Colby Armstrong data dump

This isn’t pretty. I’m also only doing this because @67sound asked me to, but I looked at the Leafs with and without Colby Armstrong over the last two seasons.

First, what is Colby Armstrong? He was an average scorer before coming to Toronto, busted up by a pretty favourable shooting percentage in Atlanta:



  G/82 SOG/GP Sh%
Pittsburgh 16.8 1.74 11.7%
Atlanta 18.8 1.51 15.1%
Toronto 9.3 1.05 10.8%
Career 16.3 1.52 13.0%

Brian Burke signed Colby Armstrong after his stint in Atlanta that ended after the 2010 season for $9M over 3 years. I don’t begrudge Armstrong for this. For sure the guy is a good, honest player who worked hard to get to where he’s at, and it isn’t on him for accepting a suitcase full of money.

What’s also interesting is his underlying numbers, which were surprisingly high:



  Corsi Rel Cor Rel QoC Ozone %
2008 8.1 0.956 39.1
2009 10.7 0.691 40.3
2010 0.9 0.603 47.1

He played some tough minutes, and managed to do better than his teammates. His Corsi/ON was slightly below zero, but that’s because he played on some pretty bad teams. His first season in Toronto was innocuous enough: he had a 2.1 Corsi Rel in slightly easier minutes (.321 Corsi Rel QoC and a 46% Ozone rate) but this last season was a disaster. If you understand underlying numbers by now, you don’t need to be told twice.

The real issue is that Armstrong had a modest play-driving and scoring ability, but he completely lost that last year. He stopped moving the puck forward, he stopped creating and generating shots, and he does nothing that makes his teammates better.

Using the timeonice scripts from here and here, I’ve sussed out how the Leafs did with Armstrong on the ice and off the ice in the last two seasons with the score tied:



  GF GA SF SA MF MA BF BA Fen% Cor%
With Colby 15 16 127 161 56 83 98 90 43.2% 45.8%
Without Colby 94 104 1048 1196 483 610 661 638 46.0% 47.3%

We know how much I like Fenwick Tied as an indicator of future success, and the Leafs were nearly 3 percentage points better with Colby off the ice. But that could be determined by tough minutes, with Armstrong creating opportunities for the Leafs to pick up the slack.

So using these pages, I sussed out the overall games that Armstrong missed. The results are more striking:



  GF GA SF SA MF MA BF BA Fen% Cor%
With Colby 55 67 553 702 267 376 394 396 43.3% 45.2%
Without Colby 54 53 622 655 272 317 365 332 48.0% 49.2%

With Colby Armstrong in the lineup, the Leafs had a Fenwick Tied rate of .433 over the last two seasons. Without, they were .480. That’s not necessarily good, but it’s a pretty big gap. Could be coincidence, who knows, but the Leafs have been decidedly a better team without Armstrong than with.

Because in Atlanta… with Armstrong on the ice and not:



  GF GA SF SA MF MA BF BA Fen% Cor%
With Colby 27 26 352 409 132 134 167 173 47.3% 47.7%
Without Colby 110 157 1435 1786 530 665 653 818 44.3% 44.3%

And with Armstrong in the lineup (since Armstrong only missed three games in 2009 and 2010, I used the Thrashers record from before the trade that brought him over in 2008 as reference, but it didn’t make a huge negative difference):



  GF GA SF SA MF MA BF BA Fen% Cor%
With Colby 97 133 1294 1543 500 521 585 705 46.3% 46.0%
Without Colby 40 50 493 652 162 278 235 286 41.5% 42.3%

So, causes? What caused a player with pretty good possession statistics in Atlanta to completely fall off the map in the last two seasons? System? Trust? Role? A flaw in advanced statistics when players move teams? Or was it just all the injuries that made it a lot tougher on Colby than we think?

I’d like to hear theories, and whether we can trust the guy to stay healthy for an entire season to regain the form he displayed in Atlanta.