Sunday, December 23, 2018

Trading With Hypotheses, Not Conclusions

Noticing some good buying activity late Thursday and a holding above the Thursday lows in the evening session, I started Friday morning with a hypothesis.  I believed we could see meaningful short covering and buying pressure if early weakness during the morning trade could hold above those overnight lows.  That did indeed happen, and we saw solid buying come into the market a little after 10 AM EST.

Before long, however, the buying rolled over and sellers began to take control.

Two ideas crossed my mind as this occurred:

1)  This shouldn't be happening if we're putting in a bottom.  An oversold market should generate more meaningful and sustained buying, reflecting the interest of value buyers operating on a larger time frame.  The failure to sustain buying goes against my hypothesis and suggests we haven't yet made a durable bottom.

2)  The Fed indeed was a game changer.  As I had pointed out in the previous post, the participation in the market completely changed after the Fed's announcement.  The inability to sustain buying from the oversold condition supported the hypothesis that this was a new and important input into the stock market.

In both cases, I have hypotheses and update my belief in those hypotheses given market action.  If I had traded as though these were firm conclusions and not hypotheses, I would not have been sufficiently flexible to first anticipate strength and then later trade the market's short side.  Being on the lookout for information that disconfirms our views--including trades that don't work--is essential to risk management.  It's also critical to our trading psychology.  We want to embrace disconfirming evidence, not become threatened by it.

So where do we get hypotheses from?  This is where quantitative analysis intersects nicely with discretionary judgment.  The quant work tells us what has occurred in the past under conditions that we're seeing at present.  Sometimes that work suggests the presence of a significant directional edge.  One excellent source for such studies is the Quantifiable Edges letter, which reviews potential edges in the stock market each day.  The most recent letter notes the very oversold condition of the market and goes back 20 years to examine what has occurred in similar circumstances.  By viewing the market through several different lenses (including recent Fed activity), Quantifiable Edges is a fertile source of hypotheses.

Right now, the work from QE suggests a high probability of a rally.  Rob Hanna, however, notes that, in situations when we *didn't* get a bounce under these circumstances, the downside was ugly.  Plus, Rob's work suggests the Fed is not being accommodative at present.  So right away this sets up a flexible thought process for Monday:  looking for evidence of putting in a bottom and seeing buyers take control, but also being sensitive to the possibility of ugliness and a deeper capitulation.  

With proper risk management, a hypothesis disconfirmed is not a failure; it can be the source of opportunity.

Further Reading: