Matt Martin is still bad and should be the odd man out.

With Josh Bailey and Cal Clutterbuck apparently not badly hurt, the Islanders are soon about to have a fully healthy forward corp for the first time all year.  Even if we ignore Conacher (not likely to get playing time if people are healthy) and Boulton (goon – also injured for now), this leaves 13 forwards for 12 spots – someone suddenly has to sit.  5 games ago, when word of Grabner’s impending return came around, it seemed clear:  Someone in the fourth line would have to sit, probably Martin or Cizikas.

As if sensing this was the case, the fourth line has played inspired hockey in the last 5 games:
4th Line

The #s above show what many have probably noticed: the fourth line has played damn good hockey over the past 5 games and have the best possession numbers of any Islanders line during this stretch.  How can you bench some players playing truly great hockey?

The answer is simple: We have a lot more than five games to go on when making this decision.  I’ve repeatedly mentioned “Bayesian Analysis” on this blog when talking about analyzing players and teams and it comes into play here as well.   Again, “Bayesian Analysis” is simply a complex name for a pretty simple context:  When evaluating the play of players or teams or anything, we don’t simply forget all that came before.  We take new information (our “signal”) and add it to the prior info we have (the “prior”) and adjust our current belief as to that thing by combining the two.  The stronger our “prior” – the larger our sample in this case – the stronger the signal it takes for us to change our belief.

To go back to our fourth line players, well, our signal here is only five games worth of data.  Yeah the fourth line, including Matt Martin and Cizikas has played incredibly well for five games, but we have multiple years of data of Martin and Cizikas besides the data from this year to judge these players on:

The above is Martin’s 6 year history.  You can see the uptick in this year’s performance pretty clearly – it’s the first time Martin has ever been above 50% in corsi, and it’s driven strongly by this past 2 week’s performance.  And well, for the 280 games prior to this season, Martin has been an awful NHL player, with the team greatly better without him on the ice.  That’s what we call an incredibly strong prior and this year’s performance – and especially the last 5 games’ performance – really isn’t nearly enough evidence to take away from that.  The odds are incredibly unlikely Martin has morphed into a plus possession player, better than either Bailey or Grabner.  And he’s a worse scorer than either of those guys as well.

The same is true of Cizikas to a lesser extent – we only have two seasons worth of prior for him – but CC has the advantage of being a Center, which Martin does not.  Martin being removed makes lineup changes easy -> out Martin, in Bailey or Grabner, voila, keep all the other lines the same.  Even after two weeks of good play, it’s still the right move – odds are if he could keep up this good play for a full season, we’d have seen a player who wasn’t god-awful for a prior season.  We haven’t.

It may be tough to justify taking out a player playing the best hockey of his career, and hell, genuinely some of the best hockey of anyone on the team over the past two weeks.  But it’s clearly the move that makes this team better more than any other.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s