Given that the Leafs are currently carrying an entire line’s worth of fighters, it’s understandable that a lot of discussion on this site has focused lately on the value, or lack thereof, of the pugilist caste. The word "meaningless" gets thrown around a lot.
Hockey certainly has had a statistical revolution in recent years, and in my case it has really enhanced the fan experience; but if significant research exists on whether fighting matters for wins and losses, it hasn’t been widely disseminated.
Unfortunately I’m no statistical revolutionary, so what I’m presenting is only the bluntest of analytic instruments, the correlation coefficient. I took seven years of data, 2005-2012, which is what hockeyfights.com has available.
The inelegant spreadsheet can be found here.
Correlation between Fighting Majors and Standings Points By Team, 2005–2012
P (2 tails) 0.031
Every one of you knows this doesn’t imply any sort of causal link, just that fighting and winning don't hang out much. Maybe better teams avoid fights (Detroit was last by a mile every year); maybe they don't have to fight; maybe fighting doesn't help. In any case it's marginal. It’s also only significant at the kind of level that psychologists and others whose n is measured in grant money and days consider significant.
But I sure wouldn’t argue that there’s some mysterious positive correlation hiding in that -.14. Goals for (-.21), goals against (-.05) and differential (-.011) pretty much tell the same story: fighting isn’t associated with winning.
"Meaningless" is about right. And I'm not comfortable ceding that many roster spots meaninglessly.