This is a deep dive into the statistics from the October 12 season-opening game between the Leafs and the Senators.  There is more to this game than the 5-4 OT score.

Terminology

Shots means all shots that happened while the player was on the ice, also known as Corsi.  Shots on goal are not under consideration here.  Fenwick, or all unblocked shots, is also not considered.

Scoring Chances (SC) and High Danger Scoring Chances (HDSC) are from Natural Stat Trick and are derived using the War on Ice formula of location and shot type.

Everything discussed is about five-on-five play only.

Looking Backwards not Forwards

A set of single game statistics and the visualizations taken from it are snapshots of time.  They are, more than anything, a check against the fallible memory and perception of the viewer.  They are a true record of what happened.

This snapshot can not predict the future.  It can't tell you the true nature of an individual player. You generally cannot even guess if you're looking at a normal result or an aberration.

Quality of Competition and Teammates

Unlike in a full season, where who you play with and against level out to varying degrees, one game's performance can be heavily influenced by both factors.

Matchups, or the competition faced, mean a lot, even more so in a road game, where the opposing team drives most of that matching.

The Game

There's more than one way to look at a game.

You can see that the Leafs won the shot differential battle from either graph, and both show you when they pulled away from the Senators.  The double line graph might better show the rate of shots for each side, but the push/pull graph gives you a better feel for the ebb and flow of the game.

The Players

This graph gives you a look at both the scale of a player's impact (small total numbers of shots don't matter to the game outcome as much as larger numbers) and how he did relative to the other players.  So while Milan Michalek and Zach Hyman are on the same line, and therefore have similar percentages, Hyman had the greater impact since he’s farther to the right.

Zach Hyman, Auston Matthews and William Nylander

Hyman played one preseason game where he put up results that were astonishing, but this was an NHL game with one of the best defenders in the league trying to stop him in the offensive zone.

Hyman faced Erik Karlsson and Mark Methot for 7:22 and 6:10 minutes, about three minutes more than he faced the second pair of Dion Phaneuf and Codi Ceci.  So he achieved a shot differential of 24-5 against the best defenders the Senators had.

He did that with some very good linemates.  Nylander and Matthews each had shots of 24-7 and 24-8 respectively, and collectively they drove the most offence on the team of any forwards.  They almost matched the top four defenders who played a lot more minutes.

Matthews had 13:19 minutes, the most of any centre, so this line were played as the top line, and matched by the Senator's defenders as the top line.

The Senators matched Kyle Turris's line against them the most, but they also saw the Brassard and Pageau lines nearly as much, which means Matthews had opportunities against lower quality forwards, more than he did lower quality defenders.

None of the Senators managed a positive shot differential against Matthews, but Jean-Gabriel Pageau came closest, proving once again that the toughest line to play against is not always the top line on the team.

One thing to watch for in the next road game is how much the matching changes.  If teams do match hard against Matthews, that leaves Tyler Bozak and his new winger Mitch Marner with more room to play.

James van Riemsdyk, Tyler Bozak and Mitch Marner

At 12:21, Bozak and his line were on the ice almost as much as Matthews.  This should be expected to be the standard practice all season, making line number for the top three virtually meaningless.

Generally what are called "zone starts", but are in fact faceoff location data, are not something that has a meaningful impact on results unless they are extreme.  Most shifts start on the fly.

In this game, Bozak had a very high percentage of offensive zone faceoffs and the fourth line had a very low percentage.  This is worth considering as seasoning on the other results, but it shouldn’t be seen as the major reason driving those results.

Bozak had a shot differential of 15-11, and his wingers were almost identical.  They mostly saw Phaneuf and Ceci, with Methot and Karlsson on the ice against them about three minutes less, so the mirror image of Matthews defensive competition.

They faced the top two forward lines about equally, Pageau's a little less, and the fourth line almost not at all.  No one had a very large positive differential against them, but  the top two lines kept them at about half and half shots for and against, while Bozak played very effectively against Pageau.

Milan Michalek, Nazem Kadri and Leo Komarov

Kadri played just under 12 minutes, Komarov just over, but Michalek had only nine, so there was some line shuffling going on between Kadri's line and the fourth line.  Connor Brown drew in for some shifts at the end of the game. Given the power play time in the game, Kadri playing the least five-on-five minutes is to be expected.

Kadri had a shot differential of 18-6, and his regular wingers were about the same, while Brown had more shots against.

They played evenly against the top four defenders at four minutes and change each, but they were much more heavily matched against Turris's line than any other forward line.  However, Kadri dominated all opponents fairly equally.

Kadri was seen by the Senators as the top line in terms of forward matching, but they focused on Matthews and Kadri in differing ways and largely left Bozak to flourish offensively without heavy opposition.  Mitch Marner had a lot of good offensive chances in this game.

Matt Martin, Peter Holland and Connor Brown

Holland played just under 11 minutes, Brown just over, and Martin had just over nine, mirroring the pattern on Kadri's line.

Holland’s shot differential was 12-11, Brown was 13-15 and Martin was 10-12. By looking at the raw numbers when icetime and shots are low, you don't get misled by dramatic looking percentage differences.  Martin’s 45% and Holland’s 52% are only three shots different.

This is where the fairly dramatic faceoff location difference should be taken into account.  It's reasonable to infer that Holland spent a lot more time in the defensive zone trying to get out while Bozak started out in the offensive zone and still had the same number of shots against, albeit in more minutes.

Holland faced the Senators' third pair defenders more than anyone else and did very well, but he was caved in by the top pair, with a differential of 2-6 against Karlsson and 1-4 against Methot.  They were in tough against the Brassard line, but did fine against Pageau and against the Phaneuf-Ceci defence pairing.

It boils down to: Fourth line done over by Erik Karlsson, news at eleven.

Morgan Rielly and Martin Marincin or Nikita Zaitsev

Rielly started out the game with Marincin, but due to the injury to Matt Hunwick, he played more minutes with Zaitsev.

This group were one, two and third on the team in minutes played.  Rielly: 22:03, Zaitsev: 19:46, Marincin: 16:58.

Rielly's overall shot numbers were 26-19, which are the poorest shots against of any defender, and with Marincin he had 5-7, while with Zaitsev it was 9-6.

Rielly played just over six minutes with Matthews, just under with Kadri, and around five with the other two lines, so no one set of linemates dominated his night, but he saw mostly the top two pairs and lines as competition.

He had positive shot differentials against most opponents, except he struggled against the Senators’ fourth line, and curiously, Phaneuf.  Rielly’s shots were 5-10 against Phaneuf, so I guess he knows his old teammate very well.

Zaitsev had shots of 25-13, and he fared better with Marincin with 10-1 than he did with Rielly.  He played 1:28 with Hunwick, so he was not a factor in what seemed to be a nervous start for Zaitsev that saw him one of the poorest shot differentials on the team after the first period.

Zaitsev saw a lot of Karlsson and the numbers were heavily in his favour: 13-6.  Only Matthews and Nylander had better results while Karlsson was on the ice.   Zaitsev had trouble with Pageau's line, but was otherwise solid.

Marincin's overall shot numbers were 24-16 and he was, as shown, better with Zaitsev than Rielly.  He also played a few shifts with Connor Carrick and was very good there too.

Marincin saw much less of Karlsson than did Zaitsev and fared more poorly.  He only had real difficulty against Brassard and Zack Smith, however.

Jake Gardiner and Connor Carrick

Gardiner had shots of 27-12 and Carrick had 29-11, and they played almost all of their shifts together barring Carrick's play with Marincin.

They played over seven minutes behind the Bozak line, marking them as the only defenders to play with a specific forward line.  Their shots as a unit were 10-4.

Gardiner and Carrick also spent around four minutes with Matthews and three each with Kadri and Holland.   They had around 50-50 shots while out with Holland's line, but otherwise played extremely well.

So that's who did what with whom and against whom.

There was one set offensive unit in Bozak plus Gardiner and Carrick, and one set checking unit in Holland plus two of the other three defenders.

Matthews and Kadri played largely similar games.  Kadri had a few more defensive faceoffs, not enough to mean much, and Matthews had more minutes.  Kadri saw tougher forward competition, but Matthews faced the top defence pair more.

As far as individual contributions to offence go, Marner (8), Marincin (8), Carrick (6) and Matthews (6) shot the puck more than anyone else, while Brown had 0 individual shots for.

As an exercise in maximizing offensive punch, this deployment worked great. Defensively the team was less effective up and down the lineup.

Kadri and Matthews had very positive scoring chance differentials (scoring chances being those shots that have a greater expectation of becoming goals), and Matthews was only on the ice for two against.  The Holland and Bozak lines and Rielly, Marincin and Zaitsev, were the culprits in allowing the most.

While The Leafs outshot the Senators by a great deal, the proportion of scoring chances is much less dramatic.  They let the Senators get more high-danger chances, and the Senators converted on some of them.  Add in a poor goaltending performance, and the overtime loss becomes easier to rationalize out of some very impressive surface numbers.

For a first game with a lot of rookies and a new goalie, it was impressive, high-speed hockey.  The next game, a home game, will show how Babcock wants to match his lines up.

Acknowledgements

All numbers come from Natural Stat Trick, unless marked otherwise in charts, with some additional information from the NHL.com shift chart and Corsica Hockey for the individual shots for.

Analysis