The Tanner Glass Experience

Whether Tanner Glass plays one game in 2016/17, or the rest of them, it probably doesn't matter much for the NYR's chances at the 2017 Stanley Cup. The Defense is just so so bad.  

But it's fun to look at players on the extreme of stats - the truly great and the truly awful. Take a guess which side of that scale Glass falls. Using, I compared Glass to all forwards who have played over 2010 5v5 minutes since the start of the 2007 season.  There's no magic to 2010 minutes - it seemed like a good way to ensure any "full time" skater would get included, even if they only play a few minutes a game (and as I've said before, Corsica's time-on-ice slider gives me nightmares).  As you'll see, for some stats, Corsica returned an "N/A", so I noted those few cases in the charts.  I then ranked Glass' results and his overall percentile. I focused mostly on rate stats to help account for time on ice. All data is score, zone, and venue adjusted.

First, Glass' individual stats:

click to enlarge.

Red is bad, and there is a LOT of red here. Noteworthy among the bad are his individual shot metrics - bottom 5% for individual shots (Corsi, Fenwick and Shots on Goal), and bottom 2% for individual Scoring Chances and individual Expected Goals. Not surprising that he's near the bottom - he's not around to directly create offense - but maybe somewhat surprising that he's SO low.

On the positive side, I guess, he has thrown a lot of hits, and has drawn more penalties than he takes. He also takes a lot of hits. These are cumulative hit stats, and I'll have more to say about his hit results at the end of this post.

Next, how Glass' team fares with him on the ice.

click to enlarge

Lots of deep, deep red again.  Glass' teams take few shots when he's on the ice, and give up a lot. Bottom 5% in CF60, CF%, FF60, SF%, etc. Shots against are bad too, but at least he gets up towards the bottom 1/3! Aim high, Tanner (except when throwing those hits, of course).

Maybe expected goals is your thing? 

click to enlarge

Hardly surprising to see a lot more red. I guess we should be happy that his teams are only just terrible defensively by this metric? Look I'm trying to be nice here. 

Ok, now the real fun - how his team fares relative to when he's not on the ice.

click to enlarge

Now we're seeing how bad, he's truly been. Like "Bottom 2 or 3 in the entire dataset."  He is DEAD LAST in Relative CorsiFor%.  2nd to DEAD LAST in Relative CF.60, Relative FenwickFor%, Relative GoalsFor per 60...and so on. You know how people say hard-to-prove stuff like "Tanner Glass Makes the Team Better?" Well here's the proof that it's true: Putting Tanner Glass in your lineup, and then immediately stapling him to the bench, makes your team better!!  [/sarcasm]

To his credit, Tanner has consistently started his faceoff-beginning shifts in his defensive zone more than others. There is some evidence that zone starts don't matter *that* much. But, ok, let's give Glass some benefit of the Zone Start penalty...toss him a few % points if you want...and he'll end up going from "ALL TIME WORST" levels to just "ASTOUNDINGLY BAD".  

"But Rick" you say, "Glass is here for his physical play, not his point or shot creation. He creates space for his teammates on the forecheck." First, if that's true, you would expect to see it somewhere in his team on-ice results, and we see instead that his team is one of the WORST at creating shots and chances when he's on the ice (bottom 2% CF/60, CF% , and xGF%; and bottom 4% for xGF /60). 

Now it is true that, by sheer numbers, he throws a lot of hits. He's 15th in hits thrown of the 520 forward in the dataset. He's also pretty high up in hits taken (77th percentile) - and some of that *could* be hits taken in the offensive zone during a dump and chase or forechecking situation. On the other hand, we also don't know how many of these are the useless (to me) "finishing your check" type hits that happen after the puck has been moved. There is so much we don't know about where hits are given/taken, and how successful they are, but he's seems to be a high-event hitter/hittee.

But he's also been around for 10 years, and has played a fair amount of games. I went back to Corsica, and got a per-game listing of all games Glass has played since the start of the 2007 season, including playoffs. 

Some key takeaways:

  • In 50% of the games he's played, he's thrown 2 or less hits. 
  • In 92% of the games he's played, he's thrown 5 or less hits.  

Let's assume he gets at least five 5v5 shifts a game (1.3 per period). On that assumption, he'd average only about 1 hit per shift.   

Now, he generally throws more hits than he takes...but:

  • In 19% of the games he's played, he has taken as many hits as he's given ("Even").  
  • In 16% of the games he's played, he has taken more hits than he's given ("Minus").

Again, it's simplistic and probably wrong to assume that each hit taken is a "negative", but in 1/3 of all his games, he's taken more hits than he's given. Maybe these are really good rate and ratios such that Glass is one of the best "hitters" in the league, but I was expecting to see many more hits/game, and many more games where he gave more hits than he took.

So what difference does it make?

Glass is bad. "One of the worst forwards of the last 10 years" bad. Whatever he and his teammates are doing when he's on the ice, it hasn't worked. 

Again, this isn't to pick on Glass. I'm sure he's a Great Guy in the Room, etc. What this really shows me is that if a guy like this can stick around for a decade, then the NHL managers and coaches still have a long-way to go to become great talent evaluators. And for the NYR's is best summed up by must-follow Alex Ellenthal:

...Again I'm right in my analysis.