Tuesday, May 04, 2010

Core Ideas in Trading Psychology: Identifying Historical Patterns in Markets

One of the core themes that runs through the TraderFeed blog is the importance of identifying historical trading patterns in the markets. I owe an appreciation of this theme to the influence of Victor Niederhoffer, whose blog and books have added greatly to the market literature.

The key idea is that, as a trader, you want to think about markets like a scientist. You make observations, you formulate theories about what is happening in markets, you express those theories as hypotheses, and you test those hypotheses with specific trades that you place. Over time, your trading experience either validates your market understanding or contradicts it, supporting or leading to modification of your basic theories.

We know that historical observations of market patterns can help generate successful mechanical trading systems. Less well appreciated is that those observations can generate hypotheses for discretionary traders. Knowing, for example, that in 17 of 20 recent occurrences where the market has made an X day low it has ended up Y% higher in the next X days does not, in itself, necessitate that you take that trade. It does, however, help you frame a trade idea if you perceive that we are in a correction within a bull market (your underlying theory).

Should the historical patterns hold, you might gain confidence in your assessment of the market's strength and trend. Should the pattern not hold, you now have concrete evidence that the market is not living up to its historical script. That could suggest that the market trend is turning: some unique factors may be at work in generating recent returns. As a scientist, you are benefiting from hypotheses that are disconfirmed as well as those that are confirmed: losing trades that were placed with a positive expectancy may be providing unique market information.

When traders identify multiple historical patterns that are independent but that point to the same anticipated market outcomes, that can provide an added measure of conviction to trades: those are strong hypotheses.

Among resources for identifying historical market patterns are the excellent Quantifiable Edges, Market Tells, Market Rewind, SentimenTrader, MarketSci, Vix and More, and CSS Analytics sites. If you check out the blogrolls for those sites, you'll see many more good resources.

If you have an interest in testing historical patterns, do investigate the resources at the excellent
Vertical Solutions site. And if you're a do-it-yourself type, the Trading Coach book has a chapter devoted to using Excel to identify historical market patterns.

For more on this theme, check out the posts on Trade Like a Scientist: Parts One, Two, and Three.