Why would you need to go further back than that? You want to backtest it through the '87 crash? The '29 crash?
I would be happy with data going back to the 2000 crash. Is there a reason that I should not want that data? I’m planning more of a buy and hold strategy with some risk controls so i need to see what my potential drawdowns are.
I agree with you. The more data you have, the better. Having tons of data doesn't mean you have to use tons. You never know what ideas you'll have in the future that would require data you currently aren't interested in--but are glad you acquired it and didn't discard it. I doubt Amazon, Facebook, Apple and the like are throwing away data saying, "meh...we don't need that much data."
There is a %daily change data file for UPRO simulated back to 1986. I downloaded it and tried to convert it to Close data. I just took an arbitrary # (100 I believe) as my starting close value and then applied the daily changes to it to simulate UPRO so that I could upload it into my software which is RealTest. When I run a simple test on the data using 200 day sma, it doesn't match up with the results that I get from using Portfolio Visualizer which can handle daily %change data. Any ideas how to get this to work better?