Quantopian's community platform is shutting down. Please read this post for more information and download your code.
Back to Community
Backtest data errors

Hi Everyone,

I'm just getting started with this and wondered what the frequency or severity of expected errors in the backtest data are? I think I've found a problem with the historical data (or maybe I'm just doing something really wrong).

Essentially I'm working on a shorting algorithm and I've found that it works ok in a backtest up till the end of October 2015. Through October 2015 the algorithm is shorting EEQ @ $0.26. Suddenly on 5th November 2015 the price for EEQ jumps to $26 and the algorithm loses millions of dollars. Obviously there is a risk of this happening in reality, but when I check Reuters historical pricing EEQ was trading at $25 - $35 historically so it looks to me like an off by 100 error in the data.

Any thoughts? Is this common? Am I missing something? Any input would be appreciated!

Thanks,
Hayden

2 responses

Hi Hayden,

Looks like an error. I checked Yahoo finance and also https://www.splithistory.com/eeq/, and there's no other explanation I see.

Once reported, Q does work to fix data problems. But there is no list of securities under review or with known problems waiting to be fixed. The most recent feedback from Q on the matter can be found at https://www.quantopian.com/posts/stocks-missing-split-adjustment.

I don't know how common errors are, since there is no overall reporting other than what users share on the forum. That said, the data aren't perfect, since problems are reported from time-to-time (every few weeks/monthly perhaps, not daily).

Grant

Hi Grant,

Thanks for the feedback, that's great. Looks like theres some more related activity on that other thread so I will report it there for the Q guys to pick up.

Thanks again!
Hayden