clock menu more-arrow no yes

Filed under:

WOWY is Worthless: Part 5

WOWY is worthless. It's things we know from Corsi plus a big dose of statistical noise. In Part 5, I show that WOWY doesn't correlate from year to year.

Lehtera and Schwartz
Lehtera and Schwartz
Jasen Vinlove-USA TODAY Sports

Part 5?

I hadn't planned on looking at WOWY this way but I found that WOWY doesn't correlate meaningfully from year to year. I looked at all player pairs with at least 1 event with and 1 event without in each year. If we look at 2011-12 and 2012-13 we get:

Pearson's product-moment correlation

data: W1112$WDiff2011 and W1112$WDiff2012

t = 3.9676, df = 8091, p-value = 7.324e-05

alternative hypothesis: true correlation is not equal to 0

95 percent confidence interval:

0.02229968 0.06578988

sample estimates:

cor

0.04406566

There is over 8000 data points, so the result is statistically significant. It simply isn't of any practical significance. Graphically, it looks like:

wowy1112

Looking at 2012-13 and 2013-14 we get:

Pearson's product-moment correlation

data: W1213$WDiff2012 and W1213$WDiff2013

t = 5.7512, df = 8046, p-value = 9.185e-09

alternative hypothesis: true correlation is not equal to 0

95 percent confidence interval:

0.04219584 0.08571346

sample estimates:

cor

0.06398506

Graphed, we get:

wowy1213

Once again, there is just nothing meaningful there.