Quantopian's community platform is shutting down. Please read this post for more information and download your code.
Back to Community
AlgorithmResult larger than Research memory

A get_backtest() call for a specific algo of mine fails for every backtest, because the kernel crashes when it runs out of memory.
I'm stopping all running notebooks before so I can't get any more memory.
Is there a way to reduce the size of an AlgorithmResult object so it doesn't crash? Would turning off logging or custom data help?

Thanks!

3 responses

My sense is that the scaling is with the number of stocks and the number of transactions. So, if you could can trim down these, you should be able to analyze longer backtests.

I'm guessing it should be able to handle ~1000 stocks with ~5% average daily turnover, over a ten-year backtest, but this is just a hunch.

It's 200 daily positions with 3.92% turnover over 10 years...

Hmm? Seems like it shouldn’t be choking.