Tuesday, February 12, 2008

NHL's best bargains for bucks

One unfortunate thing that happens when advanced statistics begin to permeate sports is that you end up with cheap imitations. An example: Bill James started with a well-thought-out framework, came up with Runs Created, and optimized it until it Runs Created correlated as well as possible with actual Runs.

But all kinds of people followed-up Runs Created with meaningless offensive statistics - I remember my 1989 baseball season preview described five or six of them and treated them as though they were equally significant. Some of the "competitors" to Runs Created used RBIs in their formulas or failed to penalize players for getting caught stealing. Simple stuff. The only statistical testing they underwent was for the person who made them up to look at the top ten and decide that it did a good enough job of picking out the best players in the league.

You can't do that in baseball anymore - the last decade has made it obvious that James knew what he was talking about, while the pretenders were doing just that - pretending to create statistics. This is not true, in hockey, unfortunately.

Witness the Toronto Star pretending to do statistical analysis:

NHL's best bargains for bucks

Long story short: they think the NHL's "best bargains" are 1) forwards who have the most points per dollar of salary; 2) goaltenders who have the most wins per dollar of salary. The list has the usual suspects: Corey Perry, Ryan Getzlaf, Alexander Ovechkin, Evgeni Malkin, Sidney Crosby, Patrick Sharp, Paul Stastny and Nicklas Backstrom.

It's hard to say how preposterous this is. It's obvious that the list will be filled with players with under three years experience - though not necessarily young players: Sharp is 26 - and guys who play the power-play. It's also hard to say that these guys are still bargains given what they'll get paid next year:

Sharp: $4M for 4 years
Crosby: $9M for 5 years
Ovechkin: $9M for 13 years
Getzlaf: $5M for 5 years

The others haven't signed extensions yet, but it seems insane to say that four players who are owed $210 million by their respective teams are bargains.

At any rate, this article is arithmetic masquerading as statistical analysis. It's obvious that developing young, talented players will give you the most bang for your buck. That's always been true, though it's less so now that the CBA forces teams to renegotiate after a player's first three seasons. But to effectively claim that power-play points per dollar of salary is a measure of a player's worth is patently ridiculous, not to mention ignorant of the entire rest of the game.

Labels: ,

Couldn't agree with you more. It's no news to anyone that young players who produce big numbers can still be had for smaller amounts.

When I think of value, I think of teams that locked up star players who had some leverage, and did so with long-term contracts for those players peak seasons, a Rick Nash, an Ales Hemsky. I also think of Chris Pronger, Martin Brodeur.

But I haven't looked at this in detail, either, so I can't be too hard on the Toronto Star.
Yeah, I laughed at that article as well. About the nicest thing you can say about it is that hopefully the author will turn toward more accurate metrics in the future.
Post a Comment

Links to this post:

Create a Link

<< Home

This page is powered by Blogger. Isn't yours?

Subscribe to Posts [Atom]