 |
EdL
2013-02-11 11:35:24
|
|
Hi,
I have run the simulator on 2 basic trading systems and the resulting Sharpe Ratios from the backtests are -7.068 and 32.244.
Apart from these appearing to be extremely low and high (respectively), as mentioned in this previous post http://www.quantshare.com/title-908-sharpe-ratio-of-backtest-simulation-aapears-to-be-too-high ,
there are also conflicting results when running the exact same trading systems using the Combined Trading Systems tool available from Simulator -> Tools dropdown -> Combined Trading Systems
When I select the 2 systems and press Analyze, the resulting Sharpe Ratios are now -13.2 and 26.22!
The Performance, Drawdown and % Winners metrics are all the same, it is just the Sharpe Ratuos that are different.
So my question is what is the reason for this difference? Or is it a bug?
Thanks
Ed
|
|
|
 |
QuantShare
2013-02-11 12:30:25
0
|
|
Hi,
Each time you run a backtest, the report is saved in your hard drive (if it is not big) so that you can load it quickly using "Simulator Manager -> Tools -> Trading System Report".
The combined tool uses that report instead of creating another backtest.
The changes you see can occur if for example, you update your trading system, run the combined tool (before a new backtest) then backtest the new trading system again.
|
|
|
|
|
EdL
2013-02-11 12:43:32
0
|
|
I run a standard backtest and then run the combined tool without updating anything and I get the results I described. Where are the reports stored on my hard drive and can I delete them and see if that makes a difference? Currently I am just getting different Sharpe Ratios when running these 2 tools when they should be identical.
Thanks
|
|
|
|
|
QuantShare
2013-02-11 12:48:32
0
|
|
To get the report path, click on "Help -> Trading Object Paths -> Get Path -> Trading System"
The report has the following name format:
[Trading System Name]_metrics.dat
Please delete it then try again.
|
|
|
|
|
EdL
2013-02-11 13:10:58
0
|
|
I have followed the instructions, deleted the file, re-run a backtest and then immediately run a combined system test on the same system but the problem still persists.
I have also tried this on several trading systems and they all have the same issue.
When you run a normal backtest followed by a combined trading systems backtest do you get the same Sharpe Ratio?
|
|
|
|
|
QuantShare
2013-02-11 18:19:49
0
|
|
Yes, I do get the same Sharpe Ratio.
Please send your trading system to support [at] quantshare [dot] com, so that we can check it.
|
|
|
|
|
EdL
2013-02-11 19:03:47
0
|
|
I have emailed over the trading system in question, the corresponding 1min bar data and a screenshot showing the 2 different Sharpe Ratios
Thanks
Ed
|
|
|
|
|
EdL
2013-02-12 19:22:39
0
|
|
Best Answer
The QuantShare Team responded to my email with the following:
Hi,
I discovered why you are getting different sharpe ratio.
The combined tool uses a default trading system template to make calculations.
In your case, your trading system uses 0% as risk free rate. The simulator uses that. However, the combine tool uses 5% as risk free rate.
This will be fixed in the next release. Thank you for reporting this.
Regards,
The QuantShare Team
|
|
|
 |
Dave W.
2014-02-18 20:58:56
0
|
|
Hi. This is an old thread, but I'm getting a somewhat similar issue.
In my case, I've run the Trading System Combination tool on the same 4 systems on two different days. The combination tool gives different sharpe ratios on day 1 than on day 2. The systems are the same (super small differences) and the annual return, win %, and drawdown are the same (fractional differences), but the sharpe ratios are very different.
Attached are screen shots.
https://drive.google.com/file/d/0BxSjaSuq_9wgLUZlbkIzZFplM0E/edit?usp=sharing
https://drive.google.com/file/d/0BxSjaSuq_9wgTWRKeU9STGJPUUk/edit?usp=sharing
|
|
|
|
|
QuantShare
2014-02-19 12:17:20
0
|
|
This can happen for many reasons. I do not see any issue here.
Just check detailed stats for each system to see how they are different.
|
|
|
|
|
0
