“Moneyball is horseshit.”
Not Brian Burke’s exact quote, and he wasn’t exactly talking about statistical analysis as a whole when Burke mentioned at this year’s Sloan Sports Analytics Conference, but we can reasonably say that the Toronto Maple Leafs general manager of four years doesn’t think too highly of advanced analysis over traditional methods of player research.
“Everybody is looking for these ‘Moneyball’ breakthroughs. … I have yet to see anything that has value in terms of an alternative way of evaluating players.”
Looking back on these comments, and then reading Ellen’s piece on the Los Angeles Kings yesterday, and you realize that maybe the biggest opponents of any sort of sabermetric (‘sabremetric’ in Canada) applications in sports are probably those who misunderstood the concepts outlined in Moneyball.
The story goes, or so the movie tells is, is that the Oakland A’s, a baseball club with not a whole lot of money, loses three star players and using nothing but numbers, finds cheap replacements who are cast-offs from their previous teams and finds success.
Michael Lewis’ narrative ignored key aspects of Oakland’s success. While the team was reasonably successful getting on base, the book ignores that their two best hitters were likely Miguel Tejada and Eric Chavez. It also doesn’t even mention their “Big Three” pitchers in Barry Zito, Tim Hudson and Mark Mulder.
Any fan of the book is already aware of some of the criticism, but part of the reason for the reluctance of people to understand Moneyball is that Moneyball makes understanding the 2002 Oakland A’s problematic.
Moneyball, and Moneypuck, at its root, isn’t about mere statistics and discovering new ways to evaluate hockey players. The evaluation period is done. Fedor Tyutin had a fantastic season in Columbus controlling possession against top competition, but he already signed his big money 6-year, $27M contract. Even Burke gave a good payout to his top two-way centreman Mikhail Grabovski, a 5-year, $27M deal.
Evaluation isn’t a problem. For somebody like Brian Burke, he has watched enough hockey, hung around with enough players and coaches, and been in the game long enough to be able to judge Grabovski’s value without a spreadsheet or even a peep at the man’s Relative Corsi number.
Scroll down the list of forwards with the highest Corsi ON rates and look for the players who have a Corsi Rel QoC of above .8 or a zone start rate of under 50%. You get names like Patrice Bergeron, Anze Kopitar, Pavel Datsyuk, Johan Franzen, Joe Thornton, Dustin Brown, Henrik Zetterberg… none of these guys are really making a lot lot less than their underlying numbers indicate, or players that even the most casual fan wouldn’t recognize.
For savvy managers, this presents a challenge, players that can influence possession aren’t the most under-valued players in hockey. The Philadelphia Flyers knew the value of Mike Richards and Jeff Carter when they dumped them off their team in a few short hours, but either wanted to go in a different direction in regards to goaltending or saw their off-ice extracurriculars as a distraction or a detriment to the team.
Whatever, said Dean Lombardi. Whatever Richards had allegedly done with himself, it wasn’t a big enough deal to avoid trading a hyped-up Brayden Schenn and Wayne Simmonds for the play-driving centreman. It probably didn’t take a mathematical genius to drop Jack Johnson in favour of Jeff Carter, either, just a willing Scott Howson who was too impatient to see his expensive, long-term roster through after a half season of bad goaltending.
I’ve already covered in this space the plays that Boston and Chicago made to win their Stanley Cup rings. Pittsburgh is one of the teams that is the most open about their advanced methods, and while it isn’t necessarily Corsi-based analysis like a lot of what is done online, there’s still some advanced variant to what the Penguins do to go a step further and find players who are more than good puck possession players. Detroit, who won before Pittsburgh, started scouting and drafting in Sweden before anybody else did and they landed a tonne of talent.
The Kings haven’t won anything yet, but they’re as close after Game 1 as any team has been in recent memory. And should New Jersey win four of the next six games, it’s not as if their own Moneyball tactics didn’t work. They picked up Dainius Zubrus, a pretty competent forward, for $3.4M off the free agent wire and traded a few peanut shells at the deadline for Alexei Ponikarovsky. While they’re known for their big-money deal for Ilya Kovalchuk, they also have a cheap defence. They picked up Marek Zidlicky for free from the Minnesota Wild who no longer saw him having any use, and their contracts to Henrik Tallinder, Anton Volchenkov and Andy Greene, while not total bargains, are reasonable.
In short, Moneyball isn’t about poring over a spreadsheet and picking players based on numerical value and not name or character. It’s the act of picking players who display an undervalued trait, or getting players for cheap because they lack a piece others overvalue. Burke is wrong in saying that Moneyball has never won a championship. I’d be surprised if any team in sports history ever won with a team where a manager never thought to acquire at least a few players that other teams passed over because they were too inexperienced, too fat, too bald or too short.
For a team to win purely based on numbers, if that has happened, no team is likely to tell us, but the overall concept of Moneyball is definitely more than just a trend. If a team can exploit a competitive advantage, they ought to exploit a competitive advantage, or they will be cursed with un-success and misfortune.
And the Kings aren’t perfect. There is probably some aspect of the game that the Kings overvalue or undervalue. If a team had found a perfect model, they’d be the best team every season. We’re past the evaluation process. The Flyers knew what they had in Carter and Richards, and likely the Oilers did with Penner. There was just one aspect of those players’ on-ice or off-ice qualities that their teams didn’t like. Something that the Kings just didn’t seem to care too much about.