Backtesting with Fixed Size vs Fixed Ratio

Quote from moron28:

How come the number of trades in the backtest changes as the position size changes? I thought you were comparing the same trading strategy (entries and exits) under various position sizing strategies. I would have expected the number of trades under each column to be the same.
More...


Hmmm... That's a good question - I didn't notice that. I'm using the $imuscript from Wealth Lap Pro and I just ran the exact same chartscript through with the different parameters... I'll have to look into this because I would've expected the fixed risk to have a fixed Sharpe at least - even if the fixed ratio didn't...

Thanks for pointing that out. I'll post why when I figure it out.
 
Ok - so it's a Wealth Lab issue. I used the default setting of 100k. When I set the capital at 10M the same number of trades was always taken. I'm too tried now to look at the various Sharpe's and such so I'll have to do it later this week.

I searched through the forums on their site and found out why and that's why I knew to try using 10M capital:

"The $imulator does not choose trades randomly. ChartScripts create the trades and the $imulator processes them. If there isn't enough cash for all trades on a given bar, then the $imulator chooses trades by Position Priority. If the ChartScript doesn't SetPositionPriority (the seventeen-liner does, and yes, it's an integral part of the system), then the $imulator creates the priority "somewhat randomly". "

Thanks again for finding that.
 
So I guess it turns out it's a moot point. When I had a big enough position (or set commissions to 0) with more than enough capital - the Sharpe's were virtually equal. So for screening purposes only to see if a strategy is worth looking further into, I plan on doing the following:

1) Make sure I have enough starting capital relative to position size

2) Make sure the position size is not so small that commissions eat into the strategy too much. I'll probably just set the commissions to 0 for screening purposes.

3) Check the Sharpe and number of trades with fixed AND ratio position sizes and make sure they're the same so that I'm sure 1 and 2 are set properly.

4) Do more in depth strategy analysis for promising strategies.

Thanks all for your help.
 
@MustPlayOptions,

the use of Monte Carlo Simulation (MCS) in the system development or backtesting process is independent from the details of your trading system or "the art of trading your system".

With the the traditional use of the MCS you "stress test" a trading system using the existing system test results (= system simulation). You can do this with your own system, but also with professional "blackbox systems", if some numbers are given, see here - there's a pdf document on my website: http://www.zentrader.de/html/monte_carlo_simulator1.html

The advantage of this is, that you see a wider range of possible profits and more important possible accumulated drawdowns, which help to decide which size of trading capital is becessary to trade your system. The disadvantage is, that this stresstest is usualy based only on the original backtest with the given historical data.

To simulate also changing market conditions (which will come in your trading future...), you need also to simulate the test data (a process which I call data simulation, also known as data scrambling). And with the so resulting additional system test reports you have to run new system simulation stress tests.
And so on...

So in my opinion Monte Carlo Simulations are necessary and a very important process step for every trading system development process and there's no direct connection with fixed size or fixed ratio!

bye,
Volker
 

Attachments

You must log in or register to reply here.
Back
Top