A New Proprietary Indicator

Click on Chart to Enlarge

This post will be the first part of probably 2 or 3 concerning a new indicator that I have come up with. I have closely followed put/call ratios for quite a while and have come up with a few insights here and there which I feel add to or expand the existing ways in which the data is used as a timing indicator. Also, in the past I have worked on developing a multifaceted indicator based off of several observations of the VIX, however, I have never got anything that I felt was quite what I wanted.

Most indicators based off of the VIX or put/call ratios compare VIX to VIX and P/C's to P/C's. So it's apples to apples. My idea with this indicator was to compare the VIX to P/C's. So more like a fruit to fruit comparison. The VIX is used as a contrary indicator, and the equity and total put/call ratios are also used as contrary indicators.

There were several observations regarding the VIX, equity put/call ratio, and total put/call ratio that led to a hypothesis regarding a relation between them that might prove useful. Some of the observations....

-The VIX tends to rise in tandem with market declines and spike at market bottoms

-The equity put/call ratio tends to rise in tandem with market declines and spike at market bottoms

-The total put/call ratio tends to rise in tandem with the market declines but often tends to diverge at market bottoms

-The total put/call ratio factors in a significant portion of "smart money" volume trading on index options, and this group may be diverging from the general trend (and dumb money) as the market gets stretched too far in any direction and hence near inflection points.

-There are ample ways to track whether the VIX is "too high" or "too low" and whether the total put/call ratio is "too high" or "too low", but I have never seen the question asked "Is the VIX too high or too low compared to the put/call ratio?"


Taking some of these ideas into account, I set out to try to answer that question. I decided to use the total put/call ratio in the study rather than the equity put/call ratio, because it seemed to me that conceptually I wanted to see if the fear premium on options was excessive or inadequate relative to the sentiment of the market at large (smart and dumb). My hunch was that while the fear premium (VIX) was hitting an extreme in one direction the smart market participants would actually consistently be taking a contrary position and thus causing the VIX to be too high at market lows (and vice versa) in relation to the total market sentiment as evidenced on the total put/call ratio.

After toying with some different ideas on how to draw this out, I decided to build on a commonly used method for evaluating the VIX which is to compare the closing value of the VIX to its 10 day average. So I calculated that out and also did the same thing for the total put/call ratio. so now I had numbers showing how stretched the VIX and total put/call ratios were from their 10 day averages. Then I subtracted the value calculated for the put/call ratio from the value calculated for the VIX. By doing this, times when the VIX was more stretched from its average than the total put/call ratio was, would lead to positive numbers and vice versa. My hypothesis was that relatively large positive numbers would be correlated with lows in stocks of some degree in that fear premiums had gotten ahead of things. The opposite would be the case for relatively large negative numbers. In other words, I was hoping to be able to identify an efficiency of sorts between two data sets that should be strongly related. Then you could take a position betting on a quick re-balancing of the data.

The chart above shows a 3 day average of the (VIX deviation - TPC deviation) spread in blue. Then there are +/- 1 standard deviation (from the 21 day average of the spread) bands around the data which help to filter out relatively high and low values. I have put a chart of the S&P for the corresponding time frame there as well.

After going through the data, it seems that only taking signals in the direction of a moving average like the 20 day average of price, will help to stay with the larger trend. I have placed arrows indicating approximate buy sell signals from the indicator, but have only put pink lines connecting successive signals in the direction of the 20 day average.

Of note is that price has consistently responded quickly to these signals. Because of this, my next post may look at the data without a moving average, and simply use the raw daily data with some standard deviation bands and then see what happens to price just 1 or 2 days out rather than waiting for an extreme in the opposite direction to exit the trade.