Re: Evaluating parameters for EMA crossover trading: A rigorous approach (kind of)
Love the depth of the OP 
I backtested Bitstamp data for 1 and 6 hours candles, didn't change the fee assumption even if there is a lot smaller fee to pay (0.4% vs 0.6%) but a bit more slippage that may compensate that. One thing I noticed, the "strange beast" EMA5/1 gives me negative results for 6 hours...
EMA 20/6 is still quite good even for 6 hours candles, while hourly candles are weak in general, as I would expect. Would like to test 2h intervals at some point.
I backtested Bitstamp data for 1 and 6 hours candles, didn't change the fee assumption even if there is a lot smaller fee to pay (0.4% vs 0.6%) but a bit more slippage that may compensate that. One thing I noticed, the "strange beast" EMA5/1 gives me negative results for 6 hours...
EMA 20/6 is still quite good even for 6 hours candles, while hourly candles are weak in general, as I would expect. Would like to test 2h intervals at some point.