Q: So you're claiming every single trade based on signals from your method was profitable?
A: Nope. There were two unprofitable signals, as can be seen in the table above: in 2011, a sell signal came in at 10.75, only to be shortly followed by a buy signal at 11.77, and at the end of
Was there only 1 unprofitable signal?
The exact targets in the OP aren't up to date anymore, those were for the first version of the algorithm. In the backtest of the version I've been using for a while now, there was one trade incurring a real loss (-17% BTC), and another one that was pretty much exactly profit neutral, but under the original assumption ("few trades because of larger position"), that would mean a small loss after probable slippage. Plus the recent 'live' loss (-16% USD), for a grand total of -16% in live profits
