Hello - I'm wondering what kind of workarounds might exist for inflated pricing related to reverse split assets like UVXY?
The historical price appears to be upwards of $100,000,000,000 per share back in 2011 when accounting for splits, which obviously impacts trading in backtesting. I hadn't thought how this might affect other strategies (albeit, infrequently), and I'm curious if there are recommended approaches to this (beyond awkwardly inflating initial capital)?
Thanks
The historical price appears to be upwards of $100,000,000,000 per share back in 2011 when accounting for splits, which obviously impacts trading in backtesting. I hadn't thought how this might affect other strategies (albeit, infrequently), and I'm curious if there are recommended approaches to this (beyond awkwardly inflating initial capital)?
Thanks
Rename
There's no out-of-the box solution for that. In the old days, Wealth-Data had a series UVXY.20230622.S to eliminate that effect of any splits after 2023-06-22. But that series is no longer supported. (And the "S" designation in general won't be.)
You'll have to work for this one. If you're only interested in testing for one series, I'd do it like this.
1. Make a script that calculates a new BarHistory by reversing the effect of all splits. That BarHistory will have huge "cliffs".
2. Export the data to ASCII file for each section - from one ex-date to the day before the next ex-date. Name each file differently, like using the last date in the section, e.g., UVXY (from the last split), UVXY-20240410.txt, UVXY-20220622.txt, etc.
3. Using the split dates, at the same time, write a helper file in the format required for the comma-delimited dynamic DataSet ranges.
4. Create a new DataSet of your ASCII files and make it dynamic, copying/pasting the range data.
You can still use UVXY series for for indicators like Stochastics, RSI, and the like, but if you need to compare close to indicators like moving averages, you'll need to think about how to adjust those series so that they're valid immediately on the transitions.
You'll have to work for this one. If you're only interested in testing for one series, I'd do it like this.
1. Make a script that calculates a new BarHistory by reversing the effect of all splits. That BarHistory will have huge "cliffs".
2. Export the data to ASCII file for each section - from one ex-date to the day before the next ex-date. Name each file differently, like using the last date in the section, e.g., UVXY (from the last split), UVXY-20240410.txt, UVXY-20220622.txt, etc.
3. Using the split dates, at the same time, write a helper file in the format required for the comma-delimited dynamic DataSet ranges.
4. Create a new DataSet of your ASCII files and make it dynamic, copying/pasting the range data.
You can still use UVXY series for for indicators like Stochastics, RSI, and the like, but if you need to compare close to indicators like moving averages, you'll need to think about how to adjust those series so that they're valid immediately on the transitions.
Thanks Cone!
Does feel easier to just backtest with $1,000,000,000,000 for now.
Does feel easier to just backtest with $1,000,000,000,000 for now.
Your Response
Post
Edit Post
Login is required