The Fool finished off a recent article with something to the effect that "statistics has proved that bumblebees can't fly".
As a pedantic science teacher, this sort of thing is mere sauce to the goose of my pedantry -- especially the bumblebee example, as people often use it to argue as follows:
"Science? Hah! Science proved that bees can't fly, that shows you all you need to know about science!"
So, if you don't mind, I'll pedant a little... and there is a point at the end, I promise.
Firstly, statistics strongly suggests that bees CAN fly. A statistical study relies upon much data -- and after long and exhaustive research of many bees I can confirm that bees have been discovered which can fly. Statistics therefore shows that bees can fly.
Statistics only tells us "how many". Science tries to say "What will happen next?" -- and it even tries for a Why from time to time.
So it wasn't statisticians. Who was it?
It was the physicists who "showed that bees can't fly".
Now, the more important part (as I'm a physicist and would rather see the statisticians take the blame, and having cleared them I now have a pressing need to clear the physicists).
I could take the easy route and say that it was the theoretical physicists (that's some pressure taken off) -- though out of fairness I'll have a go at defending them.
Theoretical physicists applied the equations of flight which have been so successful in designing Harriers, Stealth Bombers, 747s, Migs and so on, to the bee, and they found that the equation predicts the bee falls out of the sky.
It's important to understand that this does not mean that bees can't fly -- it is obvious that they can -- it means that the equations are wrong. This sort of thing happens all the time in science. Science NEVER proves anything to be true, not even "When I drop my pen it will fall" -- what it does is ATTEMPT to falsify. And the more a theory is NOT falsified then the more it's trusted. I have never seen a pen not fall when dropped, and so I trust the theory implicitly -- however if a pen ever did hover then the theory would be wrong, at worst, or some other factor would have been overlooked, at best. (Best and worst will switch depending whether I found the error in my theory or yours).
This is what scientists do -- make predictions and try and prove them wrong. One wrong prediction* and the theory is incorrect, though it may be 'good enough'.
For example, Newton is incorrect. Newtonian Physics makes some predictions which are wrong, e.g. that the strength of gravity diminishes as 1/r^2 -- however there is no correction to allow for the speed at which a change in the position of the mass is transmitted to the gravitational field. Nevertheless, Newtonian ideas about gravity were good enough to land a man on the moon and slingshot several spacecraft around the planets in the solar system in order to send them onward to the stars.
Newton is wrong, but in most cases he is "good enough".
Back to the bumblebee. The failure of the equation is not a failure of science -- it's a failure of a particular idea when applied to bumblebees. The equations make certain approximations which fail miserably as soon as you move the wing, but do not matter for a fixed wing (until you get to the cutting edge).
This tells the physicists that they need to refine their ideas and try again -- they did, and can now deal with cases where wings move (if I recall correctly) -- and this will eventually lead back to a further understanding of aircraft... until the next time a theory mispredicts something. Science, in short, would not have progressed without looking for failure, and being honest about it when it happens (though human nature being what it is, hoping you don't find it in your own work but in other people's!)
The equations are "good enough" for many examples. We now know more about their limitations, but for most people the equations will remain "good enough".
How does this all tie in? Well, the Foolish Four mechanical investing approach has been tested 39 times, if memory recalls, and failed 5 times (in the sense that it produced losses). The widely diversified tracking strategy failed 8 times.
Both theories of picking stock are 'false' in the sense that one counterexample, that of falling wealth, has been found. Fortunately, like that other false theory, Newtonian gravity, one of them (or both, or neither) MAY be "good enough" for a particular purpose (in this case, becoming rich beyond our wildest dreams).
In the interest of fairness it should be remembered that the FF methods were produced with benefit of hindsight, and if they hadn't worked historically they would not have been considered. It IS possible that the method is flukish for your set of data, as, probably, are other methods -- the real test will be a long term comparison starting where the data was not known in advance. Unfortunately this would take at least another twenty years, ideally longer -- and most of us cannot afford to wait and see. Especially as however many years we have with spectacular performance we can never guarantee that the method will always work next year (though we might gain confidence).
So, we take our best guess, go with what we believe will be "good enough" and hope that the bumblebee does not fall to the floor too often -- and if it does, we must be prepared to revise our ideas.
* Assuming that they didn't make a silly error in the calculation and took everything into account -- most discrepancies are due to not taking things into account, like that faulty wire or that bit of friction.
Fool USA
