Wisdom of the collective

In 1906, Sir Francis Galton came upon a contest in a livestock fair where any person could guess the weight of an ox and the winner would take home the ox.

Some of these contestants were what you could call experts - farmers, butchers and the like. Others were less experienced. He found that while the range of guesses was wide apart and no one could guess the right weight, the mean was extremely close.

This is a phenomena that has been observed and noted by various people - how groups of diverse individuals with varying degrees of knowledge and understanding of a field can still outguess experts.

It explains diverse phenomena such as the higher than expected accuracy rate of the ask the audience lifeline in Who wants to be a millionaire and the reason why markets function efficiently with prices close to the implied values in most cases.

The basis of this phenomenon lies in the fact that when a large heterogeneous group of people use varying methods to arrive at a conclusion and those methods are uncorrelated, the errors cancel themselves out.

This is a very powerful concept and has found homes in various fields industries the military and the financial sector, but it stands an equal chance of being overused or misused by being shoehorned where it does not fit.

This brings me to the point of this post - if several stock market experts state their expectations for the index level for the next year, and a pool taken from site readers also gives an answer for the same, can these numbers be trusted?

Images courtesy www.freefoto.com

Winning Streaks and Mutual fund NFOs

For reasons beyond my ken, I have been asked, repeatedly, by friends and family on the wisdom of investing in one fund NFO or another. Depending on whether I know them well or not, my answer is either a snarled no, followed by a mini-lecture, or the lecture itself delivered with a "why me" expression. I'll probably print out cards pointing to this the next time to save my mouth the wear.

First, lets get the "but this fund is available at 10 bucks, while the other that you recomment costs 260" argument out of the way. Fund NFOs aren't cheap just because they are at par. Stocks might be, especially in IPOs since the market does not have a prior valuation of that offering. A fund is merely a shell over a set of assets, and the number of units and price are inversely proportional. A fund at 10 has no history of performance, while one at 260 has been around long enough to get there and its initial stakeholders have been rewarded 26-fold.

The easiest way to pick a good fund is to find the funds that have survived (first and foremost) over a decent period, then see which ones have been consistently outperforming the market, then narrow it down to those that meets your goals (assets under management, style, charter, investment rationale, etc) and pick the best amongst them.

The truth is that in a reasonable period of time (10 years?), most funds that get launched underperform the market and very few consistently outperform the index. You could say that this could be pure chance on their part, or that historical performance is no indicator of the future.

I'd say yes, that could be true. But it very well might not be.

Lets talk about winning streaks. Pick any game (Golf? Woods. Tennis? Federer) and you will find some sportsmen and some statistics that will defy probability. Dig a bit deeper and you'll find that such is not necessarily the case. In most cases these stats are generated by men and women who dominate the sport they play, and the winning percentage for them is so high that the streak itself lies well within the realms of probability.

Take cricket and 1000 runs in a season as an example - the only players who do it are exceptional batsmen. The funny part is that many of them have done it more than once. A lot of the current lot in that list are safe bets to do it again. You'd do better to bet on them than on a rookie yet to pay his first match, if both bets were offered at the same odds.

And thats all there is to it.


Images courtesy of http://www.freefoto.com/