My back test results
- Thread starter Huskeez
- Start date
Some big things that back testers don't think about are commissions, slippage and interest earned on the account (the latter is particularly important if you are trading futures where so much of the account isn't being utilized for margin).
Try running your system through a worst case scenario to see how it stands up.
Set your commissions to $20 per trade or $0.01/share
Set slippage to some % of ATR
Set interest to 0.00% (since that's pretty much what you'll earn)
Good luck.
Try running your system through a worst case scenario to see how it stands up.
Set your commissions to $20 per trade or $0.01/share
Set slippage to some % of ATR
Set interest to 0.00% (since that's pretty much what you'll earn)
Good luck.
This is no good Huskeez. Certainly survivorship bias but maybe that's "some."
Good calmar ratios are APR/DD should >3 for world class models or 2> for very competitive 1 for something worth using <0.5 probably wouldn't bother and especially 10/70 from where you started I wouldn't even look at. Your commissions and slippage don't appear reasonable either, and I'd be upset if I couldn't get those returns in 5 years with less than 20% drawdown.
Wow, thanks for the info!
Im still working away , I am able to achieve a calmar ratio of .5 which is a lot better then the .14 I achieved in my first backtest posting.
I am beginning to think I need better software to back test more "specific" details rather than [Buy when Close price crosses through XXSMA] [Sell when Close price crosses below XXEMA]
Can you give me any helpful tips ?
These backtests were performed on daily bars with an average holding time of 2 weeks.
Quote from Huskeez:
The commissions are set to $15 per trade each way, which is the brokerage fee here in Australia which is found in the bottom right corner of the results I uploaded. I cant adjust the slippage nor the interest %
If that's the case, then I would:
A: Find a different method of backtesting which will allow you to control the "environment".
or
B: Seriously handicap your results. What I mean by this is adjust your CAGR (annual rate of return) downward and bump up your average or worst draw down.
Determine what MAR (CAGR/worst draw down) you would like to have at a minimum.
If you're test has a MAR of 0.5 (20% / 40.0%), then haircut your CAGR by 30% and bump up your worst DD by 50%. Now, you can assume in the highly variable world we occupy, that real world results may afford you a MAR of 0.23 (14%/60%)...Not that great.
By the way, looking at your test results, a max DD of 74.9% is pretty big. I suggest a gut check to see if that's something you can actually sit through and continue operating the system perfectly, day in day out. Easier said than done.
Also look at DD duration. Your equity curve looks like it has a few 2-3 year DDs. Ask yourself if you really won't be tempted to try a new system at the 23 month mark. Basically, if your system hasn't generated a return in 700 days, you might feel like it has stopped working or you might not have the wherewithal to continue.
Best of luck.