clock menu more-arrow no yes mobile

Filed under:

Thrashers: With Or Without You

New, 7 comments

In the 1980s the rock band U2 had a mega hit titled "With or Without You" in which the singer lamented his romantic troubles. This post is not about romance, but it is about how the Thrashers perform with or without certain guys in their lineup.

It is a pretty simple idea really. What is the team winning percentage when _____ is in the lineup and what is the winning percentage when they are out of the line up (winning % = points earned per game). I'll save the caveats until later, for now let's look at the numbers. In the table below "+ Bogosian" means Bogosian in the line up and "- Bogosian" means Bogosian not in the line up.

Atlanta Thrashers Defensemen

Player Games Point Wining %

TeamPoints Pace

+ Bogosian 28 .411 67
- Bogosian 31 .406 66
+ Schneider 46 .402 66
- Schneider 13 .423 69
+ Valabik


.424 70
- Valabik 26 .385 63
+ Exelby 36 .431 71
- Exelby 23 .370 61
+ Oystrick 39 .410 67
- Oystrick 20 .400 66

Comments: The Thrashers winning percentage is basically unchanged regardless of whether Bogosian and Oystrick dress. The biggest gains come when the two physical defensemen Valabik and Exelby dress.

People sometimes say to me that "stats can't measure everything" which is true in a  limited sense. However, if physical play IMPROVES a team they it will show up in at least one stat-winning percentage. However, the Thrashers improved winning percentage with Valabik and Exelby in the lineup is rather puzzling in this respect. I do believe that physical play defensive defenseman can have an impact on the game--but I expect that impact to show upon the defensive stats.

But if you look at the table below, you'll see that the Thrashers have an IMPROVED record when Valabik and Exelby dress because the OFFENSE is better. When those two guys dress the offense averages over 3 goals per game and when they are not dressed the offense falls to 2.65 GFA. In fact, when Exelby dresses the team GAA is actually worse and the PK is MUCH worse. At least in the case of Valabik the GAA and PK% are slightly improved.

Players Status GFA GAA PP% PK% ST% SV%
+ Valabik 3.09 3.48 22.3 75.6 11.0 88.8
- Valabik 2.65 3.54 17.9 74.8 9.6 89.6
+ Exelby 3.06 3.56 21.7 72.1 11.2 89.5
- Exelby 2.65 3.43 17.9 80.2 9.2 89.6

So what explains the Thrashers improved performance? Better offensive execution. The Shooting % is higher and the Power Play % is much better. Considering that neither Valabik or Exelby shoot much or play on the PP units, I have a hard time drawing a firm causal arrow between their presence and the team's improved performance. While the team has won more when they have dressed this season, their influence does not show up in the expected categories and therefore I'm more inclined to give credit to the offensive players than these two guys.

Atlanta Thrashers Forwards

Thankfully the Thrashers best offensive forwards (Kovalchuk, Little, Kozlov, White) have essentially missed zero games so they do not show up in the table below. Instead we see mostly checking line forwards and offensive players who have failed to live up to that label (Christensen, J. Williams)

Player Status Games Played Point Win % Team Points Pace
+ Peverley 16 .469 77
- Peverley 43 .384 63
+ J.Williams 41 .390 64
- J.Williams 18 .444 73
+ Christensen 42 .429 70
- Christensen 17 .353 58
+ Slater 46 .424 70
- Slater 13 .346 57
+ C. Stuart 10 .450 74
- C. Stuart 49 .398 65
+ Crabb 26 .288 47
- Crabb 33 .500 82

Comments: Rich Peverley and Jason Williams are essentially opposites so far this year. One guy departed just as another guy left. The Thrashers have been a much better team in the 16 games since Peverley has replaced Williams. The Thrashers are playing at 77 point pace with Peverley which is a big improvement--but even a 77 game pace puts the Thrashers just one point ahead of 26th ranked Toronto. Even with Rich Peverley this team is not even close to being a playoff club.

The team is also significantly better with Christensen and Slater in the line up. Honestly, this comes as a major surprise to me. In fact, this pattern causes me to question the very utility of using winning percentage. Christensen has contributed almost nothing offensively all year and yet when he dresses the team's offense is better? Is he the cause or is this a coincidence? I think it is the latter. Likewise when Slater dresses the PK is MUCH worse and the offense is much better--given the fact that Slater's line rarely scores and he is PKer--I'm not sure how much credit he deserves.

Finally, in the AHL call up battle between Crabb and Stuart, Colin Stuart is the run away winner, but his sample size is just 10 games so take it with a huge grain of salt. The team was simply dreadful during Crabb's 26 NHL games but I have a hard time putting all the blame on a guy who averaged just 12 minutes of ice time per night.

Atlanta Thrashers Goalies

Fundamentally the goalies only control 1/2 of the game--preventing goals. And they control less than half since the skaters largely determine shot prevention for the most part. Goalies must do their best to stop those shots.

Players Status Games Played Point Win % Team Points Pace
+ Lehtonen 32 .375 62
- Lehtonen 27 .444 73
+ Hedberg 24 .417 68
- Heberg 35 .400 66
+ Pavelec 11 .273 45
- Pavelec 48 .438 72

Comments: The team has a better SV% when Lehtonen plays in a game, but they simply don't score for Lehtonen. That's the story here. Consider the average offense for these three goalies: Hedberg 3.46 GFA, Lehtonen 2.63, Pavelec 2.55. Hedberg has had the benefit of nearly a full goal more of offense to work with. If Lehtonen got that level of offensive support his winning percentage would be much better.

Conclusion: I started doing this I had hopes that it might prove insightful, but now that I've calculated the "with or without" numbers I'm less enthusiastic. My guess is that the sample size is too small for most of these guys and there are confounding factors (like the power play being more productive when defensive guys are dressed) which make it very difficult to sort out cause and effect. I almost chose not to publish this, but I figured some readers might find it interesting.