000lot, daily, 20year backtest
total P/L $140,000

GBPUSD 400,000lot daily, 20 year backtest
total P/L $150,000

various stocks = 1000% ROI over 20 years on a daily TF





============================

but this is besides the point, this is just not possible....if it is its the holy grail

i'm afraid that my backtested entrys (enter bar after crossing of MAs) are leading the actual "close of the bar" that caused the MAs to cross in the first place if that makes sense....

edit: they are not displaced

 

[shadowninja]

Forget backtesting. You can curve fit the rules. Try forward testing for 6 months.

 

[DDI]

Forget backtesting. You can curve fit the rules. Try forward testing for 6 months.

you can curve fit for one instrument in one timeframe..pretty much impossible to curve fit over a wide range of markets...

good work wildshoe, if your system is producing this over most markets then you are on to something interesting...

IMO the performance results of backtesting your system on a market that is completely different from the one that you've designed it around, are as reliable as forward testing...

 

[wildshoetwt]

BUT? my question is: how do I know if entering the BAR after the averages cross is WHEN they would have crossed? how do I know that it wasnt the close (or 2) after that brought the MA's across?

 

[amnonco]

simply calculate the averages yourself on an excel sheet.
you can export the OHLC data to excel, and calculate any MA you can think of...

i have some comments though:
1. if you are using the term "holy grail" as a system one can find - you've got it wrong. there is no such holy grail system.
2. MA systems on a 4 hr time frame will never produce the equity curve that you have shown here. so i'd double and triple check that.
MA systems will ALWAYS (no matter what time frame or MAs you use) show a (relatively) pretty large drawdown on the end of trades OR on whipsaws. due to the lag effect.