Though Alex Ovechkin took home his second straight Hart Trophy in 2008-09, it was regular season points leader Evgeni Malkin who ended up lifting both the Conn Smythe Trophy and Stanley Cup by postseason’s end. If it were ever in doubt to begin with, this playoff performance cemented Malkin’s status among the elite of the elite in the NHL. In the regular season, Malkin’s 24.1 Goals Versus Threshold (GVT) ranked fifth among skaters behind Ovechkin (27.1 GVT), Pavel Datsyuk (26.9 GVT), Zach Parise (26.3 GVT) and Mike Green (25.9 GVT), and tenth overall.
As you can imagine, the young superstar’s value has not changed appreciably this season. Though missing a handful of games due to injury, Malkin’s 6.6 GVT still ranks among the top 50 players in the NHL, and it’s bound to improve from there. The 23 year old center currently places second on the mighty Penguins to Sidney Crosby’s incredible 11.4 GVT.
So it goes without saying: in the real world, you might not put more than Sidney Crosby and Alex Ovechkin ahead of Evgeni Malkin on a team’s depth chart. But in the world of fantasy hockey, there are other considerations, such as a player’s number of games in the upcoming fantasy week (primarily for those not playing in daily fantasy leagues), strength of opposition (particularly defensively), number of home versus away games and number of back-to-back games. In the fantasy hockey world, it’s not uncommon to find scenarios where benching Malkin may be the correct “coach’s decision”. For instance, if your other fantasy centers happen to be Atlanta’s Rich Peverley, Florida’s Stephen Weiss and Ottawa’s Mike Fisher, you should let Mr. Malkin ride the pine this week. Does that seem odd? Let’s take a look at how this can be the case:
OGA - Opponents' Goals Against - Average GA per game of this week's opponents
mOGA - OGA modified by home ice advantage and back-to-back game effects
Games - Games in the upcoming fantasy week, Sunday - Saturday
G - Goals prediction for the upcoming week
A - Assists prediction for the upcoming week
P - Points prediction for the upcoming week
Expected production of select centers, Fantasy Week 11
Name Team Pos Week's opponents OGA mOGA Games G A P
Rich Peverley ATL C @NYR @FLA DAL* NJD 2.87 2.90 4 1.63 2.48 4.11
Sidney Crosby PIT C PHI @PHI BUF 2.66 2.70 3 1.75 1.74 3.49
Stephen Weiss FLA C @NYI ATL, CAR @CAR* 3.34 3.37 4 1.32 2.10 3.42
Mike Fisher OTT C @TOR BUF @NJD MIN* 2.71 2.73 4 1.92 1.49 3.41
Evgeni Malkin PIT C PHI @PHI BUF 2.66 2.70 3 0.91 2.37 3.28
Derek Roy BUF C @MTL @OTT TOR PIT* 3.02 3.05 4 0.99 2.14 3.12
John Tavares NYI C FLA NYR @NYR* MTL 3.06 3.15 4 1.87 1.23 3.10
Nik Antropov ATL C @NYR @FLA DAL* NJD 2.87 2.90 4 0.47 2.03 2.50
Jonathan Toews CHI C STL BOS DET 2.65 2.87 3 1.09 1.01 2.09
Joe Thornton SJS C ANA 3.22 3.44 1 0.28 1.41 1.69
*Second game of back-to-back games
Just by eyeballing the table above you can see the dual importance of a player’s ability and the number of games he will play that week. Above average forwards can often be bumped ahead of elite forwards on the strength of an extra game, as in the case of Peverley over Crosby and Fisher over Malkin in Fantasy Week 11. (In the same vein, you can see exactly how dominant Sidney Crosby’s Hart Trophy-caliber season really has been, as three games of Sidney Crosby project better than four games for most top six forwards). So for those of you in leagues where you set your lineup weekly, the first step in optimal lineup selection requires a simple number crunching exercise, multiplying a player’s per game production times the number of games that week. If you already make that calculation, you’ve taken care of the two most important considerations in setting your lineup, which should have you in good shape as it is.
How do you determine per game production? Many of you may automatically think that an average over the course of the season is most accurate. But how do you take fluctuating levels of individual and team performance into account, while retaining a large enough sample size? Back in the early days of the Shots On Goal column, we discovered that looking at the last 27 games (9 fantasy weeks) seemed best for future prediction of goals, while looking at the last 18 games (6 fantasy weeks) seemed best for future prediction of assists. Taking this research into account, the table above bases predicted goals per game on the past 27 games’ performance and predicted assists per game on the past 18 games’ performance.
The next most important consideration in setting your lineup is the upcoming week’s level of opposition; in the case of skaters’ production, this means the level of the oppositions’ defense. To account for varying levels of opposition, the predicted rates of goals, assists and points can be modified by the ratio of Opponents’ Goals Against (each opponent’s average goals against per game) to the NHL average Goals Against (GA), which currently sits at 2.85 GA per game. For example, Malkin’s third game of the week will be against the Sabres, who are just about the stingiest team in the NHL in GA at 2.17 (Chicago’s 2.16 GA has now passed both New Jersey’s 2.19 and Buffalo’s 2.17). Therefore, Malkin’s current goals per game level of 0.32 and assists per game level of 0.83 would be expected to decrease by a factor of 2.87/2.21, to 0.24 goals and 0.63 assists for the Buffalo game.
Finally, adjustments should be made for home ice advantage and back-to-back games, effects that Tom Awad recently showed were on the order of 0.26 Goal Difference each. The effect of these adjustments to OGA is shown as modified Opponents’ Goals Against (mOGA) in the table above.
With all that information under our belts, let’s return to this week’s decision on Evgeni Malkin. While three games for the Penguins’ center should produce more points (3.28) than four games of Buffalo’s Derek Roy (3.12), New York’s John Tavares (3.10) and Atlanta’s Nik Antropov (2.50), he is not likely to produce more points than four games of Atlanta’s Rich Peverley (4.11), Florida’s Stephen Weiss (3.42) or Ottawa’s Mike Fisher (3.41). So it’s really true: even with a typical three games this fantasy week, you’re better off starting above average centers like Peverley, Weiss and Fisher instead of Malkin, because of the extra game.
Well, that’s nearly true. For this week, you’ll also notice that while additional game evens the playing surface, it is strength of opposition that pushes the alternatives ahead of Malkin. While Malkin faces off against above average defenses of 2.70 mOGA (particularly on the road against Ryan Miller and Buffalo), Weiss gets to look forward to an easy week of 3.37 mOGA opposition (including the leagues’ worst defense of Carolina, twice), while Roy and Tavares also a relatively soft schedule of 3.05 mOGA and 3.15 mOGA opposition.
So no Malkin – What about Blackhawks captain Jonathan Toews and San Jose first line center Joe Thornton? As you might guess, three games for Toews (2.09) and one game for Thornton (1.69) don’t even put them on the radar screen as viable fantasy choices for this week, probably no matter what your roster consists of.
What about daily leagues? Daily leagues require a surprisingly different set of skills from a fantasy owner than weekly leagues, as I’ve learned as a rookie in the SportsJudge Fantasy Hockey Challenge. Except for Fridays and Saturdays, when most NHL teams are on the ice, there are really very few choices about which players to play. On weeknights, every player with a scheduled game is plugged into your lineup. Therefore, skillful roster construction is proportionately that more important in daily leagues, whereas weekly leagues require more number crunching and counterintuitive selections, like we’ve covered today. Select your players based on per game production, without multiplying by number of games in the upcoming week.
Timo Seppa runs the statistical hockey site Ice Hockey Metrics. Follow Timo on Twitter at @timoseppa.
Timo Seppa is an author of Hockey Prospectus.
You can contact Timo by clicking here or click here to see Timo's other articles.