Early on in this blog I wrote a rant on how the traditional method of looking at special teams effectiveness (PP% +PK%) was nonsense. Nevermind. Here is how the rankings would stack up if you used the traditional method vs. my proposed method (PPGF + SHGF - PPGA - SHGA).
Detroit: (1) 120.7, (1) 17
Toronto: (3) 114.1, (2) 11
Minnesota: (2) 120.3, (3) 10
Colorado: (4) 112.1 (4) 6
Ottawa: (6) 108.1 (5) 6
Trust me, it is the same for the ones at the bottom. However, notice one thing. An aggregate score of 100 gets you no where - it is now distinctly middle of the pack. To have truly good special teams your PP% + PK% needs to be above 105. We will see if this holds at the total number of power plays goes down as the season moves on.
Putting on my "what if" hat here, the methodologies seem to treat scoring on the power play, and being scored on while shorthanded, as a wash. Should they? If one team scores on the powerplay and is scored on while short-handed, are they equally effective as a team that does neither "just because" the scoreboard shows the same net result?
ReplyDeleteI've seen teams play "keep away" on the power play and I wouldn't call them effective, even if their penalty kill is reasonably effective.
Do the rankings change if we emphasize one variable over another? Perhaps something like 1.5*PPGF + 3*SHGF - PPGA - 3*SHGA. To emphasize power-play scoring, really reward short-handed goals, and punish short-handed goals against.
The "percentage" formula, of course, normalizes the equation for "opportunities": if you get more power plays (more penalties against your opponent), you'll get a higher PPGF for the same PP%. And in measuring effectiveness, it seems satisfying to normalize for your opponent although if it's your team drawing the penalties it "feels" less clear cut.
Do the rankings change if we normalize the PPGF? (1.5*PPGF/PP + ...)
Inquiring minds, and all that ...