As promised, after having had time to digest Alan Ryder's epic 50-page
2008 NHL Review, I'm here to chime in with thoughts on the overall analysis, as well as a look at how the Nashville Predators fared in particular.
First, a bit of an introduction to Ryder's work. Alan's research articles, found in full over at
HockeyAnalytics.com, aim not just to apply statistical analysis to the study of NHL hockey, but also to properly balance the various factors within the game, giving appropriate emphasis to goaltending vs. power-play production, for example. His basic approach within this review is to start from the foundation of team accomplishment (measured as points in the standings), then to divide that up between members of that team according to their individual
Player Contribution (PC). This PC value is calibrated such that 10 PC points are equivalent to one earned point in the NHL standings. Thus the Detroit Red Wings, who won the Presidents Trophy with 115 points, have 1150 PC to distribute among individual players based on their role in team success (the numbers don't exactly marry up due to rounding and players moving from team to team).
Each individual is then allotted points based on their production at even strength, on the power play and penalty kill, with allowances for ice time in each situation (if you get more ice time, your production is expected to be higher). There is also consideration for
penalties taken and drawn (new this year thanks to the NHL providing that data). Goalies are judged on their save percentage as adjusted for the Shot Quality that their team gives up. As we'll see below, Florida's Tomas Vokoun gets an allowance for doing yeoman's work behind an awful defense corps. There are also factors included for performance during the shootout.
Like any analysis out there, the limitations of this model need to be well understood. Certain aspects of the game, like performance on faceoffs, are ignored. Skaters are measured in terms of goals for and against in various situations, which would seem to be somewhat inconsistent with the goalies who are measured across all shots. Also, the quality of opposition isn't accounted for. If two players on the same team produce the same results across the same amount of ice time, but one player did it mostly matched up against elite opponents, while the second played against scrubs, their PC values would be the same, but you'd certainly argue that the first was more valuable. That said, Ryder's is probably the most sophisticated broad-based analysis out there, and provides some surprising insight into the relative performance of various NHL stars.
HighlightsThe first observation, which should be obvious to most NHL fans, is that goaltender is by far the most important position on a given team. This is reflected in the fact that 18 of the top 20 spots in the PC rankings go to goaltenders, led by Tomas Vokoun of the Florida Panthers. The top skaters are Alexander Ovechkin (12th) and Jarome Iginla (18th). The difference between the top goalies like Vokoun and Martin Brodeur (283 and 276 PC, respectively) and an average performer like Kari Lehtonen or Martin Gerber (173 & 156) can be 10-15 points in the standings, enough to make or break a team's postseason hopes singlehandedly.
Speaking of goalies, the normally soft-spoken Ryder makes no bones about the Vezina Trophy candidacy of San Jose's Evgeni Nabokov, who
led the NHL in wins this year and drew raves reviews from some circles:
"It slays me that Nabakov was a Vezina Trophy
finalist. What got him there was a league leading 46
wins. Of course he got a lot of wins. He played a lot
(77 games) for an outstanding team (second best in
the NHL). Goaltender wins is the most useless
statistic in the NHL."
I couldn't agree more; Nabokov had a fine season at first glance, but he played just about every night behind one of the NHL's top teams. In Ryder's Defensive Index rankings (which bring together Shots Against with the Quality of the shots being given up) San Jose was 1st in the league, even ahead of Detroit. In other words, San Jose's goaltenders had the easiest job in the NHL last season, while Florida's had the toughest.
Moving over to the skaters, Alex Ovechkin (162 PC) comes out the clear winner ahead of Jarome Iginla (141) and Pavel Datsyuk (132), no surprise there. Among defensemen, Nick Lidstrom (121) leads the way followed by Brent Burns (108) and Brian Rafalski (105). Burns may jump out as a surprise there, but his three Shootout goals earned him 16 PC points in that category alone, another indication of how impactful the shootout has become to NHL standings. Other standout
Shootout Specialists were Chicago's Patrick Kane (25 PC from shootouts, 80 PC overall) and San Jose's Joe Pavelski (22 PC from shootouts, 82 overall). With the NHL considering changes to bring about
even more shootouts, could we see a day when a roster spot is reserved for a shootout specialist who plays only a few minutes of regular ice time? The argument in favor of that appears to get stronger and stronger.
The Nashville AngleSo how do the Nashville Predators shake out in all of this? In summary, the Preds came out as a very average NHL team last year, which, given the expectations that the hockey world held as David Poile was forced to slash payroll, is a
remarkable accomplishment. With the 17th overall points in the standings, Nashville came in 12th in Marginal Goals - Offense (an overall measure of offensive performance), and 15th in the Defensive Index (Shots Against adjusted for Shot Quality). Marginal Goals - Goaltending came in 13th, a mixture of Dan Ellis' surprising performance and Chris Mason's failure to hold the top job.
Individually, Dan Ellis came in 16th among goaltenders, which is pretty good considering he played in only 41 games. On the blue line, Dan Hamhuis (53 PC) and Ryan Suter (52) led the way overall. In terms of pure defensive performance, Greg Zanon (45 PC for defense) was tops on the team, but his offensive production was dreadful (-7 PC, meaning that he crossed the marginal performance barrier, a symbolic level that represents what a replacement player from waivers or the AHL might offer). The recently departed Marek Zidlicky was by far the top offensive threat on the power play (14 PC in that aspect, nearly twice Shea Weber's 8). In penalty killing, Hamhuis, Zanon, and Greg de Vries were the standouts. Ville Koistinen, who played in 48 games last year, came in with 32 PC overall; if he can build on that performance in 2008-9 and play something closer to 70-80 games, he may develop into a pivotal cog within that group.
Up front, the centers were led by Jason Arnott (55 PC), who excelled at Even Strength (46 PC). Not much else is notable within this group except for the fact that of Radek Bonk's 22 PC, 11 were earned during the shootout, furthering my point that his line was a weak point in the Preds attack last season, and a revamped 3rd line is essential to improving the team. I highly doubt his intended role on the Predators is to serve as a shootout specialist like I posited above.
On the wings, J.P. Dumont's career-best 29 goals made him the top Nashville forward overall with 73 PC, 59 of which were earned at Even Strength. Martin Erat came in second with 61, based not only on good ES play but also 10 PC points earned through his
Penalty Plus/Minus performance. Alexander Radulov's 3rd-place performance is interesting, in that he tallied 56 PC at Even Strength, but lackluster results on special teams docked him 10 points in total, leaving him with 46 overall.
Sound familiar? The Predators basically came in with a very strong performance at Even Strength, but dismal performance on the power play held them back as a group. Heading into the 2008-9 NHL season, putting together a stronger power play is probably the single largest opportunity for team improvement, and as I mentioned recently, with
Marek Zidlicky having been traded to Minnesota, I wonder just how they'll try to do it.
Labels: nashville predators, statistics
