Monday, October 09, 2006

Something unique....!

I've received a good deal of responses to my earlier two posts on this topic below. Some had sent in their bios showing interest in participating with me and Joe while others have kept themselves updated by occasional email exchanges. We welcome all and appreciate their interest.
It's not easy to set up a hedge fund with a consistent strategy that could be outlined upfront. The space is so very dynamic that to lay down a few statistical dictums for the algorithm to follow would rob the very purpose of generating above average returns to humans who are not mere statistic. So while we can use traditional risk weighted formulae to a good measure to arrive at trading strategies, it needs to be rationalised with a fair bit of reason - a difficult task indeed, but that's what we aim to achieve. That's what will set us apart.
While Joe is busy meeting potential investors to his hedge fund, I'm trying to pool together some brilliant minds to work around devising some unique trading strategies which is a combination of Math, Statistics, Market dynamics and phew....human reasoning.
Here's my take on how our eventual output will be like....
All of our strategies have a value component, a component that looks at earnings expectations, and a component that keeps us from 'arguing with the market' for long periods.

All proprietary quantitative portfolios are rebalanced monthly, which strikes a good balance between being responsive to changing conditions and not overreacting to short-term price dislocations.

Our strategies only include selection rules that satisfy two criteria: They have to to be effective in historical simulations AND they have to make good sense. The latter is very important, because statistical research alone will turn up many trading rules that are based on data anomalies and will not work going forward.

Quantitative strategies have been gaining momentum over the last decade. There are now quite a number of quant funds and money managers that employ at least partially quantitative stock selection models. Modern computers have easily enough power to sift through a large amount of information quickly, which has made some of the more complicated strategies possible.

An advantage of quant funds are that stock selection is automatically based on objectively measurable criteria because a computer cannot analyze anything else. Rule-based trading decisions take the emotion out of the process and ensure discipline. Especially the latter has been shown to be a vital ingredient of strategies that are successful in the long run.