LGD – Game 52: Leafs @ Avalanche – Corsi splits

Another short LGD (but some fun statistical stuff) as there’s a quick turnaround between Phoenix and Colorado. The Avalanche are the brethren of the Leafs: got off to a super hot start (the Avs were the only team the Leafs lost to in their first seven games) mostly fuelled by goaltending, but both teams have come back to earth since October. On November 1st, the teams were first in their respective conferences in points rate. 

Since November 1st, Colorado is 6th in the West and Toronto is 14th—ahead of just the New York Islanders and Buffalo—in the East. Interestingly, it’s been the Avalanche who have a much more noticeable drop in goaltending: their team save percentage was .955 in October and is just .912 since. The Leafs were .937 after October and .913 from November 1st through last night.

The Avalanche are still a young team with a lot of talented young forwards, and still hold on to third spot in a relatively weak Central Division and will probably stay there with no real competition. We’re not sure about the Leafs just yet, but Anthony Petrelli had a great characterization in his Leafs Notebook post from this week.

Lots of stuff below today. This is all coming from various parts of Extra Skater, just a fantastic ressource. If you are unfamiliar what Corsi Close measures, check out this tutorial from Justin Azevedo.

Here’s Anthony’s characterization:

It’s tough to disagree. The Leafs are picking a good half season to slump. It’s odd to say “slump” in the midst of a five-game win streak, but it’s this five-game win streak which is why the Leafs have more points since November 1 than Buffalo (The Leafs have 37 points in 37 games, the Sabres have 28 in 32 games). Toronto is a team that should expect to win about half its games over an 82-game season, so you’re essentially dealing with a weighted coin. There’s a great statistical anecdote about a college professor who asked his students to flip a coin 200 times in a row and record the result. He could tell when students had faked the assignment because some of them didn’t account for long streaks. Never expect to go win-loss-win-loss-win-loss for a .500 season. If I wanted sports fans to understand any two concepts, it would be regression to the mean, and probabilistic distributions. You don’t need a lot of math or numbers to understand those concepts, just know that one great month, record-wise, doesn’t always mean a great second month. A lot of fans mix up “what has happened” with “what can plausibly happen next” without looking back to seeing whether anything is sustainable or even remotely indicative of the future.

And, really, when comparing the Avalanche and Maple Leafs from October and from November-on, you’re looking at two completely different teams in both cases. As a wise man once said:

Perhaps Colorado was doing some good things in October when they jumped out to their record reminiscent of 2001, but all it takes is an elementary understanding of sports to know that some of them won’t last:

Anyway, onto what we can expect from this game…

Colorado isn’t a real good possession team this season, but they’ve done okay against Eastern teams while at home. Corsi Close % is a good approximation of how good a team is at controlling the puck, and when comparing it to a team’s record you can see which squads are sort of living off goaltending, but they’re very rarely split. Tyler Dellow did some good work looking at how teams play on days of rest. I put together a simple chart on the Avalanche Corsi Close splits based on location and opposition:

Basically, you’re looking at a team that’s pretty good at home against the lousy Conference. As you might expect, they play their worst on the road against the West, but to them, the Conference has been more important than the location. Similar to the Leafs in that regard:

The caveat is the small sample which makes it tough to decipher between our four situations. The obvious (and predictable) trend is that the Leafs play better against the East than the West. I teased this on Twitter a little. Basically, I’ve arranged Leafs games from worst to best, but Corsi Close % in the game, and highlighted them by colour:

 

The takeaway is that there’s a lot of blue on the left (bad) side and a lot of red on the right (good) side. Being the second half of a back-to-back as well, this doesn’t look good from the outset, and we’ll need a big game out of James Reimer to stay in this. But that’s not a new theme. It’s just more important that he does it in this game than he does in any other game. The Avalanche are on two days rest and the Leafs are on none. Historically, the home team has gotten 53% of the overall Corsi events in these games, meaning they’ve had the puck 53% of the time, and the split would be greater if you only looked at “close” situations.

There’s question about long-term sustainability, but I find those numbers against the East and West interesting and splits aren’t often brought up in the analytical world. Extra Skater has game-by-game team logs going back to the 2011-2012 season, so we’re at the point where just curious people without programming knowledge can mess around and find cool things.

Toronto’s lineup should be known a little before game time. There isn’t much to discuss by way of the lineup. Should be the same one. Here’s Colorado’s:

Gabriel Landeskog – Paul Stastny – Nathan MacKinnon
Ryan O’Reilly – Matt Duchene – Jamie McGinn
Cody McLeod – John Mitchell – Max Talbot
Patrick Bordeleau – Marc-Andre Cliche – PA Parenteau

Jan Hejda – Erik Johnson
Ryan Wilson – Tyson Barrie
Nick Holden – Andre Benoit

The Top Six is actually quite good. If you want to look at a difference in possession between Colorado and a team above 50% like, say, Anaheim, look at it this way:

Hejda has always been an underrated defenceman and he’s helped Erik Johnson along tremendously, who is probably having a bit of a career year on his own. Those two are getting a heavy share of ice-time as well as defensive zone starts. Patrick Roy has them playing heavy zone time rather than matching them up specifically against individual opponents.

That top line with Landeskog and Stastny are also soaking up tough minutes and doing very well in them. It’s been a tough learning curve for Nate MacKinnon, but Gabriel Landeskog has been an awesome possession guy for a couple of years now.

So if Toronto’s going to generate any play, it will probably be against those bottom two lines.

Lots of stuff on the site today. We have Steve’s LFR, our 16-20 prospects ranked midterm from Justin, and Blake Murphy’s fun take on small sample sizes. Game story today, plus some other stuff and splits on Cody Franson tomorrow. Maple Leafs and Colorado start at 9 Eastern on Sportsnet Ontario. Lineup information from Daily Faceoff.