Since I am still yet unready to dive into a look at the Winnipeg Jets’ past season due to the team’s poor performance in the post-season, I thought I would sink my teeth into something else to distract from the misery. Thus, here I am again looking at how well advanced statistics can tell the story of a playoff series and how accurately expected goals can predict the final outcome.
I am well aware that the expected goals (xG) stat is never really supposed to be dead-on with their numbers because while it is trying to reflect the likeliness of a goal being scored, its formula is using an NHL average shooter and an NHL average goaltender in its calculations to determine how often a puck would enter the net. While the xG stat doesn’t take into consideration the impact elite skaters like Nathan MacKinnon, Connor McDavid, & Nikita Kucherov have on goal scoring, I still feel that it does an adequate job at reflecting the actual flow of play. Multiple low danger shot attempts will barely move the needle on the expected goal numbers, so teams need to get quality opportunities to actual improve their stats. That being said, there are always situations that make xG’s a little worse at telling how a contest played out, such as when one team gets an early & healthy lead in a game, making them sit back and allow the trailing team to control more of the puck. At the end of those matches, it isn’t uncommon to see the losing team win the xG battle despite never really being in the game.
Flaws and all, since it is the best stat I have to work with, I decided to go ahead and look at a comparison between the expected & actual goals for each playoff series. Starting off by using the information from the hockey nerds over at MoneyPuck to create the graphics below, I got to work looking at the data to see what jumped out to me.
With only one game left to play in the opening round (Vegas @ Dallas in deciding game 7), I have opted to go from the most statistically lop-sided series to the closest. Unfortunately, the leaves me starting off with my squad’s dismal performance in Round 1. ***planning on updating article after that VEG-DAL game***
The Colorado – Winnipeg series…
The fact that the stats indicate that the Avalanche dominated their series against the Jets certainly isn’t surprising to me, but seeing the down-right awful numbers is still unenjoyable. Winnipeg’s offense had 2 decent contests and 1 good game, but their defense could only stop Colorado from 4 expected goals on one occasion. While the numbers also support Connor Hellebuyck had to play a lot better (actual goals higher than expected goals), they also highlight the absolute train-wreck the NHL’s best defensive club in the regular season had become in the post-season. Like most goaltenders, Bucky is better at stopping low & medium danger attempts (has elite numbers over his career), but his style of play has historically led to high danger looks beating him at a higher rate when compared to other elite goalies (Shesterkin, Demko, Vasilevskiy, etc). Winnipeg did such a great job at limiting these opportunities during the season, but seemed unable to stop the Avalanche when their season was on the line.
The Winnipeg Jets were the only team to lose the xG battle in each & every playoff contest, while finishing with the worst expected (-1.77) & actual goal differentials (-2.60) in the post-season. Of course that means the Colorado Avalanche are the NHL’s most dominant playoff team at this point, as they wait to see whether they will host the Knights or travel to Dallas to start round 2.
Los Angeles – Edmonton series…
The Avs weren’t the only team to explode offensively in the 1st round, as the Edmonton Oilers seemed to have no problems creating chances against the Los Angeles Kings. LA was similarly built as Winnipeg (relied on strong defense to aid below average offenses in regular season), but this strategy didn’t pay dividends for the Kings this year. The Oilers won the xG battle in 4 out of 5 games, though they did lose a game they weren’t expected to (Gm 2) and won another the numbers say they shouldn’t have (Gm 4). Edmonton also did a great job of reducing chances against, with Los Angeles only slightly above Winnipeg for the least expected goals in the playoffs.
The Oilers also put up the highest xG number in one playoff game (8.49) though they were only able to score 6 in the contest. They also had the 7th highest xG of the post-season with 5.08 in the opening match. Oddly, when the Kings played their perfect game in the 4th match and limited the high-powered Oilers to the 2nd lowest xG number of the post-season (1.43), they still ended up losing thanks to a shutout performance by Stuart Skinner. This was one series where MoneyPuck was almost dead-on with their goal projections, with LA being -0.04 goals away on average and Edmonton being +0.02 from expected.
Washington – NY Rangers series…
The only series that ended up being a sweep, the New York Rangers won the xG battle in 3 of the 4 contests. The Washington Capitals offense disappeared in the opening match, but they really did their best to try to win the 2nd game at Madison Square by putting up higher xG numbers. The NY Rangers did what they needed defensively and in net during the series, allowing an average of +1.00 Goals Saved Above Average per game, while also scoring close to expectations (+0.11).
Washington was able to create offense in games 2 & 3, but they struggled to penetrate the Rangers’ defensive structure in the other two contests. And when they did, they still had to find a way to beat Igor Shesterkin and that proved difficult (.929 save %).
I didn’t end up watching much of this match up personally, but it seems that the Rangers for the most part deserved an easy victory. With their breezy first series in the books, NY sits near the top of the expected & actual goal charts and will have a much tougher test when they face-off against the Carolina Hurricanes in round 2.
Tampa Bay – Florida series…..
Having watched most of this series, I kind of expected the numbers to be a bit closer, but it seems that the Florida Panthers routinely won the xG battle (4 out of 5 games). I believe this is the first and only series that all of the wins actually went to the winner of MoneyPuck’s expected goal predictions. That being said, it was a very good match up to watch with a lot of “robberies” by both goaltenders so I was a bit surprised that the expected & actual goals were fairly close when the dust settled on the series.
The Tampa Bay Lightning were only able to limit the Cats to under 3.00 xG goals once in the 5 contests, allowing them to snag their only win of the 2023-24 playoffs. Offensively, both squads were able to create dangerous looks – the Bolts were above 3.00 xG in 4 out of 5 games, while the Panthers cleared the 4.00 xG mark on three occasions. However, it was Florida that routinely found ways to beat the opposing tender and that helped them to finish with a 4 goal per game average, good enough for 3rd best in the NHL playoffs.
While this series ended up being a bit lop-sided in the win-loss column, either of those first two games could easily have gone Tampa’s way and that would have made things a lot more difficult for the Panthers. Paul Maurice will lead his Florida franchise against the Boston Bruins in what could end up being quite the goalie duel (Bobrovsky vs Swayman).
NY Islanders – Carolina series….
Another 5 game series in the books and both my eyes & the stats support that the Carolina Hurricanes were the better team. The New York Islanders barely squeaked into the playoffs only to be mis-matched against a formidable Hurricane squad and that showed up in the stats. To give NY their props, they did manage to win the xG battle in game 3 when they stifled Carolina to a post-season worst 1.11 expected goals, but that happened to be a game when Ilya Sorokin let them down in his only start. The Isles’ Semyon Varlamov did end up helping to steal game 4 from Carolina when he had a +2.71 Goals Saved Above Average (GSAA) rating. New York also had their highest offensive output that game, with their series-high 3.85 xG and tying their series-high of 3 actual goals.
The Canes are the 4th highest scoring team (3.80 per game) in the post-season, while the Islanders finished as the 4th worst scoring team (2.40 per game). The Isles had the 20th worst goals against in the regular season and their defense didn’t improve in the playoffs (allowed Carolina to have over 3.25 expected goals in 4 out of 5 games). With both the Rangers & Hurricanes producing offensively in the post-season, the next series may come down to goaltending (edge NYR +1.0 to +0.46 GSAA)
Toronto – Boston series….
One of two series to go to the 7th game, the Toronto Maple Leafs have to deal with another heart-breaking loss to the Boston Bruins. Not sure if it will make Leaf fans feel any better, but this has been the only series so far in 2023-24 where the franchise with the lower expected goals total ended up victorious (TOR +0.59 xG differential). Toronto ended up winning the xG battle in 5 out of the 7 games, but it was the play of Jeremy Swayman in the series that ended up being the difference. The B’s goalie absolutely stole games 1 & 3 with GSAA’s of +6.21 & +2.66 respectively. Besides the explosion in the opener, the Leafs had 4 other solid (above 3.00 xG) performances offensively, but hiccups in key games (4 & 7) saw those numbers much closer to 2.00 xG.
In terms of predicting the outcome of the games in this series, MoneyPuck’s expected goals had a 5-2 record with Bruins’ wins in games 1 & 3 being the misses. Boston was fairly consistent with their attack, finishing between 2.16 & 3.71 xG’s in the seven contests, but you might worry how they will match up defensively against the Florida Panthers in round 2. To date, the Bruins are allowing an average of 3.00 xG and have a 2.57 GAA, so that confirms that Swayman is playing above average and likely deserves to be the go-to starter for the time being.
Vegas – Dallas series….
The second 7 game series of the NHL’s opening round, but so far we only have data from 6 contests in our graphic. This has been an interesting back & forth affair and I expect the final entry to be just as entertaining as the first 6 chapters. The Vegas Golden Knights & Dallas Stars started the playoffs like they were trying to prove that the expected goal stat was useless, as the xG winner ended up losing the first 3 matches. The Stars held a slight advantage in the first two games in Dallas (under +0.30 xG diff) and the Knights dominated game three (+3.92 xG diff), but none came away with victories. However, a switch flipped in game 4 and now the team that wins the xG battle has won the last three contests.
This has been one of the lowest scoring matches of 2023-24 so far, with both teams’ goals per game putting them in the squeezing the stick category (VEG 2.50 & DAL 2.33). Both franchises’ goalies are playing well and have positive GSAA ratings (DAL +1.01 & VEG +0.54).
Entering into the final game, Vegas & Dallas have the 2nd smallest margin of xG differential (Knights’ +0.30), so the gap is close enough that tonight’s winner might end up leading in the expected goals category for the series. We will have to wait and see how it plays out, but either way MoneyPuck’s data seems to confirm that this has been an extremely hard-fought series.
Nashville – Vancouver series…
And we get to the final series of the opening round and what is so far the closest statistically (0.14 xG differential), despite it going less than the full seven games. MoneyPuck’s expected goals correctly projected the winner in only 50% of the games in the Nashville Predators and Vancouver Canucks series and since I didn’t watch much, I have no visual information to compare to the data above. It appears that Juuse Saros stole the 2nd game with a +2.16 GSAA rating in the contest, but both franchises’ goalies finished with positive GSAA in the series. The Nucks’ Arturs Silovs responded with a game of his own at a key point, putting up a shutout and a +3.47 GSAA rating in game 6 to help Vancouver move onto the second round.
The Preds and Canucks weren’t great offensively in the opening round with the 3rd & 4th worst goals per game average (2.00 & 2.17 GF respectively). Nashville only topped the 3.00 xG mark once in the series, while Vancouver barely eclipsed it in the first 2 contests.
While you would have thought the injury to Thatcher Demko being the biggest hurdle for the Canucks, it appears that scoring goals may move into the top priority spot. Especially when you are facing the Edmonton Oilers in an all-Canadian second round match up.
Here are some other interesting tidbits:
- 7 out of 8 playoff series had at least one game where the team with the lower expected goals won the contest.
- Only the Panthers-Lightning series was always correctly predicted by xG’s.
- The xG upsets occurred in 14 out of 42 post-season matches (33.3%) in the opening round so far.
In case you weren’t done with a desire to look at endless numbers, here are a couple more graphics showing the rankings for expected goal and actual goal differentials.
Once again our Winnipeg Jets find themselves cemented to the bottom of each table, but it should be interesting to see if any of the elite franchises (expected goals-wise) like the Colorado Avalanche, Edmonton Oilers, New York Rangers, and Florida Panthers will go on to win the Stanley Cup. Fortunately it keeps getting easier to compile this data as the NHL playoffs progress (less series, less teams, less games each round), so I might as well keep at it until some franchise lifts Lord Stanley’s Cup.
Looking for an easy way to support Arctic Ice Hockey Use our Affiliate Link when shopping hockey merch this season!
