By #jontent (@yakovmironov)
Hi there! It’s your ol’ pal Jon. I feel that my content has become quite stale lately and wanted to take on something more meaty than my previous couple of posts of “Jon’s thoughts on things” and “here I am playing with a draft tool.” Not that you shouldn’t read those posts, my thoughts are always gold, but I felt like playing around with stats today.
Basically what I’ve set out to do is create my own happy little player usage chart based on the simple little four quadrant graph below.
This isn’t anything revolutionary by any means, it’s not far off what a lot of people are already doing from stats perspective already (read basically my version of a player usage chart), so I’ll try to shut up about it quickly.
Using various stats found on Corsica.Hockey, I want to create 4 quadrants…
- (Upper Right) Players I feel comfortable playing in any situation
- (Lower Right) Players who are stronger offensively and need sheltering
- (Upper Left) Players who benefit the team defensively
- (Lower Left) Players who either need time to develop or gotsa go.
I want to avoid surprises like Kadri being considered garbage or Roman Polak being an offensive dynamo, but I also want to check my biases on Leafs players I’m not particularly fond of like Matt Martin, Tyler Bozak, etc.
I also wanted to try to create single measures or start down to the path to creating single measures that I use to measure offense and defense in one quick glance. This is probably where I’m getting myself in trouble, but I want to arrive at a healthy list of assumptions that I personally feel comfortable with when evaluating players.
Using the range of NHL Players in 2015-16, create a relative score for each of CF60, xGF60, and P60 then total them together for an offensive value score. Between points and expected goals I feel the results side is adequately covered off, while Corsi For and expected goals reflect the driving of offence and make it more predictive.
The defensive axis was a bit more complicated because I felt the need to add some context to the defensive numbers. Basically CA60 and xGA60 are the key defensive stats, but on their own they heavily favour sheltered players. Since I wanted some context, I included TOI%, and CFQoC% as well. Again using league relative numbers for all stats to create an overall defensive value.
The decision to include quality of competition might be a red flag for some, since there is very little difference between a poor CFQoC% of 48.89% and a great one of 50.6%. Why I like this stat and included it is because when you look at the top to bottom list of players the stat makes a ton of sense, even if everyone averages out pretty close.
Just to make sure I’m comfortable with what I’m doing I dropped in the standard Glass to Crosby comparison, a couple of other stats darlings like Bergeron, Kessel, and Karlsson, and because they are somewhat topical at the moment, Weber, Subban, and Luke Glendennig.
I can certainly live with those results. I think some work is still required for me to be completely comfortable with what is happening on the defensive axis, but it identifies Bergeron on as a God so it will do for a first attempt.
Yay!!! I think I have something that looks reasonable. Nothing sticks out as an incredibly gross surprise, but there are a few interesting takeaways so it serves a purpose other than to just state the obvious.
Milan Michalek is the wild card here. Looking at the numbers that led to this, it’s primarily his xGA60 in tougher usage that led to him looking like a defensive specialist. His CA60 is actually below average for NHL forwards.
Counter to that, it did capture the strength of Peter Holland and Martin Marincin in their defensive roles albeit in more sheltered situations. Josh Leivo was insanely sheltered in the little time he was with the Leafs, and all players with small samples are identified in red to help take them with a grain of salt.
Since we’re talking small samples we’ll jump across to the offensive quadrant which contains William Nylander and Connor Carrick. I think it’s hard to argue that Nylander, especially as a center will require as much sheltering that can be allowed with a young team, and Carrick is someone who can benefit from some of the same or one of the Leafs stronger partners like Marincin, Gardiner, or Rielly. Luckily as a right shot, he’s pretty much guaranteed one of them.
Greening is the other player in that quadrant and has generally strong numbers across the board. His reason for falling short is that he didn’t go up against as tough competition, which doesn’t mean that he can’t be effective in a bottom six role. So far Holland and Greening look like 2/3rds of a solid 3rd line.
The top right quadrant is pretty much where you want your core to be, and that’s where the most of it lands. Rielly and Gardiner are not surprisingly the defensemen there. Kadri, Komarov, and van Riemsdyk being the most dominant forwards (although I’ll assume JVR’s defensive score is heavily influenced by Kadri and Komarov).
Tyler Bozak’s offense not dropping without Kessel makes him a reasonably strong offensive performer, but like Michalek, his defensive numbers are too heavily skewed by usage, as he was one of the Leafs forwards at shot suppression.
Zach Hyman is the surprise guest in this quadrant, but since the sample on him is small, his minutes sheltered, and his 5v5 points unimpressive, I wouldn’t count on him living up to this, but it’s worth noting he did well in what we saw of him.
Having Polak, Hunwick, Martin, Laich, and Froese end up here makes me feel like I’ve done something right. It’s a clear patch of players that the Leafs are probably better off without.
Froese and Martin both performed well enough defensively against the lower competition they faced, and that is likely why they’ll be 2/3rds of the fourth line and likely penalty kill candidates.
Polak, and Hunwick were completely underwhelming, but given the shortage of right-handed shot defensemen in the organization, Polak’s probably got a third pairing/penalty killing gig locked up, while Hunwick probably enters the season as the 7th defender.
Frank Corrado also falls into this group and since the Leafs invested the money in Polak being the 3rd pairing right-handed shot, Corrado looks like a strong candidate to be the odd man out this year.
The less than exciting numbers attached to Kerby Rychel mainly speak to the Leafs needing to take sometime in developing him further or simply utilizing him in a different way than Columbus did. There was a reason he was unhappy in Columbus and a reason why they were willing to let him go, and these numbers reflect him being a reclamation project.
This seems to help with building out a preliminary lineup card a little. Odds are we’ll see Kadri and JVR stick together because of the success that they had, and the evidence that Kadri can cover up van Riemsdyk’s defensive short comings, but with incoming rookies, it’s likely that Komarov will be cut loose to help out another line. Bozak in a sheltered scoring role seems like a good fit, and Colin Greening looks like the perfect person to go with him. If Holland sticks at center, he looks like a great fit for the fourth line role, which could be defensively solid with probably Zach Hyman and Matt Martin on his wing.
Defensively the lesson seems to be that the Leafs strength resides the left side, with Connor Carrick being the only righty showing some flashes of potential, but of course hope still remains strong that Zaitsev is as advertised. Similar to Matt Martin’s acquisition, the Leafs went out and acquired Roman Polak, and while chart doesn’t particularly like him, he’ll probably get used because they paid for him.
While prospects will certainly be the main way that the Leafs fill roster holes, along with players like Martin and Polak (who are going to happen whether we like it or not) I figured why not take another look at some of the available unrestricted free agents to see who could help plug holes in the Leafs roster. I selected a couple few free agents, some who have been rumoured to the Leafs, others who I was kind of interested in, and others who I know very little about but was curious…
the end result is that I wound up hating everyone on this list and the fact that Luke Schenn looked like the best offensive defensemen option almost made me weep.
I still have some interest in Jiri Tlusty, who looks like he could be good, cheap bottom six filler and can help cover for the multitude of sins young players will make when they first enter the league. Tlusty played significant 5v5 time against strong competition and was a low event player and he could make a lot of sense if you’re not worried about taking a potential roster spot away from the kids.
It’s also worth noting that Kris Russell is awful, and actually worse than Roman Polak, so if we need to claim a win in free agency, it’s probably that.
While this is far from perfect and I plan on spending a good chunk of the summer refining it until it makes me happy, I feel it’s given a somewhat fair assessment of where the Leafs are today and has maybe helped me appreciate both James van Riemsdyk and Tyler Bozak a little more.
It hasn’t helped me come to terms on Matt Martin, which I was hoping it somehow would, but it has left me somewhat intrigued by what the Leafs might have in Milan Michalek.
Assuming the Leafs go with 7 or 8 experienced forwards as part of their line up, it’s seems reasonable that 4-6 rookies/younger players will fill in the gaps, and that probably means the forward group is pretty much set.
The blue line has a greater number of question marks as the left side looks set, but any of Corrado, Polak, Carrick, and Zaitsev could find themselves anywhere and it could be even more competitive if Rielly remains on the right side (please for the love of all things good in the world move him back to the left side).
The roster is beginning to look close to completion, but it’s still looking far from good. While watching Marner, Matthews, and Nylander play full seasons in the NHL seems appealing and adds a wild card element to the team, generally the Leafs still look like they have a tough year ahead.
The next step in all of this is looking at what makes each player unique. What results are they achieving that others on the team are not? What is it that makes a GM and a coach who have a history of success say that they are going to invest in Matt Martin when most of the readily available says either A: It’s a bad idea; B: It’s too much money, C: It’s too much term, or D: All of the Above.