Monday, December 13, 2010

The Put Call Ratio, continued...

The major market indexes are at new highs for the year, and continue to close above .8 standard deviations above the 10 day moving average; so there is no bearish sell signal yet, however sentiment according to a variety of measures has once again gotten extreme indicating that the money to be made in the next few weeks ahead will be on the short side. Further strengthening this argument are divergence in indicators or breadth and momentum.

As I've been noting a for a week or so now, the put call ratio has been signaling that we should expect a correction soon. As the 5 DMA worked it's way back to the one standard deviation mark however, I wrote that I expected another setup to emerge, which we are seeing now.


Chart 1:  Total Put Call Ratio - 5 DMA
We can see here that speculators are pouring into calls. The signal line (5 DMA) is only slightly in the setup area so it may very well get even more extreme. That being said in my experience these kinds of speculators are wrong 99% of the time. It's always amazing for me to see that after the rally, people want to bet on the upside, after the move has already happened. 

I also mention previously that I will also look at the 10 DMA of the put call ratio, which will sometimes weed out some of the false signals that can occur right before a true signal.

Chart 2: Put Call 10 DMA
Using longer time periods has it's advantages and disadvantages. While it may weed out some bad signals, it can also weed out some good ones, like the buy signal that occurred in early September. In this case however it would have weeded out that false sell signal that occurred in mid November. I actually consider that "false" signal to be part of the same setup, but back to the chart, we can see that there have only been 3 signals this year including this one. One in January and one in April, right before the flash crash. Both signals lead to very fast declines, and currently the 10 DMA is in a similar spot to where it was right before the flash crash. As always, an extreme signal can always get even more extreme. But what this does tell you is being long is no longer a smart bet.

2 comments:

  1. how does the data look using ema instead of ma?

    ReplyDelete
  2. Very similar, in fact almost identically in the same position. I'm sure there is statistically some variation in method of calculation, weighting, and time period that gives you the best signals, but after starting with the simple 5 and 10, then conducting an exhaustive search to find the "best" signal, I've gone all the way back to just using the 5 and 10 simple moving average.

    ReplyDelete

Followers