Matt Martin is a Terrible Player Right Now...no, really: A statistical look at the 21-year-old rookie

There are players who are known for delivering crushing hits, causing opposing players to always look over their shoulders in fear, weakening their resolve.  And when these players do deliver hits, they're well timed, so that they knock the puck loose and result in less opposing scoring chances, resulting in better team performance.  The best of these players can also add in the scoring department as well. 

Matt Martin is not one of these players.  A statistic of hockey prospectus that attempts to measure the total value of a player called GVT (Goals Versus Threshold) - the hockey equivalent to WAR (Wins Above Replacement) - tabs Martin as a below replacement player; that is, it says that he's worse than just a random player available for every team to sign from the AHL to come and play on the big team.  That's pretty bad, actually.  And aside from Rick DiPietro, it's the worst on the team (aside from Jon Sim, who no longer is on the team).

Now, GVT has a big flaw...it doesn't take into account context and misses a whole bunch of other statistics...it also ignores the role of good or bad luck.  Thus I don't take it too seriously (it also heavily values goalies, even though goalie performance varies a ton due to small sample sizes).  But well, a look at other hockey statistics does not paint a pretty picture about Mr. Martin.

Table 1:Matt Martin and Zenon Konopka's Even Strength Statistics: 
Table Legend:
Relative Corsi = A Player's Corsi while they're on the ice MINUS the team's corsi when the player is off the ice. 
Corsi = Total Shots directed toward opponent's net MINUS Total Shots directed toward Your OWN Net (Per 60)
Normalized Corsi = A Player's Corsi rating normalized to show how the corsi rating would be if a player started an equal amount of time in the offensive zone as the player does in the defensive zone.
Relative +/- = A players +/- while on the ice MINUS their +/- while off the ice (with some minutes adjustment)
% of Starts in Offensive Zone = The percentage of faceoffs you're on the ice that are in the offensive zone (not including neutral zone faceoffs)
%Corsi Rel Qual Comp = A measure of the skill of the average opposing player on the ice.  (Basically it's the average Corsi Relative of the opposing players on the ice while this player was on the ice.)
%Corsi Rel Qual Team= A measure of the skill of the average teammate on the ice.  (Basically it's the average Corsi Relative of the linemates and D-Men who are on the ice at the same time of this player)
---------

Like I said above, GVT fails in a major way: it doesn't measure context.  Thus l've selected for comparison Matt Martin's most frequent linemate: Zenon Konopka.  After all, the poor performance of the fourth line could be as much Konopka's as Martin's.  There's also the fact  that Martin for 2 minutes a game is on a line with Trevor Gillies, who truly stinks at the non-fighting parts of hockey.  But, well at 2 minutes a game, that shouldn't impact Martin's statistics too much, so I've left Gillies' putrid #s off of the table above. 

In addition, I've chosen several statistics that measure the context of when Martin is on the ice.

• Corsi Rel Qual Comp measures the quality of each player's competition by looking at how good his opponent's are (through Relative Corsi).  Matt Martin is in the middle of the pack among Islander forwards in this measure, showing that he hasn't played especially tough or especially easy opposition.
  
• Corsi Rel Qual Team measures the quality of teammates by looking at Relative Corsi.  This measure states that Martin has the 2nd worst teammates of any other player....but it's misleading a little bit, in that it's counting Martin's own contributions as he's very frequently on the same line as Zenon Konopka.  And this measure is misleading in another way because it undervalues Konopka, as I said earlier. 
  
• % of Starts in Offensive Zone (also known as Zone-Start) measures the ratio of faceoffs on which a player is on the ice in the offensive zone as compared to how frequently he's taking defensive zone faceoffs.  So Konopka, who is a faceoff specialist, is super-frequently used in defensive zone faceoffs, and has the lowest zone-start % in the league.  Matt Martin's zone-start % is actually about average for the Islanders. 
  
  As you might expect, a player who starts more faceoffs in the defensive zone will have worse statistics, particularly in corsi, as that it's hard to have a positive shot differential if you're on the ice all the time for defensive zone faceoffs.  
• Relative Statistics:  To measure a player who plays on a team, the easiest thing to do is to measure a team's performance with that player on the ice and compare it to their performance off the ice.  Thus Relative +/- shows the difference in +/- per 60 minutes when a player is on the ice as compared to off the ice.  Relative Corsi does the same thing with Corsi (instead of +/-).  Thus, when Matt Martin is on the ice, the team's +/- (goal differential) is 1 whole goal WORSE than with him OFF the ice. 
• Normalized Corsi is a measure that well, let me explain:

So, the Zone-Start % is a major factor in influencing corsi (shot differential) for obvious reasons outlined above.  What's been found is that each additional offensive zone start results in an increase in a player's total corsi of +0.8...in other words, for each additional offensive zone faceoff taken by a team, the team puts .8 more shots on an opponent's net than the opponent puts on their own net.  This can be a big effect for extreme players, like Zenon Konopka (who as previously mentioned sees more defensive faceoffs than anyone else in the entire NHL).

So what I've done is I've used this figure (+.8 per additional offensive zone faceoff) and calculated what Konopka and Martin's Corsi would be if they were on the ice for an equal amount of offensive faceoffs as defensive faceoff (a zone-start of 50%).  This eliminates the effect of zone-start on corsi and makes a big difference here. This is listed in the table as Normalized Corsi

See, Matt Martin and Zenon Konopka are tied for the worst relative corsi numbers on the ENTIRE ISLANDERS.  And it's a bad relative corsi, when either player is on the ice, the Islanders are outshot by 21.1 MORE SHOTS Per 60 minutes than they are when the two players are OFF the ice.  That's pretty bad (and the Isles are outshot when both players are off the ice as well).  This of course shows up in their normal Corsi #s, in Table 1, which are the two worst on the team:  when each player is on the ice, the Isles are outshot by over 26 shots per 60 minutes.

However, as I've said above, corsi and relative corsi are heavily affected by zone-start %s, so I've calculated the Normalized Corsi of both Matt Martin and Zenon Konopka.  For Zenon, there's a huge change...his corsi per 60 improves by over 21 points.  In other words, the teams'  poor shot-differential while Zenon is on the ice is almost entirely (21/27~=75%) the result of how Zenon is used: as a faceoff specialist meant to limit the damage of defensive faceoffs.  That's not his fault, and it's not the fault of his own play...it's just the fault of how the guy is used by Gordon/Capuano. 

But for Matt Martin, as you should have expected right now, there is little change (the change in shot-differential is just over 3 shots per 60), showing that only 1/9 of Martin's terrible Corsi ratings are attributable to how the kid is used. 

So What Does This (and the other numbers up there) Tell Us?

What this tells us is that taking context into effect, the team's horrible performance while Martin is on the ice is NOT the result of poor play of his most frequent teammate (Zenon Konopka), but is almost certainly due to the poor play of Martin himself. 

And poor play it is - the Islanders are massively outshot, as discussed above, with Martin in the game, and it would seem to mainly be his own fault.  He can't even use the quality of his opponents' as an excuse for this shot differential - as mentioned above, Martin's faced only average opposition. 

In addition to causing the Isles to be massively outshot, Martin commits 1.6 minor penalties per 60, but only draws 0.6 penalties per 60, a difference of 1 whole power play per 60 minutes.  Remember, a power play results in a goal roughly 20% of the time, so every 60 minutes, Martin is causing the Islanders to be down an additional 0.2 goals.  If we pro-rate his play over 80 games, we find that Martin would roughly cost the Islanders 2.5 goals over an entire season, simply due to taking so many penalties.  That's obviously not good - hockey statisticians basically estimate that 6 goals = a win (in baseball 10 runs = a win for stats like WAR), so from penalties alone, Martin costs the Islanders almost a half a win. 

And of course, if you want to use +/- (which I don't love either) and dont like corsi, Relative +/- states that for every 60 minutes Martin is on the ice, the Isles are outscored by 1 more goal than they'd be if he was off the Ice.  Thus his ice time result in the Isles being outscored by 13 goals per 80 games MORE than they'd be if he was off the ice.  Pretty Miserable. 

Conclusion:

Matt Martin is 21 years old.  He is a young kid, who could mature into a player who has some use to a hockey team outside of fighting.  But right now, Matt Martin is a terrible terrible player, whose play would seem to result in the Isles being massively outshot and outscored, and who doesn't help the team by taking so many penalties. 

And there is a place for young players, who have potential to grow into good NHL players, but are currently horrible in the NHL:  It's called Bridgeport.  That's where Matt Martin should be, and that's where he should stay until Martin can prove that he can handle the NHL.  And the team shouldn't consider Martin as a key cog of the future until he can show that he has greatly improved.