So you’ve decided to go with James Reimer

Brian Burke addressed the media today in Toronto. There is lots to go through, but the big thing is the goaltenders. From our pal Sean Fitz-Gerald:

Is there any reason this is anything more than damage control at this point and saving face? Well, maybe. Burke has talked about adding a goalie, but truly, unless Burke was willing to pay an expensive premium for Roberto Luongo, nobody really popped up.

I think going with Reimer is the right move, and the vision of many a Leaf fan is clouded from the perception of Reimer after last season.

So you’ve decided to go with James Reimer. What we really need to know about him is that he had a very strong first season and a very weak second season.

The “Sophomore Slump” is athlete-speak for what we know as “regression to the mean”. This pertains to things outside of sports. In school, for instance, half the class of students will take one edition of a test and the other half will take a second edition. The next day, the students switch tests. It’s found that the students who rank the highest and lowest on the first day will wind up closer to the middle on the second day.

I’m sure somebody with a math background could explain this a little better, but when applied to sports, regression to the mean explains things from the Sports Illustrated cover jinx to the sophomore slump. The first, because the athletes placed on the cover of SI are usually there because they’re in the middle of a hot streak or a hot run, and when their luck comes to an end, they “perform” worse than they were. The second, because often when a rookie comes into the league with excellent numbers, some of it is luck-driven as well, and if it goes the other way, you get James Reimer’s season.

It’s important to look at both years for Reimer when assessing his performance. I looked at a few goaltenders who, in their first seasons, played about the same amount of games as Reimer, and checked to see whether their performance in their third year was influenced more by their second year or their career to date.

Here is the “total”, representing Years 1 and 2.

The r-squared, or correlation, between the Years 1 and 2 numbers and Year 3 was .66387.

But how did those same goalies do in their second year? Well, of the 13 goalies in the sample, just five improved on their first year numbers. This is likely due to “survivor bias”, and goalies who fared poorly in their first season didn’t have a chance to recover. Brian Boucher had the most notable drop from Year 1 (.918) to Year 2 (.876) but recovered in Year 3 with a .905.

The r-squared, or correlation, between the Year 2 numbers and Year 3 was .50158.

In short, if you look at Reimer’s statistics to try and predict his third year performance, it’s best to also include his first season, wherein he went 20-10-5 with a .921 save percentage and a 2.60 goals against average, along with his second season where he went 14-14-4 with a .900 save percentage and a .310 goals against average.

It isn’t a question of “which James Reimer will show up in Year 3”, it’s that James Reimer is closer to being a 34-24-9 goalie with a .911 save percentage and a .283 goals against average than he is to either of the two extremes.

If you extrapolate his record across 50 games, using both seasons it looks far more promising than just the second:

Unfortunately, it’s been a while since we’ve seen Reimer perform at a very high level, so naturally our vision is clouded. Reimer had a rough season, but he was still about the league average at even strength. He’s a capable netminder who has the capacity to be a starter going ahead if his health is there. I’m worried more about another concussion spelling his doom than I am a repeat .900 save percentage.