How can sentiment data dictate when to be contrarian - and what's it saying now?

Many investors are aware of the idea of contrarian positioning. Bullish sentiment on stock markets has increased recently, especially among small and larger speculators, and many commercial traders are growing more cautious recently too. But how do you know when to run with the crowd versus detecting sentiment extremes to go contrarian? There are methods for detecting extremes, such as very high or very low put/call ratios. Another I've used and shown here is marking channel trendlines in the volatility index ($VIX), as well as seeing when it's stretched far from its moving averages. There are also "Commitments of Traders" data tables and graphs; plus a few others. I've included several sites featuring good sentiment data sources in the list at the right side of the page here. Readers should browse those from time to time. Today, let's take a look at one of them, in the middle of my paragraph 9 in that list:

9. Weekly COT (commitments of traders) reports - sadly, they only come out each Friday afternoon reporting on COT positions as of the preceding Tuesday, so always a bit dated: COT (Commitments of Traders). There's also COTS Timer blogspot. And check out the Option Pain CBOE (Max-Pain) Calculator from OptionPain.com.

Below is an example of a COT graph from the first site in that paragraph, the graph for S&P 500 (SP). It's usually thought that you want to position with the commercial traders, and against the large and small speculators. But if you positioned contrarian too soon, you'd be in a losing position. (And there are times when being parked in cash is good too.). How to parse the COT data? This weekend you can read up on how the COT data are incorporated into one trading model, at COTs Timer Commitments of Traders | Cracking the Market Code, located at http://cotstimer.blogspot.com/. Notice that their system gives signals far in advance - so for example, their new signal for most equities (bearish, except for the Nikkei) will kick in on November 2. You'll have to read up over there how it works. Point is, that you want to have some method for going with the trend when it's trending, but if you want to pick a trend change (like the B wave topping) and use sentiment data for that, then you'll want a method that detects extremes. Also, those extremes will have to match your trading timeframe. The COTsTimer blogspot uses a leisurely swing trading approach. So if that's your style, you should take a look at what these folks are seeing and saying now about equities, gold, and natural gas.