The First Law of Collaboration and The First Law of Anomalies, 01/06/06
A colleague of mine sent me a study the other day that sent me off on yet another unexpected tangent. After a couple of days inside of R looking at data, I have an entirely new class of anomalies to work with.
When a colleague sends me something that I build upon or that sets me off on a new and interesting direction, I always try to reciprocate with a relevant finding or study as way of thanks. What I give up in secrecy comes back many fold in strong relationships and flow of ideas that are the lifeblood of the creator. In the case above, I sent my colleague and friend all my R code so he could easily reproduce my work and a note that said, "Hey, look at this!"
Which leads to:
The First Law of CollaborationReciprocate.
Corollary - The First Don't of CollaborationWhen a colleague says, "Don't pass this on." Don't.
The relationship with another creator is too valuable, and the present value of future ideas too great to abuse a trust.
Returning to the new class of anomalies:
The First Law of AnomaliesA new class of anomalies adds a portfolio effect to an existing class or classes.Corollary
This is because it is unlikely that multiple classes of anomalies will decay at the same rate. Similar to not expecting a portfolio of stocks to show the same volatility as a single stock.Budget and spend time looking for new anomaly classes to create the portfolio effect w/in a set of trading systems to add robustness to overall results.
Where do you look for new classes of anomalies?
See the First Law of Collaboration : )
And that, not so coincidentally, is why this journal exists: so that we can build and expand our networks of colleagues which will increase the odds that we will continue to find new classes of anomalies and add further robustness to our work.