The first screenshot is a strategy running 5 years of daily data with QQQ as benchmark. The APR and Exposure for the benchmark look correct.
The next image is the same strategy running 5 years of 1-min data. You can see the APR and Exposure are wildly incorrect.
This image is the equity curve with the 1-min data, which shows that gains are not being calculated for the benchmark until the very end of the run. This would explain both the APR and exposure results.
The next image is the same strategy running 5 years of 1-min data. You can see the APR and Exposure are wildly incorrect.
This image is the equity curve with the 1-min data, which shows that gains are not being calculated for the benchmark until the very end of the run. This would explain both the APR and exposure results.
Rename
First screenshot
Next screenshot
Next screenshot
You should arrange your historical providers so that the best intraday provider appears first in the list in the data manager. It looks like your benchmark is coming from a provider that can only provide a limited amount of data 👍
@Glitch you nailed it.
I am using IB for historical 1-min data, but I am trading live with Tradier. So I had put Tradier at the top of the historical provider list. That worked fine for backtest positions and performance, but apparently the benchmark uses the historical provider at the top of the list. Tradier only has 1 week of 1-min historical data. For the purposes of benchmark history, WL ignored the "Obtain data from selected DataSet only" selection but used if for the regular backtesting.
Thanks for the solve!
I am using IB for historical 1-min data, but I am trading live with Tradier. So I had put Tradier at the top of the historical provider list. That worked fine for backtest positions and performance, but apparently the benchmark uses the historical provider at the top of the list. Tradier only has 1 week of 1-min historical data. For the purposes of benchmark history, WL ignored the "Obtain data from selected DataSet only" selection but used if for the regular backtesting.
Thanks for the solve!
Your Response
Post
Edit Post
Login is required