1/ The article does not back-test the RSI + Bollinger. It simply tries to filter the RSI by using a moving average on the Bollinger spread. This is extremely different than an RSI + Bollinger strategy.
2/ The article does not provide a trading strategy, rather, an idea to filter the RSI signals that theoretically only work in ranging markets. It is simply discussing the signals' quality. A trading strategy is much more complex than that.
3/ The results presented have shown better results by using the filter but didn't say positive results. This can only mean that when creating a strategy that uses the RSI as one of the components, it may be interesting to try out the Bollinger filter to enhance the RSI signals within the trading system.
4/ Using Bollinger + RSI that you have described shouldn't really work. If trading and analysis was that easy, we'd all be millionnaires. Exploring new horizons through proper risk management is the way to go.