2017-18 Season Tracked Data and Analysis

Hey all, it’s been a while. After finally getting through my (actual job’s) busy season and taking some time to settle in to a normal 40 hour work week again, I finally have some time to post something.

Like I’ve said in the past, I’m more interested in observing larger trends than covering any single game or series. Hockey is probably the least predictable and most random of the major sports and so observing or dwelling on small samples can be very misleading and leave people an emotional wreck. This team has been pretty up and down this year and that’s why it’s good to look at the bigger picture.

So far for the 2017-18 season I have compiled the data for 13 games, unfortunately all of which occurred during the first half of the season (i.e. before the break/Army series). Those games include the UMD/Union icebreaker, PSU home series, North Dakota series, Friday OSU road, WI home series, Notre Dame road Saturday (the Friday game isn’t available to stream), Michigan road series, and Michigan State Friday home game. It’s a fairly even split of home/road games in which they went 5-7-1. I might track a few more games just to get at least 2 for every B1G conference team (minus Michigan State because I don’t want to fall asleep from boredom) as well.

While the team has undergone some changes during the second half, most notably the emergence of Mat Robson as the starting goalie, one long term trend remains the same. Other than a very good offensive showing against the Badgers in Madison a couple weeks back as well as 7 goals in the series vs a bad Michigan State team, the Gophers’ offense has mostly struggled all year and not much on that front has changed since I wrote about it in a previous post. The Gophers goal totals in the other series were 3 vs SCSU (plus an empty netter to make it 4 on the weekend), 4 against Michigan, 2 against Notre Dame, 3 vs OSU, and 3 this past weekend. Amazingly they managed to split (as well as a win and a tie vs OSU) all but the Michigan series with that offensive output. More recently against Penn State they scored a whopping 3 goals in the first series and then had an offensive outburst today to get to 8 goals on the weekend. While last night’s game an anomaly, the team did show some improvement offensively.

It is worth mentioning that there have been some notable improvements in terms of puck possession on the year and the team is no longer a horrible possession team getting badly out shot every night like they were back during the earlier parts of the season. They have since moved up to being mediocre in that area and are now 28th out of 59 NCAA teams in even strength shot differential. While this isn’t exceptional by any means it’s definitely an improvement from what it was earlier and it is worth noting that this team has played the most difficult schedule of any NCAA team this year. This improvement as well as the emergence of Robson, who has posted a .943 sv% on the season, is probably the biggest factor in this team’s success during the 2nd half of the season and the recent 6-1-1 run they had before the PSU series. Based on Don Lucia’s comments this team seemed to have embraced a defensive style of play (so much for that these last two weekends) and their reward is a high percentage of making the NCAA tournament, even after the recent debacles against Penn State.

For the most part the data I track is probably best used for shedding light on the neutral zone play as well as the offense and that’s what this post will focus on. So without further ado let’s get into the data for what’s been tracked this year.

Here are the zone entries for this year as well as a breakdown of shot attempts and scoring chances generated off those entries.

 

2017-18_entries

A few things jump out from this data. Unsurprisingly Casey Mittlestadt leads this category by a wide margin and with his speed, playmaking, and vision, it’s not all that shocking that he generates a lot of clean entries as well as shots and scoring chances off of them. We can also see that there’s some correlation between getting clean zone entries and shots per entry. The guys in the top tier certainly average more shot attempts and chances per entry than the bottom tier, though it obviously doesn’t explain everything as Romanko and Szmatula aren’t getting many dangerous scoring chances from their entries (more on that later). On average though this team is still far below what they were last year in this area averaging a controlled entry on 41.1% of all entries.

The last column also highlights some big differences from last season for a few players. Pitlick, Sheehy, and Bristedt are all significantly worse in terms of generating controlled entries than they were a year ago. While there are definitely some variables and possibly some variance due to the small sample size of data (last year’s sample only contained 8 games), it isn’t all that shocking to me that 2 of these 3 players are producing less offense than they were a year ago and Pitlick hasn’t taken a step forward yet. These guys are speedy players who did well to generate rush chances last year and this year that’s gone by the wayside. It’s hard to tell if this is a symptom of anything Lucia is teaching the players but it does seem like the rush chances both for and against have been less common this year and based on some of his comments this seems likely.

The other thing that’s obvious is that none of the defensemen generate many clean zone entries but Zuhlsdorf is leading the pack there. The thought when he was recruited was that he was a good puck rusher who could generate these kinds of plays for us and contribute offensively. It’s good to see his improvement in that area and hopefully he can continue to grow there and become an offensive presence. But as noted plenty of times by various people who cover the Gophers, this team clearly doesn’t have a guy like Mike Reilly who can really rush the puck or generate consistent offense.

For zone exits we have the following:

2017-18_exits

On the team level this isn’t all that far off from last year, for last year’s team 52.5% of zone exits were controlled, this year it’s 50.6%. Sheehy once again sticks out like a sore thumb here for his difference compared to last year. At the very least earlier on in the year it was really obvious that Sheehy was making significantly fewer controlled plays with the puck at both ends of the rink. It wasn’t all that shocking to see it in the data once it was finally compiled. Lindgren has gotten slightly worse this year though his role as the top defenseman on the team has increased from last year and he is essentially replacing Bischoff as the leader of the top D pairing. It’s also worth noting that he was injured before the season, recovering much of the offseason, and one can assume that it put him at a bit of a disadvantage early on. Sadek and Johnson improved slightly in teams of controlled break outs and although it’s a tiny sample Rossini does look good in this area. From what I’ve seen and the data shows, Sadek is probably the best puck mover on the team and while Nanne has his share of turnovers and defensive mistakes, he is pretty solid at moving the puck up ice and transitioning.

Finally we have the passing data from this year. I haven’t talked much in the past about passing data but essentially I started tracking this due to the work done by Ryan Stimson (who initially started the passing project research), Corey Sznajder (the inspiration for much of the tracking I do), and Matt Cane who assisted Ryan with the research in the upcoming link. Passing data is important because at a higher level it examines how you set up your offense and examines your offensive chemistry. This research by Ryan Stimson and Matt Cane shows that shots preceded by a single pass are more likely to end up scoring and shots preceded by multiple passes are even more likely to find twine.sloan_sports_passing

As you can also see they show that passing is a repeatable skill and not just something that has a ton of random year to year variance. While it’s pretty intuitive to anyone who watches hockey that passing is a skill, it’s good to know that someone has done the work to prove this scientifically.

Ryan and others have also shown that shots that originate from a pass from behind the net (defined as behind the goal line), or a pass across the center of the slot (i.e. the royal road) are more likely to score. Tracking these plays can tell us more about the playmaking ability of the various players on a team and tell us more about who is creating the dangerous chances.

Royal-Road

As such I wanted to do the same thing for the Gophers. See below for a breakdown of the 5 on 5 passing data. The data I have for passing is broken down by strength (5v5, 5v4, 4v4, etc.) and for this piece we’ll just focus on 5 on 5 because it explains the majority of most hockey games while also comparing apples to apples. Not every player gets time on the power play and special teams are essentially their own separate entity.

passing_data_17-18

First, the passes column in beige basically tracks the total number of passes (either the primary, secondary, or the third pass of a passing sequence that lead to a shot). Unsurprisingly guys like Mittelstadt and Novak are leading the team in this category, on defense Sadek and Lindgren clearly are the top pairing on this team and get the most ice time so it’s not surprising that they’re up there in this category. Interestingly enough, Nanne is second among defensemen and you can kinda see why the coaches would prefer to play him on the back end despite some of his turnovers and defensive gaffes, he’s clearly one of the best puck moving defensemen this team has. This combined with the zone exit data above give you a good sense of which defensemen are good at moving the puck from the back end as well as facilitating offense.

While Casey Mittlestadt and Rem Pitlick are clearly both first liners who likely get a lot of ice time, Scott Reedy is very high in terms of the number of scoring chances and 2nd on the team overall (third grey column). That column is calculated as anyone who took a shot attempt in the area generally defined as the scoring chance area (i.e. the home plate area of the offensive zone). Reedy has bounced around the lineup from Mittlestadt’s line to even the fourth line at times and yet he’s finding a way to create a lot of scoring chances in the minutes he gets. I think during most of the games I tracked he was playing a lot of minutes on a line with Mittlestadt and Pitlick but you can also see from the number of passes and shot assists that he isn’t merely a passenger on that line.

Another thing that jumps out is the combination of Leon Bristedt and Mike Szmatula who have mostly played on the same line this year. For Bristedt, who has 75 shots on the year in all situations, it’s interesting to see him next to Romanko (a fourth liner not particularly known for his playmaking or scoring ability) in terms of shots and primary assists at 5 on 5. Leon seems like a guy who could do alright on the wing of a more talented playmaker like Kloos or Mittlestadt (or last year’s version of Vinni Lettieri whose line he played on much of the previous season) but he’s kind of one dimensional and dependent on the talent around him for his production. Leon isn’t an elite sniper but when he puts up a ton of shot volume he can still be a productive player as shown in year’s past. This year his shot volume is significantly down, by over 25 shots in fact, from last year and down from the year before that as well.  This year being on a line with Szmatula for most of the year it’s not all that surprising that his production is down from a year ago (see Szmatula near the bottom of the chart above). These guys are both very low on the chart in terms of passes that lead to a shot, primary shot assists, and dangerous passing sequences (Szmatula is better in the last area, Leon not so much). Obviously these stats aren’t adjusted for ice time and I have no way of knowing how much any player gets per game but it’s not all that shocking that neither of these guys have been all that productive this year. Neither seems capable of carrying a line and when put together it’s just not very pretty.

Another column to look at is the beige column that shows the shots + primary shot assists (i.e. shots + 1A column). Basically this column contains any shot attempt either taken by the player himself or instances where that player passed the puck to someone who took a shot. This column also subtracts the number of solo shots (i.e. the number of shots taken in which a player just took a shot and never passed the puck or received a pass) because the percent chance of scoring on a solo shot is generally very low as shown from the screenshot from the Sloan Analytics conference above. This is another area where the usual suspects of Mittlestadt, Novak, and the rest of the forwards who regularly get top 6 minutes (Sheehy, Pitlick, and Gates) are higher up on the list. Reedy once again seems to shine here as well. Seeing Gates as high as he is on the list is interesting and I do feel like his playmaking ability did improve this year and he’s one of the few guys that I would argue improved offensively since last season. It seemed from what I watched that he was better at making both simple and dangerous passes to either get to the offensive zone or help finish once they got there.

gates_pass

 

Seeing Sadek, Zuhlsdorf, and Johnson up there isn’t all that shocking either as they’re the most likely guys to be actively involved in the offensive zone.

So putting all of this together what does this tell us? We already know that this team has struggled to score goals this year and at 41st in the nation in scoring (even after this past weekend), this is clearly a huge drop off from last year in which the team was 5th. While it can’t be ignored how much better the power play was last year (I plan to look into this year’s PP data more at some point), Lucia has lamented the lack of 5 on 5 scoring all year as well. Looking at how much worse this team has been in terms generating clean zone entries compared to a year ago, it’s not all that shocking that the team is creating less offense at even strength. The Gophers historically have generated a lot of chances off odd man rush situations but that has mostly dried up this year. The other factor here to consider is the forecheck, while I don’t have any great way to measure the forecheck, based on what I’ve seen this team just hasn’t been all that effective at it and often times goes back into a passive trap trying to create counter chances through the neutral zone. Having seen this team have a really effective aggressive forecheck back during the more successful years from 2012-2014 (I went back and watched some of the highlights from the regional vs North Dakota) and as someone who watches the Pittsburgh Penguins regularly, I feel like I know what a fast effective forecheck looks like. I don’t think its been a strength of this team and unless the team is down a goal it doesn’t seem like it was all that aggressive, at least earlier in the season.

If this Gophers team had a rock solid blue line I think sitting back and playing more of a counter attacking trap style would be effective. If we’re being honest with ourselves though I would say that this blue line is inconsistent at best in terms of its ability to play shut down defense and we saw that the past 2 weekends against a team that knows how to score goals (3rd in the nation). I know some of these defensemen were known for their puck carry ability coming in, specifically Zuhlsdorf and yet none of them are doing that as you can see from the zone entry data. Reading between the lines I don’t think the coaches want these guys doing that very often, and this was extremely obvious after the team lost the first game to Duluth after a denied carry-in attempt by Zuhlsdorf wound up in the back of their own net in OT. While I get that coaches want to minimize mistakes that result in a goal against, you sometimes have to wonder if this staff is trying to hide its flaws rather than playing to its strengths. Making controlled plays can sometimes carry more risk but I would argue the potential reward significantly outweighs the risk. Johnson, Sadek, Zuhlsdorf, and Lindgren have all shown that they can be competent puck movers and yet the coaching staff seems like it’d rather just have them throw it away and try to create turnovers with their ineffective forecheck. Some of their comments after last weekend referred to the players getting “too cute” with the puck. While I can’t say I know exactly what they meant by that, I feel like they’d rather see their guys make the safe zone exit or entry even if that means just throwing the puck away. As I’ve heard Ryan Wilson say, writer/podcaster for Hockey Buzz who covers the Penguins, “you work so hard to get the puck from the other team, why would you just throw it away and potentially give it back to them?”. That’s not to say teams should never dump the puck but the Gophers have done it somewhat excessively in the games I’ve tracked and significantly more than they did in the games I tracked last year.

The other thing that you can see from the passing data is that this team seems pretty top heavy and especially so when the lines are assembled like they have been at times this year. Putting all your eggs in one basket with Mittlestadt/Pitlick/Reedy on the same line makes for one really good line but they can only play 20 minutes a game at most. Spreading that out and having at least one guy who can carry each line has been shown to be more effective. The problem is that not many of these guys can really carry a line outside of those 3. Novak is a guy who can do it but after him you don’t have much in terms of centers with playmaking ability and realistically Szmatula needs a solid winger to help carry his line. Don tried this briefly switching Bristedt to Mittlestadt’s wing and Pitlick on Szmatula’s but it didn’t last long. Along these same lines you kinda wonder what the deal is with Luke Notermann, a guy who showed some promise last year in the third line minutes he played last season and the data I tracked. He seemed like a speedy guy with decent vision who could make controlled plays and seemed like an ok passer (I didn’t track passing data last year but he was decent at zone entries at least). At this point it’s too late to just throw him in the lineup and see what happens but would it really have hurt to get him in the lineup more? We kind of know what we have in Ramsey and Norman at this point and there’s no rule that says you have to play traditional 4th liners.

While this team will still likely make the NCAA tournament despite 4 straight losses to the same team, it’s hard to see this season as any kind of improvement over what we saw a year ago both in terms of success and entertainment value. Luckily the NCAA tournament is formatted in such a way that favors underdogs, the amount of variance in a single hockey game is very high and last I checked higher than any of the major North American sports. If there’s any hope for the Gophers to salvage this season by doing well in the NCAA tournament, it’s due to the small sample nature of it. It’ll be interesting to see how this team plays as the season winds down and what adjustments the coaches will make moving forward. While losing 4 in a row to Penn State is obviously disappointing, all is not lost yet (the Gophers chances of making the NCAA tournament are still above 90%) and the team has shown some signs of improvement offensively at least in terms of the eye test and scoring. If they play Robson and the offense/power play plays anything like it did last night (Mittlestadt was phenomenal) then this team could potentially make some noise in the NCAA’s.

But even if they do it’s hard not to look at the bigger picture of this season, especially as compared to a year ago and especially when compared to where this program was when I was just entering college (the 2006 squad that had a 19 game unbeaten streak and won the final five on Wheeler’s iconic goal), and wonder what exactly the plan is going forward with this coaching staff and the various other issues the program faces. Don Lucia has one more year left on his contract after this season, there isn’t much excitement around the program (season tickets and actual attendance are significantly down), and it’s becoming increasingly clear that something needs to change. I hope for the best in the NCAA tournament and look forward to watching this team (assuming they get there) in a few weeks but it’s easy to see why people are frustrated.

Thanks for reading!

 

Quick notes: Big shout out to Corey Sznajder and Ryan Stimson for responding to my emails which allowed me to track this data and compile it. Without them this isn’t possible.

Most of this post was written before the Penn State series this weekend and some of it even before last weekend’s series against the same team. I edited it a bit to reflect what happened this past weekend. It’s amazing how quickly some of the narratives can change, especially after your team concedes 21 goals in 4 games and to that point had previously let in that many in 10 games.

 

Advertisements

The Michigan Series

So I promised my next post would delve more into the overall stats of the team and while I don’t do it often, I’m going to break that promise. Unfortunately with work being as busy as it is and the fact that we’re pretty consistently being asked to do 18 hours of OT up until the new year (and possibly into January), I’m pretty far behind in terms of tracking games. So far I have the Icebreaker challenge, PSU series, UND series, Michigan Series, and the Friday night Mich St. games tracked. My hope is that over the Gophers’ break I can swing back and do the Clarkson series and Notre Dame as well.

Back on topic, I decided to write about the Michigan series for obvious reasons, this was a series that probably should’ve ended in a sweep given that they had a 3-0 lead midway through the 2nd period of the Friday night game and a 6-3 lead partway through the 3rd period of the Saturday game, which had previously been a 4-0 lead earlier. The topic of blown leads has been talked about with this team in the past, last year the team blew a 5-2 lead and ended up losing to SCSU in OT in their first home game of the season. Last year vs Notre Dame was another where they blew a smaller 2-0 lead in a more important game to end their season. There was a similar trend with these games, the team made less controlled plays through the neutral zone both in terms of entering the opponent’s zone as well as exiting their own zone.

All stats here are at 5 on 5 and show the Gophers controlled entry and exit percentages by period in the game vs SCSU. For a brief recap, what happened in this game was the Gophers scored 1 in the first and had a very productive 2nd period where they outscored SCSU 3-2. Their 5 on 5 play wasn’t particularly amazing, especially in terms of exiting the zone, but they managed to get 2 goals off power plays. They then got a 5-2 lead early in the 3rd off of Novak’s goal. After this point they managed to get controlled entries on 30% of their entries (6 out of 20) including in OT.

The Notre Dame game was similar:

Team_stats

In this game the trend isn’t quite as obvious but before Notre Dame made it 2-1 late in the 2nd, the Gophers were entering the zone cleanly about 57% of the time. That went down to about 45% of the time after ND scored goal #1, a trend that got worse in the third period. This one wasn’t so much about failing to protect a lead, it was that they became more conservative towards the end of a game they had mostly been dominating territorially during the first half. The clean entry percentage went way down and the percentage of clean breakouts was down as well. Now this was partially due to Notre Dame all out defending towards the end of that game and going into their classic shell much like they did last Friday. For them being conservative actually worked well in this one because their goalie was on fire made a few huge saves on odd man opportunities and other quality scoring chances. But at the same time you can see evidence of less controlled plays by the Gophers and this became a huge problem when their powerplay didn’t produce any results in the third.

After seeing a 3 goal lead in the first home game and a smaller lead get blown in the final game of last season, it was a bit disheartening to see 3 and 4 goal leads blown in one weekend on the road vs Michigan this season. Looking at the data gives us some idea of what the team was doing and what caused this to happen

Friday_NZ_Michigan

On Friday there isn’t a super clear trend to how the Gophers chose to defend the lead, if anything this was a game where they played somewhat sloppy at 5 on 5 all game and it came back to bite them in the ass later on. The one thing that’s apparent though is that the Gophers got worse in the third period in terms of exiting their zone cleanly. The zone exit number you see there is the percentage of the time in which they cleanly exited the zone. The entry data is the percentage of the time they attempted a controlled entry, the vast majority of the time which they completed (I can post these numbers as well if people are interested). Michigan was kinda ugly at exiting the zone all game but the Gophers had some costly turnovers in the neutral zone. They also got power plays late in the tied game as well that they managed to score on, including in OT. But one part of their gameplan where they were overall better was zone entries, the entry numbers you see there is the percentage of the time in which each team attempted a controlled zone entry either via carrying it or passing. You can see that Michigan more often than not throughout the game was attempting to carry the puck in and it paid off, they outchanced the Gophers at 5 on 5 and had a 9-6 advantage in that area.*

Saturday the Gophers had a 4 goal lead as well as a 3 goal lead with about 16 to minutes left in the game.

NZ_Data_Mich_Saturday2017

The trend in Saturday’s game is more obvious, the Gophers got really conservative in the third period and the entry percentage in the third doesn’t quite do justice to how much they were dumping the puck. After Pitlick scored to make the score 6-3, the Gophers had 5 controlled entries on 22 attempts the rest of the period for a 22.7% rate. For those of you who are fans of the Minnesota Wild, if you put an entire team of Ryan White on the ice for a period you would probably achieve this result. For a Gopher frame of reference, Jack Ramsey gets a controlled entry 28% of the time in the games I’ve tracked this year.

Now as I’ve stated in a previous post, controlled entries are far more efficient and you’re far more likely to get a shot or a scoring chance off a controlled entry than if you dump the puck and I cited previous studies that have shown this to be true. They can be more dangerous, a turnover at the blue line can cause a break the other way in which the other team potentially gets an odd man rush or a clean entry themselves that generates either a shot or a scoring opportunity. But with risk comes reward and for a team that is smaller and faster, it probably makes sense to try and generate controlled entries whenever possible. Completely changing your game plan to dump the puck twice as often is probably a mistake and the shot data bares this out. At 5 on 5 Michigan ended up with 49 shot attempts to the Gophers’ 37. After Pitlick scored in the third, Michigan had a shot attempt advantage of 15-5 and the shots on net were 8-2. Oddly enough during that time the scoring chances were an even 2-2 (all of this data pertains to even strength play) and Michigan got 3 goals to the Gophers’ 0 to tie the game and bring it to OT.

Now I’m not going to pretend to have any insider knowledge of the team or know what the coaches are telling the players. It’s entirely possible that the coaches were telling the players to carry the puck in more often late on Saturday’s game, or on the weekend as a whole, and they’re choosing not to do this. But in interviews Don Lucia consistently stresses the importance of “getting pucks deep” and reading between the lines, I have to assume he means dumping the puck when he says that. And obviously that’s what the team ended up doing late in Saturday’s game as well as more often than Michigan overall on Friday. We all know how it turned out and the Gophers currently sit 7 points behind Notre Dame in the Big Ten Standings. Now I will say that Michigan does play on an NHL sized rink and the Gophers are more used to playing at home on an Olympic sized ice sheet that is 15 feet wider and more conducive to clean zone entries. If I had to guess, this plays into the coaches’ gameplan for the weekend. The problem is it’s just not a very efficient one for the reasons I outlined and while the results of the weekend aren’t solely due to dumping the puck, in my opinion it definitely played a big part. Whether this is the coaches’ doing or entirely on the players, hopefully we see less of it as the season goes on and we improve on making controlled plays. In the games I’ve tracked this year the Gophers are getting controlled entries far less often than last year, 39% of the time to be exact. For a frame of reference this would be lower than any team in the NHL and much lower than the last year’s rate of 48% in the games I tracked (mostly home games but the games vs Duluth and Notre Dame on smaller surfaces had similar numbers, i.e. 47.7%). In the NHL that would be more towards the middle of the league.

To be fair, this was one weekend out of a 32 game season and if you include the games I cited from last year, we’re talking 4 games out of a 54 game sample. These games were picked for the very specific reason that the Gophers blew big leads in 3 of the 4 games and one was picked because it ended our season last April. With such a small cherry-picked sample I’m not going to pretend that this should necessarily be considered a trend. But at the very least, in these games, it’s a bit disturbing to see what appears to be an important factor in these blown leads as well as the fact that the third period was their worst in all of them mostly because they decided to throw the puck around at both ends of the rink more often than not.

On a more positive note, the Gophers have been playing better at 5 on 5 in two of their last three games, even if they had some trouble scoring. It’s hard to score 6 goals on 27 shots like in the Michigan series or 4 goals on 26 shots like they did against Harvard on Friday, eventually you’re gonna run into a good goalie like the various Notre Dame goalies we’ve faced over the last year or so. And if the Gophers play like they did this past Friday vs Notre Dame, they will most certainly win far more games than they lose and be a contender. This team has a ton of skill and all it takes is them playing well at the right time of year to win it all.

Thanks for reading!

 

*One note with the Friday Data, I was relying on the University of Michigan’s stream to compile this data and there were multiple issues. The camera at one point just focused on the Michigan crease for about 5 minutes of the game sometime in the third because someone likely forgot to change the camera. Also the stream’s frame rate slowed down drastically sometime after Szmatula scored the first goal of the game. So as such the Friday data doesn’t totally align with the actual shot totals and Tommy Novak’s goal got missed entirely in the data for instance. As such it’s possible I missed some shots/scoring chances for each team in this one.

Early Season Trends

Hello Gopher fans! I haven’t forgotten or abandoned this project and have been hard at work tracking and compiling stats for the 2017-18 season. So far I have the first six games tracked and plan to get to the Michigan State series eventually. I skipped Clarkson so that I had some time to compile the data already tracked and might get to it during the Gophers’ winter break but we’ll see.

While the team has lost some key players since last year in Justin Kloos and Vinni Lettieri, we all know this is a team that doesn’t tend to rebuild and instead reloads. Before the year there was a ton of excitement around the fact that one of the newest recruits, Casey Mittlestadt, a top 8 pick in the NHL draft, is now with the team. With the acquisition of him, Scott Reedy, and Brannon McMannus to name a few, the preseason outlook of the forward group was very favorable. The defense looks to be fairly similar to last year’s with the losses of two upper classmen, Jake Bischoff and Ryan Collins, whose icetime has mostly been replaced by Tyler Nanne and Jack Glover. And with new goalie Nate Robson now on the roster (who is eligible to play later in the season), the goalie outlook looked strong as well between him and Eric Schierhorn. As such the preseason expectations were very high with this team.

So far this year the brightest spot has to be goaltending and the improvement of Eric Schierhorn. It’s still early but as of today he is second only to Jake Keilly of Clarkson in save percentage for goalies who have played 10 games or more. If you include goalies that have played 8+ games he is currently 5th and within .001 of of the top 5. Given UMN’s difficult schedule to start the year and how the team has consistently been outshot this is really impressive and a good sign going forward. Whether this is repeatable is a fair question to ask given that his .933 sv% is a big jump from his save percentage over his .906 sv% over the 74 games that preceded this year. But a lot has been made of his preparation before the season and the fact that he now has better competition in Nate Robson for the spot. It will be interesting to see how this progresses going forward but as of right now goaltending is probably the team’s biggest strength thus far and by far the biggest reason for their current record of 7-3.

As for what we’re seeing from the rest of the team, admittedly it’s hard to be down on them given their current record and the fact that their schedule has been pretty tough including 3 difficult road games vs North Dakota and Duluth. But there are some trends that are a little bit concerning both from a sustainability and a pure entertainment standpoint. Let’s dig into the numbers a bit.

One of the better predictors of success at the NHL level is even strength shot attempts for and against. A lot this has to do with the fact that the majority of all hockey games is played at 5 on 5 and shot rates tend to correlate with scoring chances.  In the link you can see that while expected goals is slightly better, even strength shot attempt differential tends to be pretty close and is much better than goal differential. At the collegiate level we don’t have expected goal metrics available to us and another limitation is that we also don’t have score adjusted shot metrics to take out some of the noise of score effects. That said it’s still more ideal to be on the right end of that spectrum and consistently out-shooting your opponents more often than not.

Before the Michigan State series the Gophers sat in the bottom 10 of all NCAA teams in even strength shot differential, next to the likes of Holy Cross, Brown, and Alaska. As of today they’re 47th out of 60 teams. To get a sense of which teams are in the top 10, we can see that some of the best teams in the polls that have higher expectations such as North Dakota, Denver, Duluth, Mankato, Penn State, and Harvard, all sit in the top 10. Even a team like St. Cloud who has been been a bit more pedestrian in this area is right at about 50% and in the middle of the pack of all NCAA teams. For the Gophers it’s still very early in the season and this team has had some players recovering from injuries that happened last year as well as this year but this will be something to watch going forward as it doesn’t tend to bode well for future success. A lot has been made of the special teams difficulties this team had especially in the series vs North Dakota but to me the 5 on 5 play is a bigger concern simply because the vast majority of all games is played at even strength and this team doesn’t lack shooting talent, the power play was bound to bounce back at some point. For frame of reference, last year’s Gopher team ended up 19th in the NCAA in this metric and Denver, the team that ended up winning it all, was 6th (and due to the difficulty of their schedule vs some of the WCHA/Atlantic teams ahead of them, 6th is probably under selling how good they were in this area). The year before North Dakota was 4th in this metric and the Gophers, who eventually missed the tournament, were 19th (keep in mind the B1G was terrible that year and only one team made the NCAA tournament, their shot differential was undoubtedly worse vs non-conference teams). As such it’s definitely good to see this team winning as many games as it has but it’s also fair to say that this team has been outplayed more often than I expected, especially against the NCHC opponents.

While it’s obviously somewhat concerning that this team is consistently being outshot at even strength, I would argue that what’s driving it is even more alarming. Before Saturday’s game vs the Spartans the team was 5th worst in the entire NCAA in terms of EV shots/game ahead of Brown, LSSU, Sacred Heart, and Anchorage in that order*. Even if you just want to look at goals per game this team currently ranks 39th. For a team that has mostly reloaded in terms of its roster and has kept much of its key talent up front and acquired Casey Mittlestadt, this isn’t exactly what I would’ve expected going into the year. So while the team is currently getting the tangible results it needs to get an NCAA birth and that’s a good thing, it is interesting to see this team struggling on offense and playing somewhat conservatively.

Another way to gain some insight into the strategy the coaches are trying to implement is by looking at team pace, ie when you add up the shots for and against per game, is this that total on the higher or lower end of the spectrum? Generally speaking teams with higher numbers here tend to be playing more of a run and gun style, in the NHL the last couple years the Penguins and the Maple Leafs were on the higher end and teams like New Jersey and Vancouver (this year anyhow) at the bottom playing more of a trapping/defensive style. There are some exceptions but generally it gives us some insight into what teams are trying to do strategically.  As of right now the Gophers are second to last, only above Anchorage. The other teams in that realm are North Dakota and Bemidji. North Dakota is by far the best shot suppression team in the entire country and towards the middle of the pack in terms of shot generation which is probably the biggest explanation of why their pace is so low. In reality they aren’t necessarily trying to play a slow paced game, they just have a really good D core. Bemidji is a team that has been known to slow the game down in recent years and Anchorage is… well Anchorage. If I had to guess based on what I’m seeing as well as what the numbers show, this years Gopher team is playing more of a conservative defensive based system.

Now on a more positive note Dakota and Duluth are top 10 teams in terms of suppressing shots and the team won’t be playing either of them again during the regular season. As such it’s hard to imagine the team not improving in a lot of these areas. In the past 2 games vs Michigan State (which haven’t been tracked yet) the team has looked better and outshot them while looking significantly better at even strength. And I can’t deny that the team is getting the results it needs to gain some ground in the pairwise despite playing below their capabilities.

While this post probably seems a bit negative for a team that’s played a tough schedule, gotten good results, and dealt with a spate of injuries (possibly including some lingering ones), I think it is fair to try and think about the results and the overall process as separate entities. While the team has gotten outplayed quite a bit in the early part of the year, I also think that it’s really impressive when a team can manage to find ways to win games where they get thoroughly outplayed. This team is talented enough where it can pull more of those types of wins than a lesser skilled team and hopefully they can build off an MSU series where they looked better. In the next post I will post some of the team level and individual stats I’ve been tracking. Thanks for reading!

 

*Since not all teams play the exact same amount of time at even strength I tried to come up with a way to calculate this. What I did was took the total shots on goal at all strengths, divided that by the number of shots at even strength to create a coefficient for every NCAA team, and multiplied it by their EV shot rates across the board to try and figure out what every team’s even strength shots for and against per game would look like if the entirety of all their games was played at 5 on 5. I’ve also done this with using a coefficient derived from taking penalty minutes out of the equation and got similar results. This method isn’t perfect and ideally we’d have 5 on 5 ice time to work with but I’m doing the best with the limited data that exists for NCAA hockey. All data came from CHN’s website.

UMN vs Notre Dame Breakdown

After an exciting year of Gopher hockey the season ended on a somewhat sour note after the Gophers lost 3-2 to Notre Dame in the first round of the NCAA tournament. The Gophers were good enough to earn a #4 overall seed and a #1 seed in their regional but didn’t have enough in the tank to make it to the frozen four. Was this a matter of poor play, a poor process, coaching, injuries or bad luck? Maybe a combination of the above?

For the most part you all watched the game and remember how it unfolded so I won’t spend too much time recapping what happened. The Gophers were absolutely buzzing creating a ton of chances for the better part of the first 2 periods but just couldn’t get the additional goals needed to really pull away after taking a 2-0 lead. Eventually the Irish tied it late in the 2nd period and got the lead midway through the 3rd which would end up being the game winning goal. The goal here is to take a deeper dive into the numbers at both a team and individual level and see if our eye test mostly matches what the data shows. First we’ll start with the team metrics.

Team_stats

For those who don’t know CF% stands for corsi for percentage, it’s basically a measure of who generated what share of all shot attempts. FF% stands for Fenwick For percentage and it excludes blocks shots. By either metric the Gophers couldn’t have done much better and carrying these types of percentages on the year would likely make you one of the more dominant teams in the NCAA. For frame of reference, no NHL team has achieved a 60 CF% over the last decade (which is as far as the stat goes back on Corsica). The best a team has ever done in that span was the 2007-08 Red Wings who ended up winning the cup that year and carried a CF% of 58.7%. In terms of out-chancing ND at 5 on 5, the Gophers definitely did that over the entirety of this game.

Part of the problem though was the third period. They did get some good chances in the third such as the oppurtunity where Kloos was all alone in front of Peterson with a couple minutes left but just couldn’t find enough time or space to bury it. But for the most part you can see that the percentage of controlled entries was significantly down from the earlier periods for the Gophers. This also happened to correlate with them dumping the puck in more often during the 3rd, they had 10 dump-ins vs 6 in the first period and 7 in the 2nd. Only 3 of these dump ins for all 3 periods were actually recovered by the Gophers (there were some instances where ND turned the puck over when attempting to exit the zone). This isn’t super shocking because the Gophers aren’t exactly a big team who are going to have a ton of success grinding it out along the boards against other teams. ND on the other hand improved in terms of entries as the game went on.

Looking at what happened at an individual level also confirms some of the stuff that was discussed on GPL.

ND_Entries

MN_Entries

Above are the zone entries and controlled entry rates for each player on both teams. I’m sure everyone noticed Anders Bjork had a hell of a game and for my money Justin Kloos was just as good even if he and his linemates didn’t produce any points at 5 on 5. Kloos’s numbers above don’t do a good enough job of reflecting how impressive his contributions in this area were because he does such a great job of opening up space with his speed and dishing the puck to his linemates in open space allowing them to carry the puck into the zone cleanly.

Letierri did a solid job driving his line and generating chances and it might’ve been interesting to see what he could do if a better finisher like Gates was on his line instead of Cammaratta. That isn’t a knock on Cammy, he held his own in terms of zone entries and controlled play in the offensive end but he isn’t known for his shooting ability. Lucia tried to maintain a balanced approach with his lines and it’s hard to fault him too much seeing how well it worked on the season as a whole.

Another thing that’s glaringly obvious from the Gophers’ entry data, this team really lacked a dynamic defenseman who excels at skating the puck and entering the zone cleanly. For the most part it’s not expected that all defensemen are going to be good at this but having a guy like Mike Reilly or Nate Schmidt makes a huge difference in what a line is capable of in all 3 zones. Even without a guy like him though you’d think they could figure out a way to do better than 0 controlled entries over an entire game.

As for Notre Dame, who admittedly I’m not as familiar with, it became very clear early on that Bjork came to play and is an impressive player. That kinda goes without saying. Looking at the data I found it interesting to see that their third line (at least as it was listed on their line chart) generated more consistent offense than their second line during this game. They not only got more controlled entries they also generated more shot attempts.

Onto the zone exit data:

ND_Exits

MN_Exits

While Kloos’s zone entry data probably wasn’t as impressive as he actually was in that area of the game, this should give a clearer picture as to how important he is to the team. The entire game he was absolutely flying out of his own zone and through the neutral zone backing up the opposition whenever he had the puck. He easily evaded their forecheck with his speed and as such it created quite a few odd man rush opportunities in the other direction. His linemates did a solid job as well and to anyone who watched the game this probably isn’t all that shocking but this helps give a clearer picture of how good that line was. Even though the result didn’t end in their favor they can all definitely hold their heads high after this effort.

Some other guys who had a solid game in terms of exits were Sadek and Glover (the latter of whom’s weren’t quite as difficult judging by the pressure). These two will be very key next year now that we’ve learned that Collins is leaving to go pro, not to mention the loss of Bischoff. Glover had a really nice stretch pass late in the game that sprung Letierri behind the defense for a 2 on 1 but they just couldn’t get the tying goal. Zuhlsdorf did a lot of the heavy lifting on his line and fared pretty well at 50% controlled. He’s another player who will be interesting to watch next year and see if he can take the next step in becoming more dynamic and fulfilling that particular role. And Bischoff had a solid game generating a good percentage of clean breakouts despite being pressured on most of them. One guy who sticks out like a sore thumb in the stats above is Cammaratta. I went back and watched his exits out of curiousity and while a few of them were pressure situations where he didn’t have a decent outlet, on at least 3 of them he had options and just flubbed the pass. Not his best game.

Notre Dame wasn’t quite as good at exiting the zone cleanly but once again, Bjork really excelled in this area (what else is new) as he was clearly the driver of that line. The 2nd line once again wasn’t quite as impressive as the 3rd. I find it interesting to see that the Gophers put a ton of pressure on their top D pairing but not quite as much on the 2nd pair who seemed to fare better in terms of exiting the zone cleanly. I’d be curious to know how Jackson was deploying the D pairings in terms of zone starts and which forward groups he paired them with because that could potentially have made a difference in terms of these stats. If they were able to get the puck to Bjork on a breakout that would definitely make their exit stats look much better.

So despite the fact that the Gophers dominated play for the majority of the game, created more chances, and were a bit cleaner in terms of what they did with the puck, why did they end up losing this game 3-2? After all, Notre Dame is playing in the Frozen Four as I write this. One key difference was that Notre Dame has one of the best goalies in the NCAA who played 39 games this year sporting a save percent of .929. Only one goalie played as many games and had a better save percentage and that was Bemidji’s goalie, who probably doesn’t face the quality of competition that Cal Peterson did in the Hockey East conference. For frame of reference Adam Wilcox had a save percentage of .932 the year they went to the national championship. Goalies are a great way to even things up when your team either isn’t playing well or just isn’t quite as deep or skilled as another team on a given day. Hockey is a very low scoring game for the most part these days and games are won on the thinnest of margins. Having a 60-40 shot advantage should win you most games but when the difference in save percent is .939 to .895 the Gophers would’ve theoretically needed another 15 shots on net and not conceded anymore than they did. Even against Mercyhurst that would be asking a lot. That’s not to say Eric Schierhorn played badly for the Gophers, he made a bunch of good saves as well, but he just wasn’t as spectacular as Peterson was and has been a bit more average throughout his career at the U of M. People hate the “hot goalie excuse” but this wasn’t a case of the Gophers making an average goalie look like Hasek, Peterson is one of the elite NCAA goalies and he played outstanding in this one. Guys like Kloos, Sheehy, and Pitlick couldn’t bury one more of IDK how many great opportunities. In some ways it was just kind of unlucky.

Now was a hot goalie the only thing that prevented the Gophers from winning this one? I think one area the Gophers could’ve improved on and have complete control over is how they change lines when a defensive pair exits the ice. Everyone who watched the game remembers how Peterson threw the puck up ice for Notre Dame’s first goal of the game (not only was he outstanding at playing his position, he made a great offensive play as well). The Glover/Johnson pair had just been on the ice along with the 3rd line for a long shift in which they got pinned in their zone for an extended period of time. They were obviously gassed but the problem was that everyone except Johnson got off the ice at the exact same time on the long 2nd period change (ie the player benches are now further away from the D zone). If one of the forwards just stays on the ice and forechecks the goalie it’s significantly less likely that Peterson can get the puck all the way up the ice to the other blue line for a 2 on 1. This wasn’t the only instance where there was a bad line change involving Glover. Midway through the third Glover and Zuhlsdorf tried to exit the ice as Cammaratta was chipping away at a puck near the boards in the neutral zone but never had full control. Dello, a 3rd pairing Dman for the Irish made a nice pass to spring Ogelvie for a breakaway.

glover_change

glover_change2

How this type of thing happens twice in an NCAA regional where your entire season is on the line is beyond me. You can definitely blame Glover, the right defenseman, for going all the way across the ice to make a change when his teammate doesn’t even have control of the puck. But at the same time when this type of thing happens twice in a potentially season ending game do the coaches not take some blame for this? Luckily Schierhorn came up with a huge save on the breakaway above but these kinds of ridiculous mistakes can’t happen in this type of setting vs good teams.

So what do we take away from this game? Like I said above, the modern era of hockey is a low scoring affair and if you’re playing against an elite goalie in a one and done format you better hope that your’s is playing his A+ game as well and that your team isn’t making silly mistakes. Unfortunately neither was case and the Gophers had another disappointing end to an otherwise successful year. Having depth and generating chances is obviously good thing but it isn’t a prophecy that guarantees a victory. Unfortunately the Gophers learned that the hard way in this one.

* All stats recorded at 5 on 5 play

 

The Importance of Ryan Lindgren Quantified

After seeing Ryan Lindgren crash into the boards during the last regular season game of the year I knew that it was likely the last time we’d see him until next season. These suspicions were later confirmed after it was announced that he suffered a broken fibula. In these instances a lot of fans say something along the lines of, “next man up!”. In some instances that logic makes sense, every team suffers injuries throughout the year and most likely there won’t be many teams entering the playoffs with every single player healthy.  In this case though it’s probably not that simple, in fact Ryan Lindgren might already have been the best defenseman on the Gophers.

Before diving into the stats let’s just talk about the kind of season Lindgren was having and some of the narratives surrounding his play. When the Big 10 announced its awards this year Bischoff was named the defensive player of the year and his partner Ryan Lindgren made the all freshman team. He also made the world junior team and was considered one of the best USA born defensemen at his respective age. Obviously Ryan has a great pedigree and was having a good season, but just how good was it?

As stated in the previous post I have tracked 8 games so far this season and examined every 5 on 5 zone entry by both the Gophers and players of the opposing team as well as the zone exits made by the Gophers. The sample of games include the Friday SCSU game, both games vs the UND Fighting Hawks, the home Michigan series, the Mariucci Classic, and the Northstar Cup game vs Duluth. One thing to keep in mind here is that Lindgren was gone for the World Junior Championships (and got sick) which took place during the Mariucci Classic. As such Lindgren didn’t make it to either game and so his sample doesn’t get a boost like some of the others from playing against a couple of the weakest teams the Gophers would play all year (see below). PWR_current

The first thing we’ll examine is how Lindgren stacks up compared to his teammates in terms of allowing zone entries. Defenders that allow a lot of clean zone entries and have poor gap control are going to put their team at a disadvantage in terms of allowing more shots and quality opportunities. On the other hand if they can force the forwards to dump the puck in and grind along the boards to get a shot this should theoretically result in less shot attempts against. So does this hold true and how does Lindgren stack up to his teammates?

Gopher_D_Targets

There’s a couple things to look at here. Lindgren’s rate of allowing controlled entries is one of the best rates on the team. Interestingly enough he’s right in line with Collins, a guy who has been having a very nice comeback year and the other freshman Zuhlsdorf. In terms of shot attempts against it doesn’t appear that there’s any real pattern here. The 3 guys with the lowest amount of controlled entries allowed have a slightly lower shots against rate but it seems to be minuscule in this case. Perhaps there’s a better way to analyze that portion or we just need more data. But in this small sample Lindgren has been one of the better defensemen in terms of gap control and neutral zone defense. You can also see on the right that I added a quality of competition metric. This is measured by the amount of times the defenseman in question was targeted by a top 6 forward on the opposing team. This metric isn’t perfect obviously, defensemen tend to play with a partner and so if Bischoff is mostly playing against top 6 guys Lindgren would be as well seeing as they’re partners. This metric doesn’t do a great job of measuring the amount of time one the ice spent against the opposing top 6 but I thought it would be at least somewhat interesting to look at over these games.

Where Lindgren really shines though is in terms of controlled exits. As stated in the previous articles, exiting the zone cleanly allowed for more shot attempts at the opposite end of the ice per Jen Lute Costella’s study and most would agree that having a clean breakout is extremely important for any hockey team to be able to do.

zone_exits_Gophers

Glover

Lindgren’s ability to pass the puck and break out of his own zone was really impressive over this sample of games. Not only did he have the highest rate of controlled exits but his fail rate was low and he faced a good amount of pressure*.

So was the loss of Lindgren the reason we struggled and ended the season with a whimper? It probably doesn’t explain everything but I do think it at least partially explains why we got outplayed heavily by Penn State for much of the second, third period, and first OT. As for the Notre Dame game the Gophers controlled play for the most part out attempting the Irish 55-36 at even strength but only 25 of those shots actually made it to the net and the 5 on 5 SOG were pretty even at 25-23. With one of our best passing Dmen in the lineup I have a feeling we would’ve transitioned out of our zone with a bit more ease and speed and likely that shot differential goes a little more in our favor. Hockey nowadays is generally a low scoring affair and games are won on the thinnest of margins. The Gophers still could’ve beat Notre Dame yesterday without Lindgren but it was a tall task no doubt.

Thanks for reading!

 

 

 

 

*I tracked pressure situations in which a player exiting the zone had an opposing player within a stick’s length reach away from them

Welcome to Gopher Stat Blog

Hi, my name is Kris and I’m a big fan of Gopher Hockey as well as hockey analytics. As someone who has a degree in economics and has studied some statistical theory I find it fascinating that people have figured out how to use applied statistics to build and improve hockey teams. The goal of this blog is to examine and analyze zone entries, zone exits, and other data related to the University of Minnesota Golden Gophers Hockey program. Some of these stats are available at sites such as collegehockeynews.com and others like zone entries and exits are tracked manually by me. I plan to make this data public at some point for the sake of transparency as there is some degree of subjectivity to how these stats are tracked and because this is my first time doing this.

So far I have tracked zone entries and exits at 5 on 5 (even strength) for the following 8 games of the 2016-17 season. I plan to do a short write up on these games and series to further dissect and analyze it from an empirical standpoint and see what went well or what went wrong in each game.

  • Friday SCSU
  • UND series
  • Mariucci classic
  • Michigan Series
  • Duluth

If anyone is interested I could track other games as well and will likely do a writeup for the upcoming NCAA tournament game(s).

Part of my inspiration for doing this is the work of other analytics bloggers who have made huge strides in furthering peoples’ understanding of hockey while also showing their methods and making their work public. One of the original pioneers in this are is Eric Tulsky, now a full time employee of the Carolina Hurricanes. Tulsky had many contributions to hockey analytics and was one of the biggest pioneers of the subject. He performed many studies such as showing the importance of regressing shooting percentages to better predict true shooting talent, showing the relationship between outshooting the opposition and out-chancing the opposition, showing that defensemen don’t measurably impact on ice save percentage, and making all of this easy to digest while also showing his work to prove his theories. One of his most important studies showed the relationship between zone entries and shot/goal generation. This article shows that zone entries in which a team successfully carries a puck into the zone generates more shots than dump-ins. This study that Tulsky also published goes into more detail on that and shows that goals are more likely to come from the puck being carried cleanly into the zone. This is not only obvious from the statistics, it’s also rather intuitive. Clean zone entries and rush opportunities drastically increase the chance of getting at least one shot off rather than zero. As such I think this is something important to examine with the Gophers.

Another part of my inspiration for doing this was to better analyze the defense of the team and provide some hard data to analyze the performance of the Gophers’ defensemen. People often remember the one big mistake a defenseman makes while ignoring a lot of the smaller good plays they make to positively drive possession and ultimately out-chance the opposition. Tyler Dellow, who formerly worked for the Oilers and now writes for The Athletic, had a great quote about this:

If someone asked me what I think the biggest failing of the eyeball test is, I’d respond that it’s the emphasis on the big mistake. There are gigabytes of information contained in a hockey game. So much information that I think it’s difficult for anyone to take it in and organize it rationally. The way that our brains deal with that is by focusing on the big mistake.

As such it’s often better to take an empirical look at these things to avoid issues such as confirmation bias, recency bias, and the various other cognitive issues that all humans employ. Jen Lute Costella examined the relationship between zone exits and shot generation and found that exiting the zone with control created more shots at the other end of the ice. This is intuitive, teams and players that tend to throw the puck out of the zone tend to have awful possession, scoring chance, and expected goal for rates. My goal is similar in that I wish to track which Gophers defensemen are successfully exiting the zone, failing to exit the zone, and dumping the puck out of the zone.

Finally one of my biggest influences for doing this project is Corey Sznajder who writes at The Energy Line. Corey does a great job of concisely explaining, analyzing, and visualizing various games and series at the NHL level. This article might’ve been of particular interest to Wild fans after last year’s playoff exit.  While my analysis doesn’t go quite into the level of detail that he does, I hope to try and create something similar and expand what I track as time goes on and I become more efficient at it.

I realize that this post contains a ton of information and links to studies throughout the years. The main goal of this post isn’t to inundate you with information but to explain why I am choosing to track this information and explain the importance of tracking these stats. Thanks for reading and I promise to publish more content in the near future.