Closed BuddhiLW closed 11 months ago
well backtesting doesn't validate this (mainly for performance reasons, as dataframes there are potentially WAY bigger than in live) - but effectively, what you did is invalidate all backtest results you've ever had with this strategy.
The following section in your strategy rewrites the original prices. This means 2 things:
as a result - all results will have been running on top of the heikinashi candles - not on the real candles - essentially changing all assumptions backtesting makes.
heikinashi = qtpylib.heikinashi(dataframe)
dataframe['open'] = heikinashi['open']
dataframe['close'] = heikinashi['close']
dataframe['high'] = heikinashi['high']
dataframe['low'] = heikinashi['low']
None of the strategies here teaches you to do this - and you really shouldn't. It's however not a bug in freqtrade / this repo - but in your strategy - one you can pretty easily fix by changing the name to "ha_close", ... (for example).
Thank you very much. Now, I understand my mistake!
Describe your environment
python -V
)pip freeze | grep ccxt
)freqtrade -V
ordocker-compose run --rm freqtrade -V
for Freqtrade running in docker)Describe the problem:
I'm running my strategy live and it warns that there is a mismatch in dataframe's close prices:
Steps to reproduce:
Run this strategy in drymode for any market.
Observed Results:
What happened? In live mode, the strategy won't actually enter any trades.
What did you expect to happen? To run as in the hyperopt and backtest.
Relevant code exceptions or logs