What back-testing does not tell …

Once again, a superb post by Ben Carlson.

10 Things You Can’t Learn From a Backtest

Those who love back-testing data should take the results with a pinch of salt.

One line that I liked the most in this post is “Unfortunately there is no such thing as a front-test.” And add to it the disclaimer from mutual funds “past performance may or may not be sustained in future.”




Extrapolation …

One of the most common approaches to estimate future is extrapolation of the past. We have been taught to identify trends and extrapolate these into the future – sometimes into the infinite future. However, many things in life, especially the financial market returns, are cyclical and they do not move on the same trendily forever.

This is where the fallacy of forecasting lies. Very often, we are able to correctly forecast so long as the trend continues. However, identifying the trend reversal ahead of time is almost an impossibility. That makes our prediction track record unreliable. We are right some time, and wrong at other times.

Be careful of forecast based on extrapolations.

#RidingTheRollerCoaster – 229

If you cannot predict, protect …

We have written time and again on this blog the inability of most forecasters to predict the exact events, especially the extent and the timing. However, we as humans want predictability, though we must learn to live in ambiguity and uncertainty.

Once we accept our inabilities, we start looking at the solutions differently.

In financial markets, the big lesson from the history is that most experts and common men alike have failed to predict the future, especially at the turning points. Majority of them have resorted to extrapolation of the past, which fails them exactly when the events take a turn.

In such a case, history also teaches us that in the absence of the ability to predict, one must protect what one has. Diversification is one such protection available to common investors.

#RidingTheRollerCoaster – 101

Analysis or forecast

Analysis and forecast are two completely different things altogether. While forecast is about the future, analysis is done of things in the past. And still, we regularly keep hearing about the “Analyst forecast”. Isn’t it strange?

Many forecasts have nothing to do about the future, but are just extrapolation of the past.

Mutual funds know it better when they say, “Past performance may or may not be sustained in future.”

To read more about this, look at the chapter “On Forecasting” in the book “Riding The Roller Coaster – Lessons from financial market cycles we repeatedly forget”

#RidingTheRollerCoaster – 93

Extrapolation, overconfidence and turning points

It is generally at the turning point of events that many experts as well as investors get carried away with the momentum and assume continuity

So often, experts extrapolate the (recent) past to predict. So long as the trend continues, they are proved right. This gives them overconfidence and they are celebrated as heroes.

Then, the inevitable happens. The events take a turn.


#RidingTheRollerCoaster – 46