I saw a question on David Varadi’s blog about how to compute relative strength. That’s probably the wrong place for such a question, but it got me thinking about the distribution of strength as a measurement of market correlation.
First, to answer the original post, relative strength is how well an issue has done over some period, relative to other issues. For example, AXP is up 10% in the last six months, which is 30% better than the general market:
I measure relative strength using a modified Levy ratio, SMA(10) divided by SMA(131). You can see how this is a momentum indicator. I use SMA(10) instead of the latest close, to filter out recent fluctuations.
Without doing the percentile math, you can simply drop this formula into your watch list and sort on it. If you use the S&P 500, for example, the 100 that sort to the top are the quintile with relative strength 80 and above.
Here is how strength is distributed among the S&P 500 today:
This is a broad distribution, with a standard deviation of 0.09. The middle quintile spans a range from 2.1% to 5.8% above their respective moving averages. This is what is meant by a diverse “market of stocks.”
A year ago, AAPL was around $660. It had just peaked, and its Levy ratio was near 1.0. The S&P 500 were mostly trading below their six month averages, and 428 of them were within ± 2% of it. The standard deviation of relative strength was 0.01. If you were trading at the time, you will recall that the market was highly correlated. Twitter remarked on it daily.
The shape of this distribution, over time, shows the population of stocks alternately fanning out or bunching together near their trendlines. The standard deviation is a succinct measurement of this correlation effect.