Trying some adjustments to the Chicago Quantum Net Score tonight

Jeffrey Cohen, President, US Advanced Computing Infrastructure, Inc.

We run our model frequently, but tonight we thought about it in a different way.
The best stock picked is a terrible stock with a back story. However, they paid out an outrageously large dividend and the stock fell 66%, or 3x the dividend, so this stock looks amazing to our model. Huge dividend payout. High expected return. Low variance. And now a low valuation. What could go wrong for the investor?

We don't know what could go right (it recovers) or wrong (it goes bankrupt), but we know that this is not the intention of the model. So, tonight, we are marking down the 'dividend adjustment' in our model from 100% to 50%. This works to 'add back' the price drop of the full dividend that we assume the market made on the ex-dividend date. We are lowering that adjustment today to add back 50% of the dividend paid to the stock price.

The second thing we did was to manually remove 11 stocks with a 20% or more dividend payout that actually happened in the past year, when divided by the stock price. For these 11 special stocks, one of which paid the outrageous 1x dividend, and another paid out a 1x dividend in response to a short attack. If they don't recur, we can add them back in a few months.

We also found a startling discovery. When we look at our stocks and use the raw variance and co-variance, we get a set of stocks to hold in our optimized portfolio. These have the best risk-adjusted return. However, when we look at a 'side calculation' we did to be able to compare all stocks, we divided actual variance by the square of the average stock price, and this gives a nice, simple percentage like 1% or 10% or in some cases, 100%. Low risk stocks score under 10%. This simplified model works when doing an excel spreadsheet analysis.

However, we are uncertain what would happen if we adjusted all risk factors before they are put into matrix form (variance, covariance, etc.). However, we do want to 'dampen down' the massively high expected returns for individual stocks in this smoking hot market, and we want to make variance more important in the selection of stocks. The market is so hot, and many stocks are now sporting BETA values greater than 3.0, that a hot stock could show an expected return of 40% or more.

We made a few adjustments to account for this. We set an adjustment factor to cut the CQNS_Power, or the relative weighting of expected returns, from no adjustment (or 100%) to 25%. This should significantly reduce the power of expected returns (which are running hot), and increase the relative power (unadjusted) of the stock price variance. A cut from one to one-fourth is a massive cut (over  90%), so we will see if it needs to be tuned to say one-third, one-half, three quarters, or even 95%).

We also cut the stock market return ceiling from 16% to 15%. We use the CAPM process to calculate expected returns. This assumes that the only thing that matters in predicting future stock price returns is to calculate the expected return to risk (or the return to holding risk assets over risk-free assets). We have now set a 1% lower cap to the return used in that calculation, from 16% to 15%. This should be a very small change, and within the realm of empirical evidence.

What do we expect to see:
1. We expect the dividend anomalies to be mitigated and eliminated. Paying out a 1x dividend will not 'head fake' our quant model.
2. We expect there still to be high expected returns (the market is hot), but they will have less weight on the model. The model should tip towards lower risk stock portfolios in this next run.
3. We will look at the small 'excel test' that compares the CQNS score and the simplified expected return - normalized variance calculation to see if the differences are still there, and if they are less intense.

Hope you enjoyed reading our 'inside baseball' on how we model our math and analyze stocks.

P.S. We recently decide to add back the smallest of the small cap stocks to our model. There is market action afoot that is pumping up small-capitalization stocks, so we want to bring visibility to those names. For those who only invest in larger capitalization stocks (say $250M and above), we can still tailor our run for your specifications.