Chip Anderson

The $CPC Intraday Data Dilemma

$CPC is our symbol for the CBOE Put/Call ratio.  It is a market sentiment indicator that shows the total number of "Put" options (bearish) divided by the total number of "Call" options (bullish).  That means that it is an "inverse" indicator - i.e., readings below 1.0 are bullish while readings above 1.0 are bearish.  Here's a daily $CPC chart for the past couple of years: 


Ca20090304-1 Click here for a live version of this chart

Some people find the daily $CPC too volatile and use the 20-day EMA to smooth it.  If the EMA (blue line) is above 1.0, things are bearish,  if it is below 1.0, things are bullish.

As you can kind of tell, $CPC gives some pretty good longer-term signals which causes people to want to examine it using shorter and shorter time frames.  However there is an interesting problem with ratio-based indicators like $CPC when you look at their intraday datasets.  Check out this intraday chart for $CPC:

Ca20090304-2 Notice how, at the start of each day, $CPC basically "freaks out" and jumps to an "extreme" value that is almost always the high or the low for the day.  $CPC usually settles down within the first hour and then moves around in a relatively narrow range until the close.  The next day, the pattern repeats again.

The "freak out" behavior at the open is due to the fact that the initial "Put" and "Call" volumes reported by the CBOE can be skewed and/or delayed by a variety of factors.  When that happens, the division operation can results - on bearish days - in some very LARGE spikes right at the open.  Bullish days aren't quite so bad since bullish readings are restricted to between 1.0 and 0.0 - although if you are charting $CPC on a log scale chart, those same large spikes will also appear.

On the chart above, the spike that goes up to 3.70 is a very mild example of what I'm talking about.  As you can see, big spikes do not always occur.

If they appear, the spikes happen right at the open and then the $CPC almost immediately gravitates back towards 1.0.  Unfortunately, the spikes will often cause the scale on the chart to expand dramatically which makes it very hard to analyse the $CPC values during the rest of the day.  The spikes can also have a huge impact on the values of any moving averages or indicators that appear.

Because these spikes have a huge impact on the chart and because they are "artificial" in the sense that they depend mostly on the speed at which the first "Put" and "Call" data is reported, we here at StockCharts.com feel that they are extremely misleading and should be removed from the chart.  The question then becomes "How?"

Initially, we "removed" them by making $CPC an End-of-Day only index.  We simply didn't allow people to create intraday charts for $CPC and we always set the daily open, high, and low equal to $CPC's closing value.

Ultimately though that solution didn't seem fair to our users that wanted to see $CPC values during the day so recently we've change $CPC back into a full-featured intraday index.  But, that has reignited the issue of what to do about the spikes.

One solution we considered was to not show any $CPC values until after 9:45am each day.  By then, the spike (if any) would have "settled down" and wouldn't be distorting the chart as much.  The problem with that approach is that it would deny people access to the direction of the $CPC in the first minutes of each day and - while the magnitude of the spike is usually misleading - the direction (bullish or bearish) can be valuable.  So just "hiding" the spike would deprive our users of potentially critical trading information.

Instead, we decided to simply cap the value of $CPC at 5.0 and at 0.10.

This approach allows people to see the value of $CPC throughout the entire day and it limits the "misleadingness" of any spikes.  The value of 5.0 was chosen somewhat conservatively since the largest closing value for $CPC in the past 14 years was 1.70 in early 2007.  Similarly, the smallest $CPC value during that same time has been 0.30 in 1998, so a minimum value of 0.10 seems pretty reasonable.

By capping $CPC, we think we've come up with a solution that allows useful, usable charts even when intraday time periods are used.

NOTE: If you have a saved $CPC chart in your account that relied on our old "End-of-Day" data format, it may look very different as a result of our changes.  You can "fix" that problem by changing your chart from a Candlestick or OHLC chart into a "Line (thin)" chart like I've shown below:

BEFORE:
Ca20090304-3 Click here for a live version of this chart

AFTER:
Ca20090304-4 Click here for a live version of this chart

I'd love to hear from you if you have an opinion about this change.  We're always open to better solutions if the can help everyone make better investing decisions.

- Chip

« Previous Article | Next Article »

Sounds like a reasonable solution. I've gotten very valuable use out of both a daily and a weekly chart of CPC, and I will agree that the spikes could really screw up associated indicators. Best, XJB
The original version with thin lines is much clearer, easier to interpret, and better IMO. How anyone can read the new version that includes intraday prices is beyond us...
My personal preference is the original version of the CPC. The solution of using a line in lieu of a candlestick for the weekly and monthly charts distorts the data as presented in the original version. The end of the day info from Monday thru Friday on the weekly chart as presented with a candlestick is far more useful info then just a line which only shows one value for the week. Same holds true for the monthly chart for CPC. Incidently, the NALOW chart does the same thing as the CPC chart during the first hour of trading when using intraday data. Thanks for looking at more ways to improve charting.
Great adjustment! It gives us more flexibility in using the data. My solution for this kind of problem was/is to use the Chart Type on Invisible and use one or more MA/EMA lines, the spikes are smoothed off and the scale of the chart adjusts to the averages not to the real spiked value. It will be very good if you can do the same for the $CPCE (so far I could not find $CPCI being as useful as the other two). As Dan Eckert mentioned in the above commentary, there is a similar problem with the $NA/NY LOW and HGH, and not only at the beginning of the day. During the last few months we had several days with no new highs on both exchanges. This becomes a real problem 1) when one uses Log charts (basically the log charts are not available) or 2) when one uses ratio graphs as $NALOW:$NAHGH (as a parenthesis I personally found this indicator being in top five on signaling intermediary bottoms in the market, the top indicator may work as well but requires a little interpretation function of the market being in a bull or a bear type). An idea will be to limit these indicators to 1 as a minimum, both intraday and daily. Or maybe to 0.5, then will know that a real 1 is not a set limit 1 and 0.5 is in fact 0.
Most references I see on the various p/c ratios use it as a contrary indicator. Puts are the numerator and when more puts are bought it is >1 and as a contrary indicator this is considered to be bullish.Isn't it? Is this best way to use the p/c? If not, how should it be used? I see more on it lately and want to know how to best interpret it. Thanks to anyone who cares to answer.
@Jim- The Put/Call is often used as a contrary indicator, but can be used as an confirming bullish/bearish indicator as well, depending on the time frame and "how you view the picture." For example, as a contrary indicator, many folks use a major spike Downward in the daily value of the P/C as a sign of a short-term TOP (everyone and his brother has gone long/piled into Calls, time for the smart money to sell), and a major spike Upward as a sign of a short-term BOTTOM. There are many variations on how to define "major spike": Some look at absolute values, say, a P/C above 1.3-1.4 as an up-spike and a value below 0.7-0.8 as a down-spike. Others (including myself) place moving average envelopes around $CPC and use spikes outside of these envelopes as contrary signals.... very similar to what we do with the VIX. However, as Chip shows above, the Put/Call can also be a useful CONFIRMING indicator of the market's trend, simply by putting a moving average on the daily values of $CPC and watching the slope and direction of that moving average, particularly in relation to the "zero" line.
I have always viewed the $CPC on a weekly basis and in "Line Solid" Type. Further, I use the Bollinger band and the momentum oscillators (such as CCI, Full Stochastics, etc.) rather than examining the $CPC by itself. Note that when the $CPC reaches/exceeds the upper Bollinger band and the oscilators register grossly overbought readings and/or show negative divergence, that is usually associated with a low-risk, high-reward jucture point. Conversly, when the $CPC reaches/exceeds the lower Bollinger band and the oscillators register grossly oversold readings and/or show positive divergence, the market stands at a high-risk, low-reward juncture point. Leon Tuey
comments powered by Disqus