Slumps and streaks

Washington Capitals v Toronto Maple Leafs

A common theme of the early NHL season as well as one of the most boring stories has been the number of people asking “what’s wrong with Phil Kessel” or “what’s wrong with Alex Ovechkin” or “what’s wrong with [so-and-so player who has had a cold start to the season]?”

I feel like when goal-scoring slumps happen later in the season, commentators have more context to work with. “Zero goals in ten games” looks worse on paper when it comes ten games into the season than “zero goals in ten games” looks when a player has scored five goals in the first ten.

Without the first few games of the season, no discussion about a players goal scoring habits can be appropriately framed as a cold streak. It becomes a season-long catastrophe, one where summer work habits can be questioned and professional coaches can second-guess any new line mate for the scorer. Kessel has scored in consecutive games now, but he went his first ten without scoring. Ovechkin, who has scored a goal every two-and-a-half games since the start of the ’09-’10 season, went scoreless in four. He’s picked it up as well of late, also scoring in consecutive games.

And it’s interesting, because I run a Maple Leafs blog, to look through Kessel’s stats and see in which situations he catches fire. I was pretty surprised when I saw that Kessel actually scored more goals in games when he was on a scoring streak. I was actually quite surprised to see that Kessel had a higher scoring rate in games where he had scored 1 or 2 goals than after games where he had scored none.

So I had an idea about this, that maybe either all goal scorers were a little bit better after scoring goals, or, alternatively, their coaches genuinely believed in riding the hot hand and would give their player who had been coming off a game with a goal some more minutes. I looked at the last three seasons of the top ten scorers in the NHL since 2009: Ovechkin, Kessel, Steven Stamkos, Patrick Marleau, Corey Perry, Ilya Kovalchuk, Marian Gaborik, Jarome Iginla, Bobby Ryan and Daniel Sedin.

I found something I genuinely wasn’t expecting—coaches don’t necessarily play those scorers more in games after they scored a goal. Those players also don’t exhibit any signs of streakiness. I used the game logs and filtered out goals, shots and time on ice, checked to see how the player fared in their next game depending on whether they scored a goal or not in the previous game. Here’s the raw data in total games after a player scored a goal:

Player GP Goals Shots TOI
Stamkos 119 83 414 2522:31
Ovechkin 91 44 407 1913:54
Marleau 84 30 256 1766:10
Perry 88 37 312 1893:45
Kovalchuk 92 33 288 2099:23
Gaborik 83 40 258 1663:57
Iginla 83 35 253 1745:58
Ryan 77 29 226 1477:36
Sedin 82 41 275 1531:15
Kessel 83 39 338 1653:41

And after not scoring:

Player GP Goals Shots TOI
Stamkos 126 73 454 2604:07
Ovechkin 137 74 626 2867:37
Marleau 161 79 546 3339:05
Perry 155 77 520 3356:04
Kovalchuk 141 74 555 3257:40
Gaborik 136 64 474 2647:53
Iginla 162 72 539 3331:05
Ryan 167 71 502 3162:28
Sedin 134 59 442 2531:57
Kessel 150 60 569 2946:10

I looked at shots because I think shots on goal is an appropriate assessment of player offensive performance. Generally, an elite offensive player will get more than 3 shots a game, and a very good one will get more than 2. Averaged out over 10 games, a shots on goal counter will tell you something. Over 82, it tells you more. Over 2351, it’s better still.

The raw data doesn’t tell us much. I’ll break up both charts combining each player’s raw data, looking at Goals per 82 games, Shots per Game, and shooting percentage:

Goals/82 Shots/GP Sh% TOI/GP
After Scoring 38.2 3.43 13.6% 20:42
After Not Scoring 39.2 3.56 13.4% 20:27

I have a hard time looking at those charts and deducing that there’s anything different between a player coming in hot and a player coming in cold. Obviously, this doesn’t factor in players on three, four or five game slumps, but usually those are halted before they get too far, likely because it’s rare that a good goal-scorer will go into a prolonged slump of 8 or 9 games. When that happens, while there may be good explanations for why a scorer can’t find the back of the net that extends beyond variance, history tells us that you can’t predict a player’s performance based on his last game.

These ten scorers, by the way, see 20:42 of ice time in games after they score and 20:27 in games after the don’t. That’s pretty close as well, so it’s not like Corey Perry is going to see some extra shifts if he comes in on a two-game scoring streak. Over time, it balances out, and if you look at many many games over many many years, you’ll be closer to finding the players’ true talent.