pytrader icon indicating copy to clipboard operation
pytrader copied to clipboard
verified publisher icon owocki

Further analysis of initial results

Open owocki opened this issue 9 years ago • 2 comments

My test portfolio was initialized with a 1 BTC deposit, and after 2 months and 23,413 trades, exited with 0.955 BTC. The system paid 2.486 BTC in fees to poloniex.

Investigate whether I was lucky or good at trading.

  1. Were my results just because ETH had a spectacular rise ( https://github.com/owocki/pytrader/issues/5#issuecomment-204221923 ) ?
  2. How would my results have changed if I was on a low/no-volume exchange?

Bonus points 3. How does this affect how we should robustly backtest our trading algos against past data?

owocki avatar Apr 01 '16 13:04 owocki

arghh.. i typed up several paragraphs on and submitted it yesterday and it's gone now. i wonder if the flaky internet on the :bus: ate it. will try to submit something new soon.

owocki avatar Apr 06 '16 15:04 owocki

If I feel especially inclined to think about statistics (which I often am) I may devise some general tests for measuring luck vs skill. If they work they’d be useful for measuring the real profitability of future versions as well as the past.

jeff-hykin avatar Apr 14 '16 05:04 jeff-hykin