Overview
Using my xG model, I also wanted to "sanity check" it, by literally just checking who the top players and teams are in xGF, xGA, xG% and ixG over various timeframes. I was an incredibly simple methodology I simply grouped by team, or player and season, or team and season, whatever Iwas grouping by, then took the sum of xGF, ixG, or xGA and sorted by that stat. Most of the work was already done creating that model, which if you would like to read how I did that, the writeup can be found here. This will just be going over my sanity checks and showing the top players, and new vizzes.
Results
The first thing I did was simply look at which teams were the best and worst by xGF. I initially got all 2012-13 teams in the bottom and was incredibly confused for a second until I remembered that was the lockout season and I didnt adjust for games, so after adjusting for GP, here are the results:
Buffalo in their tank for McDAvid year being the worst by a noticable margin is very unsurprising. Seeing the Devils from that year below Edmonton or Arizona surprised me, but man did that team, and the basically identical team the following year struggle for offense. At least the Buffalo team was so horrid they had to change most of their roster up. Not that they were good in well... any of the surrounding years, but still. The raelly bad COL era teams being on that list don't shock me, and lots of early analytics era Wild teams, also unsurprising, I despised watching those games as a young pens fan because they would ALWAYS end up 2-1 or 1-0 it felt like. Then you see a lot more recent teams at the top, which I think is good because there has been an increase in scoring, so there should be an increase in xG as well. Looking at most of the teams you see a lot of cup contenders, Florida, Carolina, Vegas, Toronto, Edmonton, Colorado. The top teams aren't as interesting to me as the bottom ones. There aren't any teams that I remember as clearly being the best team of the last 15 years, other than the generic contenders of recent years. Washington Carolina and San Jose being the only early analytics era teams in the list surprise me, but that was probably the most interesting part.
Looking at the worst xGA teams, you see a similar/inverse trend. Almost all of the worst teams are simply the bad teams from the past few years when scoring is at its highest. It maybe would have been interesting to look at the "true" worst teams and also adjust for league scoring, but then that kind of takes away from the point of this being a sanity check, and yeah, it makes sense that all of the "worst" defensive teams come from the years when scoring is at a high point. Some bad older teams in there too, the early Tavares Isles, some bad Habs/Canes years as well. Then the best defensive teams tend to skew a lot older, but man I remember also despising those Broduer era Devils teams. Their entire system was just built to suffocate you, and then they have one of the best Gs of all time behind them, just so infuriating. Then you just see a lot of really good teams, those early analytics era Detroit teams were nasty, and fun with Datsyuk. LA, Boston both had their years, and recent Carolina is just doing recent Carolina things. Tampa as well had some great years of course, seeing them here twice didn't shock me.
Now looking at xGF% by teams, you see a lot of who you would expect. The BUF team truly was the worst team in the modern era, and may have set that franchise back a whole decade. An odd number of 2023 teams, like yeah scoring was up, so I would also expect xGF%s to have marginally more variance, but I wasn't expecting that much variance. But I can't really say it's wrong because yeah, those teams were BAD that year. Then you see some other Buffalo teams, recent Detroit, old Florida and Edmonton, and yeah those make sense. Though the 23-24 habe, were not good, but I did not think they'd be THAT bad. Then the best teams, a lot of recent CAR because their stupid, silly, obnoxious system. Colorado, Florida, Edmonton, Toronto, all very good teams, and I love seeing the old Red Wings teams up here. I have very fond memories of hating thsoe teams because they were so damn good, but then winning the cup in 09 as a pens fan is one of the best memories of my early life. The rest of the teams all mostly make sense too, Chicago won some cups, 2010-18 San Jose immediately became the best stretch of teams to never win a cup the moment Ovi won with the caps. NJ in 23 surprised me, I knew they had a really good year a few years ago, but don't remember which. Every year I feel like they're the team that I look at on paper and am terrified of, just waiting for them to put it all together with a legit goalie and not get injured and I think they have the chance to win a cup.
Now, moving on to looking at players, we look at the list of the top individual seasons in the analytics era. I really have 1 comment. Alex Ovechkin: good at hockey. The far and away two best seasons in the last 15, and another in the top 10. Those were clearly his peak years, but with the longevity he's had, truly the GOAT goal scorer (long before he beat the record too). After the next 3 of Eric Staal's career year, an insane Matthews year which was probably the best non-Ovi goal scoring season in the past decade, and Hyman's 0 ft average shot distance, McDrai fueled season there really isn't much to comment on. It is so flat towawrds the tail end of this graph that I feel like only a few shots could have made a difference of a few rankings spots, and all of those players were very good at some point in time in their careers. Shout out Dany Heatly and Big Jeff.
Looking at players careers as a whole, these are slightly confusing graphs, because I wanted to look at both the top careers total and top per game careers. Then, mostly for the per game one, I wanted to show the total number of xG in the color because it is far more impressive doing what Ovi did in his career than what Matthews or Tkachuk have done SO FAR. But yeah, Ovi being the undisputed ixG GOAT was incredibly unsurprising. His per game pace being that high for that long is just freakish. The rest of the totals also make sense, a bunch of old dudes who played forever and were starts for most of the 2010s. Tavares, Pavs, Patty Kane, Sid the Kid, Phil the Thrill, Stamkos not being higher is surprising, he's a bit younger than some of these guys, but not by much. Burns being the top d man, and also still playing to add on to his total is undurprising and incredibly funny to me. Now the per game chart is more interesting to me. I don't see Tkachuk having a super long career like Ovi with his playstyle, but I *could* see Matthews maintaining that pace for a good while longer and potentially challenging Ovi for the GOAT goal scorer status. However, he would have to play forever, and have far fewer injuries than he's had the past few years for that too happen. Then the rest of the list is a bunch of the elite goal scorers from the previious list, Sid, Tavares, Pavs. Parise surprised me. Then a bunch of young stars that with time I could see them pushing guys from the other list, McDavid and Mack of course, Svech would probably need to turn it up another notch, but I love seeing him on here, Guenzel, Pasta, and Cole are probably the 3 best pure goal scorers of the ~2018-2025 era not named Ovi or Matthews. I love seeing Lee and Gally on here as well, they won't play enough games to get the career accolades, but elite net front presences knew their roles and were incredible at it. Would have really loved to see Hornqvist on here, and he probably did come close in some seasons, but he unfortunately played too long to remain at the top of the per game list.
Now for the best shooters. I know shoting percentage above expected is flakey and I don't like using it for analysis of a whole team (other than saying they'll probably regress) or an individual player season, but if you can prove to consistently score above expected for your career, I have to assume you will probably continue to do it. This list was interesting to me because there is a mix of elite talents who generate a ton of xG and score a ton, Drai, Pasta, EP (theortically), Laine, Kuch, Marchand. But then a bunch of *good* players who had long careers but were never elite at generating chances, but finished a lot of them. Riqo, Robo is getting to star territory, but more for his overall game, not pure shooting, Hintz, Vrana. Then you do have your just pure shooters, Laine and Connor both. Laine was expected to do this coming into the league, but then Connor has always been an enigma to me looking at his stats and watching him play. He always gets all of the ice time, and produces incredibly well with it, but has relatively poor underlying numbers and doesn't wow me as much as guys like Ehlers or Scheifle did on his team. Also funny seeing Scheifele, Connor, and Laine on here, and Ehlers would probably be on a list of worst shooters it feels like. Was gonna shoutout Jiri Hudler for being on here too, but maybe he isn't the guy to shoutout... so shoutout Mark Stone instead, my stickchecking GOAT.
I also have the top players for 23-24, but really have no comments other than yeah, those are good players. Matthews and Hyman had insane seasons. Forsberg, Pasta, Mack, all stars with great goal scoring resumes. I guess Pinto and Vilardi being in the per GP list was a bit surprising. Not that those two are bad players, I love Vilardi, but just that I would not have predicted them to be top 25 in almost any stat. Val Nuke also very good player when he can be on the ice, one of my favorites to watch. Then there's Sid, in both lists still, shooting and goal scoring still not being his strongest attributes, yet he's still top 25 in the league at it at the age of 36. Top 10 C in the game still, easily.
Data
I did get most of my shot data from evolving hockey's querey feature. I hae tried to run scrapers to get data directly from the NHL api, but it's always been a mess every time I've tried to build that scraper in particular. I've built many others succesfully, but just never one for the NHL API. So the EH queries are the next best thing, and they come mostly pre-cleaned. Though they were behind a paywall, I feel like I've done enough data manipulation that I can share this, especially in a portfolio style website that I don't plan on being profitable from.