On moving on today's data.

Researchers depend on on major statistical technique for determining if a comparison is valid: Regressions and their coefficient.  But the flaw in statistics, really is determining the mean, and evaluating the mean against the population.

The problem with todays business has the flaw of not enough data.  Volatility in many markets are based on the latest data.  We speed to sell or buy based on what we hear today.  

The problem with this is figuring trends. A simple way to think about it is a question we hear at the doctor.  "Do you consume alcohol regularly?"  Well there is no exact answer for that.  If you drink a pint a day, well yea.  If you drink 5 pints one day a month, well thats regular too... somewhat.  What if you drink a bottle of scotch and get hammered once a year on your birthday, or every six months? Thats regular too. The point here is that "regular" is more about volume over time than frequency.

If we use this concept to define market data, we see big changes based on the latest data.  We try to be ahead of the curve by buying or selling at the latest data point.  Such tactic is stupid. Data tends to trend over time and blips in either direction can have no effect on the average, especially if the average is based on a broad timeline.  Two percent today, could be wiped by -2 tomorrow and +1 on the day after.   

Since there is an obvious cost to trading, it would make sense to save the transaction costs and hold.