Possession and the Preseason


The preseason is over. It’s time for teams to make their final decisions, and get ready for their season-opening puck drop. For the Toronto Maple Leafs, that’s Wednesday. But before people get to tearing apart the regular season games for every number we can squeeze out of them, I’m going to do that very thing to the preseason.

The Games

First and foremost, we’re missing one game in this equation; the first one. The first question from many will probably be “Why?” and it’s a simple answer – the Leafs and Flyers faced off in London, and the NHL’s web people basically laughed and gave the “you’re on your own”. There are no stats to scrape. The recap is shorter than this paragraph. I suppose one could find the archive and try to do everything themselves, but I’m not that stressed about it.

Here’s how the Leafs fared as a team in the other seven games:

Date Opponent Score Corsi For Corsi Against Corsi For%
23-Sep Philadelphia 4-0 W 45 38 54.2
24-Sep Ottawa (H) 4-3 SOL 45 56 44.6
24-Sep Ottawa (A) 3-2 L 35 56 38.5
26-Sep Buffalo 6-4 W 27 47 36.5
28-Sep Buffalo 3-2 SOW 50 58 46.3
29-Sep Detroit 3-0 L 25 16 60.9
03-Oct Detroit 5-1 W 43 37 53.8
TOTAL   22 GF 16 GA 270 308 46.7

Peaks and valleys, man. The Leafs end up with a 46.7 Corsi For percentage at even strength over the course of the recorded preseason, though I feel like the first game likely brings them to around 48% (all situations shots on goal were 31 to 24) it were taken into account.

Toronto’s worst showings, predictably, were in the Split Squad game that was in Ottawa (where they basically iced an AHL team) and the following game against Buffalo (slightly better roster, but still pretty bad and against the stronger half of the Sabres. 

By the numbers, Toronto’s best game was their shutout loss to the Wings on Monday, but I wonder how much of that is score effects. For those unfamiliar with score effects, the idea is that teams that lead by multiple goals will go into a defensive shell, while their opponents will take more, lower percentage shots in an effort to claw back. Toronto, had momentum in the second period but really began to put pressure on after Nick Jensen’s 1-0 goal, and carried it throughout most of the game.

Perhaps more impressive was last night’s effort; they it was their third best by CF%, but they maintained it despite the wide margin – and in fact had their highest-pressure period in the third, putting double the pucks on net that Detroit did.

It’s hard to say whether this is the pattern to expect from the Leafs in the regular season, but it would be a step in the right direction if it was. The Leafs aren’t going to be world beaters at 47 to 49 percent, but it’s a heck of a lot better than 42.8, or 44.0 from 2012/13. If the goaltending is still good on the whole, this team could fare better in the standings than last year.

The Players

  • I think that the most overrated player coming out of training camp is probably Josh Leivo. It’s the same basic point that I made in my TLN Top Prospects post on him; people are high on what he can be, but he isn’t there yet. People see him turning into Joffrey Lupul. He’s not yet Joffrey Lupul. Leivo posted a 39.3 CF% with the Leafs this preseason; worse than anybody who played three recorded games. Leivo, of course, played five, and put up three points. A shooting percentage of 33% made him look like a sniper, but this is a guy who managed to take three shots in five games was likely detrimental to the flow of the game. Let him rip it up with the Marlies – if this is any indication, he’s not to be trusted yet with top six NHL minutes. If you can’t put him high in the lineup, there isn’t another place to put him.
  • But the Leafs still need to figure out who starts on the right wing on line two, no? David Clarkson is hurt, and so is David Booth. The latter is definitely going to miss time, and the Chief of the Water Bottle Police is a question mark. My thoughts? This is where you reward William Nylander. Yes, there are parts of his game where it’s evident that he’s not ready for prime time. He makes the odd foolish giveaway, and he doesn’t seem up to snuff in terms of his physical frame. But at 52.3%, he was one of Toronto’s best possession players. He was a positive in four of six games, and generates a lot of offence on his own. I don’t know if you can trust him to put up points, but he doesn’t appear to be a liability yet. Plus, it buys you time to decide what league you want him in.
  • Brandon Kozun is the biggest breakout story at the end of the preseason, as it looks like he’ll be making the Leafs roster. This is an interesting one – I had him pegged behind Spencer Abbott, who got just one game before being sent down, in the Marlies depth chart. With that said, Kozun impressed many with his speed and lit up the scoreboard with a three point night tonight. The fancy stats have him just below the curve in terms of relative possession – the team is 0.9% better with him off the ice, but I think that his second of six games plays into that. In the Ottawa split-squad game, Kozun was on for 10 attempts for and 22 attempts against. Pull that out, and he’s even.
  • By the way, 22 attempts against? That’s only two fewer than Carter Ashton saw on the ice, for either team, this entire preseason. It’s hard to say that he was in a position to succeed when he was given two games and next to no ice time. Ashton was a horrendous 23.1% in his first game, but was also on a line with Colton Orr and Trevor Smith. He was more respectable last night, but still didn’t play very much. I assume he will be waived in the coming days, but it would have been nice to see him, you know, have a shot at making the team in a hockey-playing role.
  • Dion Phaneuf‘s game seemed to go the way of his defensive partners. With Franson and Robidas? A shade over 59%. With Polak and Granberg? Under 30%. The Leafs can’t keep trying him with immobile defensive defencemen – it doesn’t work. It’s a small sample, but seems to be a recurring pattern in Phaneuf’s career. Robidas looked both mobile and positionally intelligent tonight, however, so that may be a pair we see a lot of to start.
  • If I’m the Leafs, I’m not opposed to a third line of Daniel Winnik, Mike Santorelli, and Leo Komarov, at least to start the year. Santorelli and Komarov were Toronto’s best possession forwards. All three are capable of contributing a bit offensively, and all three are penalty killers. If Carlyle wants to recreate the magic of his Anaheim “shutdown line”, this is probably the closest thing he’ll get to it right now. 
  • The last spot in the defensive rotation seems to be pretty contested right now. Some are in favour of Korbinian Holzer, while others like Stuart Percy. Of the two, Holzer put up the better full preseason numbers, getting 50.9% of the attempts while on the ice (5.18% above the curve). Percy, on the other hand, was 1% below the curve at 45.9%. I’ve been saying that I preferred Percy’s game more throughout the preseason, but it didn’t occur to me how well Holzer actually performed. Even when he was bad, he was ahead of the team’s relative curve for in all four of his games. Percy had a team-best game against the Sabres on the 26th (Toronto’s only 50% player), but was negative relative to the team in every other game. This is where I miss having easily accessible zone start percentages; I wonder how well they come into play. In any event, this is enough to make me feel that it would be best for Percy to go down, and have Holzer run as the extra; he had a bad first go with the Leafs, but maybe he’ll be better off on Pair 3 rather than facing elite forwards with Phaneuf.


It’s hard to have anything super concrete. After all, we’re dealing with one to six game samples here. But it’s interesting to comb through this stuff and look at after the fact. Take from it what you will.

The Data

By the way, here is the big chart of possession, sorted by games played. 

GP Corsi For Corsi Against Corsi For% Relative Corsi
W. Nylander 6 67 61 52.3 7.2
B. Kozun 6 58 68 46.0 -0.8
J. Gardiner 5 80 71 52.9 8.4
J. Van Riemsdyk 5 66 59 52.8 7.7
N. Kadri 5 62 67 48.0 1.7
P. Holland 5 49 53 48.0 1.6
S. Percy 5 56 66 45.9 -1.0
J. Leivo 5 46 71 39.3 -9.2
M. Santorelli 4 43 36 54.4 8.9
L. Komarov 4 27 23 54.0 7.9
K. Holzer 4 54 52 50.9 5.1
J. Lupul 4 49 49 50.0 3.9
M. Rielly 4 54 55 49.5 3.4
D. Winnik 4 23 25 47.9 1.3
R. Polak 4 32 36 47.0 0.3
M. Frattin 4 34 39 46.5 -0.1
D. Phaneuf 4 47 54 46.5 -0.2
S. Carrick 3 17 17 50.0 3.4
C. Orr 3 16 17 48.4 1.8
T. Bozak 3 34 38 47.2 0.5
C. Franson 3 35 42 45.4 -1.4
P. Kessel 3 41 50 45.0 -1.9
P. Granberg 3 39 50 43.8 -3.4
H. Tallinder 2 20 18 52.6 6.3
G. McKegg 2 20 19 51.2 4.8
D. Booth 2 19 20 48.7 2.1
V. Loov 2 24 32 42.8 -4.2
T. Smith 2 16 30 34.7 -12.9
C. Ashton 2 8 16 33.3 -13.9
T. Bodie 2 6 13 31.5 -15.6
C. McMillan 2 6 16 27.2 -20.2
C. Brown 1 13 7 65.0 18.9
D. Broll 1 7 4 63.6 17.2
S. Robidas 1 19 12 61.2 15.4
S. Abbott 1 8 6 57.1 10.6
J. Devane 1 4 4 50.0 3.3
P. Kontiola 1 10 10 50.0 3.4
B. Mikkelson 1 15 16 48.3 1.7
D. Clarkson 1 5 6 45.4 -1.2
E. Knodel 1 9 11 45.0 -1.7
K. Marshall 1 9 12 42.8 -4.0
A. MacWilliam 1 5 15 25.0 -22.4
T. Nilsson 1 4 16 20.0 -27.6
F. McLaren 1 2 9 18.1 -29.0
B. Vail 1 2 11 15.3 -32.0

Individual games can be found on my Twitter. Here are links to individual tweets:

Sept 23 / Sept 24 / Sept 26 / Sept 28 / Sept 29 / Oct 3