I don’t understand how the fact that Heatley would have led the Wild in 12 statistical categories last season makes them a better team. Just because an individual player piles up points and leads a team in goals, points, etc. doesn’t necessarily mean he’s making that team better—just because he’s scoring more doesn’t mean the team’s scoring more and, by association, winning more. Jarome Iginla had 86 points last year and the Flames didn’t make the playoffs.

This also fails to take into account the hole left behind by Brent Burns on the blueline and the possibility that shots and goals against could increase. If Heatley/Setoguchi can’t “make up” for that, are the Wild still a better team?

The comments on that post are endlessly frustrating.

I didn’t bother to return to the debate after I left my comment, because I didn’t really think the discussion would be elevated, despite my best intentions. Frankly my point wasn’t even about Heatley or whether the Wild was a better team, just a way to think about stats and evidence and your own biases when it comes to hockey analysis.

What I found most frustrating about the whole thing was that the Minnesota poster may have had a legitimate point, had he expressed it properly or found evidence to back it up.

For example, I think it is a legitimate argument to say “For various reasons, Havlat was causing dissension in the dressing room and on the ice and had to be traded. Heatley was the best return we could get.” That’s a fair point. I doubt that if Havlat had to be traded a better offer was in the having. But all this stuff about “chemistry” and “not fitting in” was just way overblown.

Also, I think it would be a legitimate point to say that Heatley will increase shooting percentage or chance conversion rates in a margin that will exceed any decrease in the possession rates of the Wild. But there are numbers out there that will show whether that is at all possible or not. To just assert it, without providing any evidence, is just lazy.

The whole point of the “advanced statistics” (which I think is a bad way of describing them) is not to supplant observation of the games, but to try and record observations in an objective and systematic manner. This allows someone analyzing the game to test a hypothesis or prediction against a raft of observations. Since I started following this stuff in ~2006 or 2007 there has been probably a dozen times I thought “That isn’t how I remember that happening”. Then you go and check the statistics, and lo and behold my recollection or observation was faulty.

That is what I really enjoy about the writing of guys here, Flamesnation, the Oilogosphere, old Behind the Net, is that they put their predictions and information out there with observed facts. If you think that they are wrong, they are usually open to convincing if you have the data.