Artificial Intelligence and Our Trading Strategy

By: Matt Garrott

The S&P 500 finished November with the largest weekly gain in 6 years.  What does this tell us about the market?  Maybe it tells us that the October sell-off was an overreaction.  More likely it tells us the same thing the October sell-off told us:  No one can predict short-term market movements.  Further, ascribing a rationale to past market movements hinders our ability to make sound investment decisions going forward.

If we assume a past cause and effect relationship, we will be tempted to expect the same going forward.  This is ridiculous, of course.  You could conclude any number of relationships out of the recent market movements.  Buy on dips of 10%?  Sell the hawkish Fed and buy dovish remarks?  Trade on trade negotiation news or auto plant closings (or possible openings if BMW is to be believed)?

Obviously, we don’t have a “trading strategy”.  We have an investment strategy instead. If avoiding permanent impairment of capital is the top priority, then activity is overrated.  As an investor’s trading activity increases, so do the odds of losing wealth.  Actively managing taxes and expenses provide clear benefits.

Maybe Artificial Intelligence (AI) can provide a lesson in trading.  In 2013, a computer scientist tasked an AI to maximize its score in the game Tetris without losing the game.  The AI found that the easiest way to a higher score was to fill the screen with blocks and then pause the game just before losing.  Hollywood AI already taught us this 35 year ago in the movie Wargames.  The only way to win is not to play the game.

Fairway Scorecard 11-30-2018