Backtesting Question

Questions about MultiCharts and user contributed studies.
Algyros
Posts: 52
Joined: 28 Sep 2009
Location: Austin, TX
Has thanked: 1 time
Been thanked: 5 times

Backtesting Question

Postby Algyros » 26 Apr 2010

When I'm running an automated strategy, and I compare the daily period results with what is actually happening with my account in IB, the results are fairly similar. However, if, at the end of the day, if I close my workspace and then open it again (i.e., it goes into backtesting mode), the results for that day change (and are often markedly different from what actually happened).

In other words, let's say I'm running an automated strategy and at the end of the day, before I close it, my performance report says there is a certain profit. If I close the workspace and open the same one again and check the performance report, the daily profit will have changed, sometime substantially. And this version of the profit is often very different from what actually happened in my account. And yet, nothing has changed in the strategy (i.e., code, settings, etc. are exactly the same).

Can someone help me understand why this is happening? Thanks.

User avatar
Bruce DeVault
Posts: 438
Joined: 19 Jan 2010
Location: Washington DC
Been thanked: 2 times
Contact:

Postby Bruce DeVault » 26 Apr 2010

If you are set for synchronous operation, then the fills that actually happened in automation are reported back to your strategy in real-time, and the performance of your strategy will match the broker's fills. However, if you close the chart and re-open it, this information is lost and replaced with simulated fills, causing a difference in performance because now you're looking at simulated performance just like any other back-test.

janus
Posts: 835
Joined: 25 May 2009
Has thanked: 63 times
Been thanked: 104 times

Re: Backtesting Question

Postby janus » 27 Apr 2010

It also may depend on the data provider. IB for example will sometimes provide different historical OHLCV values to the real-time ones. In the case of V the differences can be huge. They recognize this but ignore to fix it.

So, when you run your study in real-time, then run it again after re-loading MC, you may get slightly different results.

User avatar
TJ
Posts: 7740
Joined: 29 Aug 2006
Location: Global Citizen
Has thanked: 1033 times
Been thanked: 2221 times

Re: Backtesting Question

Postby TJ » 27 Apr 2010

When I'm running an automated strategy, and I compare the daily period results with what is actually happening with my account in IB, the results are fairly similar. However, if, at the end of the day, if I close my workspace and then open it again (i.e., it goes into backtesting mode), the results for that day change (and are often markedly different from what actually happened).

In other words, let's say I'm running an automated strategy and at the end of the day, before I close it, my performance report says there is a certain profit. If I close the workspace and open the same one again and check the performance report, the daily profit will have changed, sometime substantially. And this version of the profit is often very different from what actually happened in my account. And yet, nothing has changed in the strategy (i.e., code, settings, etc. are exactly the same).

Can someone help me understand why this is happening? Thanks.

what is the chart resolution?
are you backtesting with intrabarordergeneration?
are you using precision backtesting? what is the resolution of the 2nd data stream?

from a distance,
can't tell you much with the limited information you have given.


Return to “MultiCharts”