A popular market breadth indicator, the McClellan oscillator is one of the tools that MoneyShow's To...
Measuring Overbought/Oversold Conditions
08/02/2013 6:00 am EST
For dedicated students of the markets curious about how to calculate overbought/oversold conditions, veteran trader Larry McMillan of McMillan Analysis Corporation offers his take on the proper way to measure these conditions.
Some analysts occasionally cite the percent that SPX is above its 20-day or 200-day simple moving averages [when measuring whether a market is overbought or oversold].The problem with using a straight percentage...is that it doesn't take volatility into account. In a very dull market, if SPX is 15% above or below a certain moving average, that may actually be more overbought than if it is 20% above or below that average during more volatile times.
The correct way to handle this, in my opinion, is to calculate the number of standard deviations (σ) that SPX is above or below the moving average. That incorporates volatility, and so can accurately compare one market environment with another.
Over time, we have received a number of requests as to how to calculate this figure, so here is a brief summary of the formulae. I realize this is mostly of interest only to self-avowed geeks and "do-it-yourselfers," but here's how I do it.
1) Calculate the 200-day historical implied volatility as the standard deviation of daily price changes. Annualize that by multiplying by the square root of 255 (the number of trading days in a year).
2) Solve the following equation for a, the number of standard deviations:
S = Peavt
Where P = 200-day moving average
S = current $SPX price
v = volatility from step 1
t = sqrt(1/365)
or a = ln(S/P) / (v * t)
By Lawrence McMillan, Founder and President, McMillan Analysis Corporation
Related Articles on TRADING
While fundamentalists delve into economic and financial data to analyze the market, technicians emp...
Being able to determine market direction is a trader’s most important skill, writes Markus He...
Markus Heitkoetter discusses reward/risk ratios and winning percentage, and why determining the dir...