San Jose's three goals in 2:50 came on shots worth about a third of one
San Jose beat Vancouver 3-2 on a burst worth 0.36 expected goals. Smith, Celebrini and Graf had the best of a five-on-five night that matched San Jose's season share, 42.0 percent.
San Jose scored three goals between 8:00 and 10:50 of the third period on shots worth 0.36 expected goals combined in MoneyPuck's game data. Rounded individually, they were worth 0.06, 0.27 and 0.03. That bought a 3-2 win over Vancouver; at 6:53 of the third, down 2-0, the Sharks had a 5.7 percent chance of winning in MoneyPuck's live model.
Will Smith's five-on-four wrist shot from 39 feet carried a 0.061 goal probability and moved San Jose's win probability to 22.4 percent. His five-on-five tying wrist shot from 12 feet carried 0.269 and moved it to 51.0 percent. Mason Marchment's five-on-five wrist shot from 31 feet, 11 seconds later, carried 0.028 and moved it to 83.8 percent. Vancouver's scoring shots were worth 0.54 and 0.05, or 0.59 expected goals across its two goals.
The third-period line#
Smith finished the night on Macklin Celebrini's line with Collin Graf. In the first two periods, Smith played mostly with Marchment and Michael Misa; that line finished with 0.09 expected goals for and 1.08 against in 5.5 five-on-five minutes.
In the third, Smith spent 6.1 of his 6.2 all-situations minutes with Celebrini. Smith, Celebrini and Graf had 0.62 expected goals for and 0.18 against in 4.9 five-on-five minutes, including Smith's tying goal. Every unblocked attempt with that line on the ice came between 3:50 and 10:39 of the third, while San Jose trailed. With Ivar Stenberg beside Celebrini and Graf, the totals were 0.07 for and 0.14 against in 7.4 five-on-five minutes.
The rest of the night#
Vancouver led the five-on-five expected goals 1.28 to 0.46 in the first period and 2.02 to 1.46 for the game. Across all situations, excluding empty-net attempts, its advantage was 3.00 to 2.29. One game is one game, even when its percentage happens to look familiar. San Jose's 42.0 percent five-on-five share matched its season share through 14 games, with 25.12 expected goals for and 34.65 against, and actual goals of 18 for and 30 against over those 14 games.
Special teams#
San Jose's power play scored on 1 of 5 opportunities, generating 0.67 expected goals on 10 five-on-four unblocked attempts; Vancouver scored on 0 of 3 power plays, generating 0.32 on 5 five-on-four unblocked attempts. Through 14 games, San Jose's power play is at 24.5 percent, scoring on 12 of 49 opportunities, and its penalty kill is at 83.0 percent, killing 39 of 47 opposing power plays.
Against me#
The arguments against this, and the first one is real.
- Score effects favor a trailing team's possession. All the unblocked attempts in the line's 4.9 five-on-five minutes came while San Jose trailed. Adjusting for score and venue drops the Sharks' five-on-five share for the game from 42.0 to 38.7 percent; while trailing, their five-on-five expected-goals edge was just 0.89 to 0.82.
- The sample is 4.9 minutes, however inviting the decimals look. Smith's tying shot supplied 0.269 of the line's 0.62 expected goals in those five-on-five minutes.
- Vancouver left San Jose last in the Pacific at 3-11-1. Against that opponent, San Jose matched its season five-on-five rate of 42.0 percent through 14 games.
Three stars#
- Macklin Celebrini. Six shots, a San Jose high, and the center of the third-period line that generated 0.62 expected goals for and 0.18 against in 4.9 five-on-five minutes, including the tying goal he set up.
- Elias Pettersson. The Vancouver center set up the opener and had the widest expected-goals margin of any forward in the game, on ice for 1.39 for and 0.41 against in 10.9 five-on-five minutes.
- Will Smith. Two goals on shots worth 0.33 expected goals combined, the power-play goal that started it and the tying goal after he joined Celebrini's line.
The NHL's order was Smith, Marchment, Pettersson. I have Pettersson second where the league has him third, and Marchment off the list where the league has him second. Marchment's winner carried a 0.028 goal probability; across his five-on-five shifts in this game, he was on ice for 0.14 expected goals for and 1.09 against.
Sources: NHL game data, box score and shift charts; MoneyPuck game and player data; NHL standings and team statistics.
