I waited a long time to write this post. While combing through stats from last season I stumbled upon some truly startling evidence that illustrates just how valuable Colby Armstrong was to the Thrashers. I sat on this data until after his contract was signed (Note: if any of NHL agents are reading this and you want to hire a stats guy to help with your non-thrashers arbitration cases my email is on my profile page). Why? Because I think Armstrong is way more valuable than the traditional stats would indicate and now I'm going to explain how the new stats prove it.

Before I talk about Colby Armstrong, first let me explain how I evaluate players. Team Goal Differential explains 93% of the NHL standings. Players can contribute to a good Goal Differential three ways: terrific PP, PK or ES (Even strength) play. If a player doesn't drive the PP, PK or have a positive ES Plus/Minus number, then that player is not helping his team win hockey games.

Ilya Kovalchuk is a great power play guy, but contributes nothing on the PK and at ES his team was outscored 73-79 last season. Kovalchuk's primary contribution is his PP scoring. At ES Kovalchuk's highlight reel goals are wiped out by his own bad defensive play and those of his teammates. Simply put at ES Kovalchuk is doing worse than breaking even. I'll have more to say about this next week.

Colby Armstrong is the opposite of Kovalchuk--his primary contribution comes not on special teams but at Even Strength. Another key difference is that Armstrong doesn't contribute with highlight reel goals but by making subtle plays that keep the puck in the opponent's end.

There is a growing consensus in the hockey stats community that "puck position" (not "puck possession") is the single most important part of ES hockey. Hockey is part skill and part luck. The puck takes funny bounces and all things being equal, if you spend more time in the opposition end of the ice you're a more likely to turn the "lucky" aspect of hockey into goals for your team if you spend more time in the good end.

Tom Benjamin points out how Detroit won the Cup in 2008 not with "puck possession" but rather with "puck position" strategies:

First, they get the puck out of their own end better than anyone else. All seven of their defensemen are good passers and six of their seven (Chelios excepted) are very mobile. Furthermore, the forwards are all very responsible so the defenseman always has a pass to make. If the defenseman is forced to chip the puck along the boards, the forward is always there. If the defenseman has some time, a forward always seems to get open in the neutral zone…

Second, once the puck is in the opponent’s end, the Red Wings make it impossible for the other team to easily move the puck up the ice. Detroit attacks the puck in a swarm all over the ice. The puck possession meme is a myth. The Wings frequently lost the puck when on the attack in the Pittsburgh end, but when they did they immediately forced the Penguin defense and often recovered it.

If puck position is crucial to winning, then smart hockey teams will gather up players who play in the opposition end of the ice. And this is where Colby Armstrong excels. For the last two NHL seasons we have data on where players started and stopped a shift that involved a faceoff. Every player has a mixture of Offensive, Neutral and Defensive Zone faceoffs (more on this mix next week), but those players who end up with more Offensive Zone faceoffs than they started with are players who are moving the puck down the ice and applying pressure at the other end. Hockey stats guys refer to this as Zone Shift. Here are Thrasher leaders for the 2008-09 Season and guess who is the runaway leader for Atlanta?

One thing that REALLY has me encouraged about next season is most of the players with negative Zone Shift numbers last season will not be back. Gone are Christensen, Williams, Perrin, Havelid, Stuart and Exelby. The only returning players from that list are Kovalchuk and the over rated 4th line (Boulton-Slater-Thorburn) who get outshot badly despite their limited ice time. Our roster is much more suited for Coach Anderson's style of play than it was last August.

A major problem for the Thrashers is that they have been massively out shot for years--basically the entire history of the franchise. When you're spending a lot of time in your own end, it is very hard to out shoot the opposition. If "puck position" matters it should show up in the terms of which players do the best at out shooting the opposition. Let's see if that is the case, we will look at Shots Fired (aka Corsi Number) which includes Goals+Shots on Goal+Missed Shots+Blocked Shots. Once again we see that Armstrong is very near the top (I'll have more to say about this table next week, Schneider and Oystrick benefited from a lot of favorable faceoff starts). Guess who leads the Thrashers ES Net Shots Fired list? Colby Armstrong again.

Another reason I consider Colby Armstrong to be the Thrasher Even Strength MVP is that he made everyone around him better last season. Colby does many little things that go unnoticed by most hockey fans. He always clears his zone when he has the chance, he makes safe outlet passes. If there is no safe outlet pass he chips it up the wall or dumps into the offensive corner and starts the forecheck to get the puck back. He almost never ices the puck, he finds a way to gain ground.

If you share the ice with a guy like Colby that means you're playing more of your minutes in the good part of the ice. You're able to spend your energy wearing out the other team's players and when they do clear their own zone the opposition is tired and much less likely to get a high quality shot. The result is a better SV% when you're wearing out the opposition in their own end. Because Colby plays more at the good end of the ice, virtually every Thrashers who played minutes with Armstrong had a better ES SV% with him on the ice than when he wasn't on the ice. In some cases the differences are enormous. Enstrom had a personal ES SV% of .957 with Colby but just .907 without Colby. Rich Peverley also greatly benefited from playing with Armstrong at the good end of the ice.

Plus/minus is a often misused and misunderstood stat. If is wrongly used as a measure of defensive ability. If you strip out the SH stuff and the Empty Net situations, ES Even Skater Plus/Minus is a very handy measure of who is playing winning hockey on your team. Good players out score the opposition. There are ways to make Plus/Minus ever more accurate, but if we just stick with ES Even Skater Plus/Minus we can see that whoever played with Colby Armstrong was much more likely to end up with a positive number in our measure of "winning" hockey.

Several players have an enormous swing depending on whether or not Armstrong was out there with them on the ice. Kozlov and Kovalchuk and Perrin all posted positive ES Plus/Minus numbers with Armstrong and double digit negative numbers without him. The same goes for Ron Hainsey a more offensive defensemen who shines best in the Offensive and Neutral Zones where he can make good use of his excellent skating ability.

Conclusion: If we could clone Armstrong and have three Colbys that would make one heck of a checking line. I certainly hope the Thrashers re-sign him at some point because last season at Even Strength Colby Armstrong was the straw that stirred the Thrashers drink among the forwards.