Volume Flow Indicators: the Chaikin Oscillator
Motivated by e-mail from Brett S.

Joe Granville published a book in 1963, Granville’s New Key to Stock Market Profits, in which he popularized (and named) a method for measuirng "volume flow".
He called it On Balance Volume. We'll call it OBV.
One starts with some OBV value (it might as well be 0) and do the following:

• If today's closing price is larger than yesterday's, we add today's volume to the OBV.
• If today's closing price is smaller than yesterday's, we subtract today's volume to the OBV.
• If today's closing price equals yesterday's, we keep yesterday's OBV.
>Why?
The idea (enunciated by Granville) is that large volume changes anticipate large price changes.
Hence we calculate the cumulative volume changes (as indicated above) and use it to anticipate price changes.
Stare intently at the left chart below to see if you can identify that predictive characteristic.

>What's the upper chart ... on the right?
Patience.

For example, if there's a large increase in volume (indicated by a rising OBV), then (presumably) there's a buying pressure that'll cause the price to increase.
The OBV is like a spring, compressed and ready to generate price changes ... so they say.

>Do you believe that?
Is that important? I'm just telling you what the OBV is supposed to indicate.

After other reincarnations of the OBV, Marc Chaikin developed the Chaikin Oscillator.
>That's the chart on the right, eh?
Pay attention!
Chaikin goes like this:

• Each day we calculate the Close Location Value: CLV = ( ( (Close - Low) - (High - Close) ) / (High - Low) )
• We multiply CLV by todays' volume.
• We keep track of the cumulative total of CLV*Volume, always adding today's value to yesterday's.
That running total is called the Accumulation/Distribution Line.
• Then, just as one does in calculating the MACD, we take the difference between the 3-day Exponential Moving Average (EMA) and the 10-day EMA.
That's the Chaikin Oscillator.
 >And that's the chart on the right, eh? Yes ... the chart on the right. Note that (High - Low) is the range of prices and it's split into two fractions: x = (High - Close)/(High - Low)   and   y = (Close - Low)/(High - Low) and x + y = 1 and ... >And CLV = x - y, right? Actually, CLV = y - x.
So CLV measures whether the Close is closer to the daily High or the daily Low. When the Close is closer to the High, CLV is postitive. When it's closer to the Low, CLV is negative.
When CLV is positive, it indicates a positive outlook on the stock and if that's associated with large volume, that suggests an upward pressure on the stock price.
On the other hand, it CLV is negative ...

>Then the stock price will drop.
I never said that!
A negative CLV indicates that the Close is closer to the Low and if there's a large volume of trading, that suggests a downward pressure on the stock price.
Because positive and negative CLV values are indicative of price movements when associated with large daily volume, that's the reason for multiplying CLV by the volume and ...

>First you say OBV is like a coiled spring. Now you talk about upward and downward pressures. Are we talking physics here?
Hmmm ... interesting analogy. It's like Newton's F = ma.
There's a force F and a mass m which provides resistance to motion and the acceleration a indicates movement and ...

>Can we get back to Chaikin? I assume that CLV is the Chaikin Oscillator.
I never said that! You haven't been listening. We gotta do the MACD thing!

Remember:
The sum of past CLVs is the Accumulation/Distribution Line which we'll call the ADL.
It measures (much like OBV) positive and negative pressure on the price.
In the words of Marc Chaikin:
The closer a stock or average closes to its high, the more accumulation there was.
Conversely, if a stock closes below its midpoint for the day, there was distribution on that day.
The closer a stock closes to its low, the more distribution there was.

Now the price could be moving South and, if a positive ADL pressure stays around for a while, the price motion will reverse and head North and ...

>You're kidding, right? Are we talking physical pressure in a Northerly direction ... on an object moving South?
Uh ... why not? So we look to see if that pressure is increasing, recently, compared to what it was in the past. That'd indicate a change in the object's momentum.
That suggests that we compare some recent average ADL to a longer historical average.
That suggests moving averages and, if we want to weight recent ADL-values more heavily, that means ...

>That means we'd use an Exponential Moving Average, eh?
You got it ... so we take the difference between the 3-day MACD (that's the recent average) and the 10-day MACD.

>Aah, finally we got us the Chaikin Oscillator. But why "accumulation" and "distribution"? Why call them that?
If the price is high, compared to the mid-point of the daily range (from Low to High), one considers this as money flowing in. If the price is low, it's flowing out.

>So the ADL is like a money flow indicator, right?
Yes. Though it attempts to measure volume flow, in and out, if we were to multiply by the stock price, that'd be "money flow" ... in and out.
The Accumulation/Distribution Line is also one of Chaikin's babies.

>Whether the flow is in or out depends upon the price being high or low compared to the midpoint of the daily range, right?
Yes, according to Chaikin ... but, of course, you can have any criterion you like.
Maybe you'd like the price to be high or low compared to, say, the 10-day moving average or maybe you'd like to compare the closing price to (Open + High + Low + Close)/4 or maybe ...

>Maybe I don't like that comparison ... to the midpint. Maybe I'd like to compare the Close to, say ...
Yes. If you'd like to generate some prescription you can do that, too. See the formula in cell O4? It says: ((CL - LO) - (HI - CL)) / (HI - LO)
You can change it to, say: CL - (HI + LO + OP + CL)/4 if you like. Then volume would be added if the CLose were greater than (HIgh + LOw + OPen + CLose)/4.
Type in your favourite formula and click a button and see what happens. It's fun!!

>But won't that change the ... uh ... the ...?
It'll change the CLV, hence the cumulative sum (which is the ADL) hence the Exponential Averages of the ADL, namely the Chaikin Oscillator.