Can you interchange polynomial regression bands and standard deviations?

Polynomial Regression Bands (PRBs) have been around for quite some time, although I know of very few traders who use them seriously. It turns out that I am a strong advocate of mean reversion and PRBs are my main tool for determining when and how price action returns to the mean. There are many other bands in the trade; Bollinger bands and the Keltner channel come to mind. But for my purposes, I’ve customized the traditional PRB into something that perfectly fits my needs. Of course, the process of refining the variables for use with the bands was a long process that took several years until I felt comfortable with the product I now use every day.

For anyone who has taken a statistics class, you will know that a standard deviation (SD) is a measure of error. You will often see that presidential polls indicate the percentage of voters who favor each candidate, and if you read the fine print at the bottom of the page, they will indicate that the poll’s accuracy is + or -3%. This means that the survey could be wrong by 3% on the high side and 3% on the low side. So far the theory of surveys.

Traders often overbought or oversell a security at any given time during the day. When the price becomes too overbought or oversold, it usually touches the inside band, which is 2 SD away from the midline, and things start to change. On certain occasions, the price action could well go as far as 3 SD. When the price goes so far out of the line, another group of traders springs into action. These traders are called arbitrage traders and they profit by trading the security against a related security. The result is usually a return to the midline for the overbought or oversold security; hence the term mean reversion. This trading behavior is especially prevalent in ES contract where there is literally an army of arbitrage traders.

You can now see that the reversion to the mean trade is the result of several trade processes. Polynomial regression bands are my attempt to quantify this complicated chain of events and quantify when the price will move towards the mean. It is a very precise trading method. I record each trade I make and indicate the type of input I used to enter the trade. Mean reversion trades have a 79% chance of moving 6 or more ticks in profit, usually much more. However, this style of trading is often overlooked in the retail ecosphere. I have no idea why.