+1 I'd say worse than noise â a disaster. The problem is you cannot backtest candlesticks with context so we're talking here about a useless method. Price patterns as originally defined by L. Williams and LBR and later placed in a well-defined framework by Mike Harris are much more useful and can be backtested. Besides, candlesticks are only a sub-category of price patterns. The scanner Mike Harris has developed allows identifying price patterns with desired statistics and analyzing their robustness and portfolio performance. That's very cool. Google "Price Action Lab".
Why don't you e-mail Mr. Nison and ask him how much he has made year-to-date trading his scanner in his live personal account?
I don't recommend you learn how to trade Japanese Candlestick patterns via computer codes. Instead, learn and apply them without codes, apply then with market context and use them as secondary (confirmation signals) analysis. Simply, you should already have a primary signal or directional analysis prior to the appearance of any Japanese Candlestick signal or any other type of trade signal. Follow that trade approach and they'll be useful. Last of all, start with 1 - 2 patterns...learn them manually, master them, manually backtest them via historical data going back a few years and then simulate trade them actively for a few months minimum prior to any real money trades.
Biggest problem with learning candlestick patterns is the difficulty unlearning them after you find how useless they are. A candlestick chart serves as a visual aid in identifying supply and demand but the individual patterns seem to me to be low probability directional indicators.