Ok, I’ll admit it, I’m obsessed with the difference between the Golden Knights’ expected goals rate and the results they are achieving in spite of it. The longer it goes on this way, the more I think about it and the more I dig and dig trying to figure out what I’m missing.
If you aren’t aware of what I’m talking about, the Golden Knights are 12-3-2 in their last 17 games and they’ve outscored their opponents 57-46 in that span, yet NaturalStatTrick.com‘s expected goals stats have them trailing 48.7-60.9 over the same timeframe. That’s a difference of eight goals for and 15 goals against, or a 23 goal delta in just the last 17 games.
Over a small sample, we see numbers like this pop up all the time, but 17 games is pushing those boundaries. Vegas has been outscored in expected goals in 14 of the last 17, yet has come out on the right side of the real scoreboard in 10 of them. At some point, it can’t be a coincidence.
The latest data set I’ve turned to is trying to understand game state’s impact on the overall numbers. One of the constants in VGK’s recent run of success has been their ability to score first. They’ve done so in 13 of the 17 and they’ve won each of the last 10 in which they have. That’s led to Vegas being ahead on the scoreboard a lot. In fact, since February 19th, the Golden Knights have led for 477 of the 1041 minutes they’ve played. That’s 28 minutes per game or nearly half of every game.
Of course, that will have an impact on the expected goals numbers. When a team is ahead, especially a Bruce Cassidy-coached team, they are expected to play a little more conservatively while the other team pushes to tie the game back up. And, while ahead, the numbers do bear that out.
When leading, VGK have been outscored 23.1-32.3 in expected goals over the last 17 games. What’s weird though is that the actual goal total mimics those numbers almost perfectly. Vegas has scored exactly 23 times when leading and they’ve conceded 30. Thus, just 2.4 goals of the 23-goal difference in expected goals come when Vegas is ahead… which is more than 45% of the game!
When they are trailing, the numbers match up as well. In 211 minutes while behind, Vegas has been outscored 10-7 in real goals and 10.9-7.9 in expected goals. As the great Mike Goldberg used to say, virtually identical.
So, that brings us to the chunk of the game when the score is tied. In the past 17 games, the Golden Knights are outscoring their opponents an insane 27-6 when the score is even. 27-6 is an 82% goal share and that is 13% better than every other team in the NHL over the same span. To put that number in context, there are seven teams this season that are 82% or worse when on the power play.
Clearly, this is where the majority of the real goal vs. expected goal discrepancy is coming in. 27-6 in real goals is paired with 17.8-17.7 in expected goals. That’s +11 in goals for and +12 in goals against.
In the first 56 games, the Golden Knights were on the opposite end of this same stat though. With the score tied, they outscored opponents 63-53 but were expected to be up 66-45. That xG% is good enough for 2nd best in the NHL. Over the recent span, VGK sit in 20th.
Add it all up, and it shows the Golden Knights were in line for a correction. It happened, it just happened in the form of an overcorrection which is why I got so confused in the first place.
As we zoom out though, dominance in tie games has been a staple for the Golden Knights this season. Over the course of the full season, VGK rank 3rd in actual goal share at 60% and 2nd in expected goals at 57%. This is really good news heading into the playoffs where the games get much tighter.
Finally, while researching all of this, I ran into one more interesting number that adds some more context. This season, the Golden Knights are second-best in the league, behind only Boston, in fending off “strikebacks.” It’s a stat kept by the website MoreHockeyStats.com and it’s a measurement of how often a team comes from behind to win. This season, the Golden Knights have only lost 10 games in which they’ve led at any point. Eight times they’ve led by one and gone on to lose, while twice the lead was two. They’ve never had a lead of three or more and lost the game.
However, the Golden Knights have allowed 48 goals against when up by exactly one goal, the 5th most in the NHL. That means 48 times the Golden Knights have conceded a lead, yet they’ve only managed to lose 10 games in which they did so. This further plays into how great they’ve been when the game is tied.
Sure, they may give up the game-tying goal, but rarely do they allow the game to get away when they do so.