Predicting the S&P and Voodoo Quantation -- a Post-Mortem

Jan 5, 2015

Happy New Year! In one of my last 2014 posts, I lambasted a particularly silly Yahoo! Finance blog titled “This Signal Has a Perfect Record of Forecasting Year-End Gains,” in which the author used a convoluted calculation based on the Market Volatility Index (VIX) to predict the S&P 500’s performance for the remainder of the year. We rarely have the opportunity to skewer such pontificating so promptly, but this “signal” no longer has a "perfect record."

The S&P close on 12/9, the day before the blog was published, was 2,059.82. Unfortunately for the aforementioned “signal,” after an exciting December (the S&P’s peak close was 2,090.57 on 12/29), the S&P ended the year at 2,058.90, down 0.92. That’s a miss of less than one-twentieth of one percent (0.0447%, to be exact), but a miss nevertheless. [As Maxwell Smart, CONTROL’s Agent 86, would have observed while holding his thumb and forefinger a tiny bit apart, “Missed it by that much.”]

Now, from a theoretical standpoint, the fact that the S&P closed down from 12/9 to the yearend is beside the point. The “signal” was nothing more than a random discovery to begin with, with no logical or causal hypothesis backing it up. So there was a 50/50 chance that the “signal” would be “correct” just because of random chance. (Actually, a little better than 50/50, since up days in the stock market outnumber down days by about 53% to 47%) But at least we won’t be hearing more about this one. I hope.

But the debunker’s work is never done. I call your attention to another Yahoo! Finance post, in which the writer bloviates about the significance of how the market performs on the second trading day of the year. Thank God for the Gregorian calendar!

“Painting with Numbers” is my effort to get people to focus on making numbers understandable. I welcome your feedback and your favorite examples. Follow me on twitter at @RandallBolten.