Here is a Bollinger bands backtest : https://decodingmarkets.com/backtesting-bollinger-bands/ As you can see it returns 5.54% a year on average, about half the return of a simple buy and hold approach (using the S&P500 as a benchmark). The trader should never take any indicator or trading idea for granted, he must backtest it first.
Thanks for the document. However, we still don't see how these results (moving averages) are better than a simple moving average crossover system (unless I am missing something).
The document doesn't show a test for simple moving averages but does claim (surprise surprise) his moving average is the best among several other nonlinear moving averages.
Yep, the old "Coke is better than Pepsi" routine, LOL Still an interesting study though, thanks again.
Ehlers is a bona fide DSP engineer. The material he presents is backed by solid science and engineering. As always, a caveat, anything from the world of physics applied to finance is or could be a misapplication of the physics. Still, better than Elliot Wave!
I absolutely agree with you. What I am saying is that moving averages work much better when you replace them with Bollinger Bands, that is, the average itself + X standard deviations. How you turn these into a trading system is a whole other deal and yes, back testing is essential. In the study you quoted, I honestly can't wrap my head around trading on a daily chart using a 5-period Bollinger Band. Daily charts imply longer maturation periods, which are not compatible with such a small average, which in turn implies some heavy volatility in terms of results. You simply can't beat long-term results by making several short-term trades, which have much higher liquidity. It's just how the statistics work, risk is always proportional to return, liquidity, maturity and volatility. Not sure if I am making sense, I tend to post here after a few drinks.
Just do a simple experiment. But need a charting system with daily data and a screening script. Run a scan for events that a stock can stay above a moving average in days. The scan results should have a long list of ticker symbols with number of days. Sort this list and do a frequency plot(count number of 1s, 2s, 3s and so on). You will notice not many can stay above a moving average for too long. Patient pays.
Yeah. I slap BB on my Heikin-Ashi candlesticks and trade the reversals off the bottom band and the midpt (20dma). And doing this on a weekly chart blasts the faster/shorter charts. I sometimes trade the daily or intraday charts as well if I'm bored, but nothing beats longer term performance and with little input from me, the huge metric is profit/effort.