NHL Predictions 2013: The Return of the SONA Index

In December of 2011, I wrote a piece called "Introducing the SONA Index." That piece discussed a new NHL statistic I had developed based on the number of registered sex offenders that lived near the arenas of various NHL teams. When I wrote it I had no idea how powerful of a tool it would turn out to be.

In March of 2012, I revisited the SONA rankings, and discovered that there was a very strong correlation between SONA rankings and the overall NHL standings. If we remove Vancouver from the equation (for filthy Canadian reasons we will discuss later) five of the top six SONA teams finished the regular season in the top six in points in the NHL, making SONA a better predictor of performance than just about any other measure we have.

Purely by accident, I discovered the most useful NHL statistic in history.

But SONA wasn't without its problems. By basing the measurement on a radius around each arena, teams with arenas in non-residential districts were unfairly penalized. The same is true for teams with arenas next to large bodies of water, since lakes and oceans tend to have fewer sex offenders living in them than other areas (though few would deny that mer-rapists are especially viscous).

To try to address the shortcomings of SONA, and to see if I could make it an even more accurate statistic, this year I made a few changes.

Methodology

The 2013 SONA rankings are based on information from the website citydata.com, which helpfully combines a bunch of different statistics about cities. The part of the website I was interested in uses the publicly available list of sex offenders in the U.S. and compares it with an area's population overall. To limit our results as best we can to the sex offenders NEAR the arena (the whole point of this stat), I used the arena's zip code rather than the city as a whole. Using this information will allow us to determine each team's SONA P/C, or "Sex Offenders Near Arena Per Capita." This should be an improvement over last year's numbers, since it will correct for the different population densities around the various arenas in the United States.

Last year's SONA Index was based on raw numbers of sex offenders in the radius around the arena, which meant that Nashville's league-leading 458 corresponded to 458 individual sex offenders. This year the numbers will be a bit different: the rankings will be based on the ratio of sex offenders to "normal" people in the zip code. So, for example, Minnesota's 2013 SONA is 1/3547, which means that one out of every 3547 people who live around the Xcel Energy Center are registered sex offenders.

Methodology, Eh?

There has been no change in Canadian law since the last time I updated SONA. Canada doesn't release information about sex offenders to the public, and I am more convinced than ever this is because more than one out of three mounties have committed sex crimes in the past.

If this ISN'T true, Canada, then release your sex offender registry and prove me wrong.

Clearly, the Canadian government doesn't care about the welfare of its citizens. And really, it's hard to blame them. I don't care about Canadians either, after all. But I DO care about SONA, and just like last time the Canadians are screwing it up. I tried a bunch of different methods to come up with a number that would allow us to incorporate Canadian teams into the overall SONA rankings (if you have suggestions or data sources, please offer them in the comments) but in the end I decided to repeat what I did last time and use Maclean's rankings of the most dangerous cities in Canada to get a rough estimate of Canadian SONA (CSONA). I'll rank the Canadian cities separately, so unfortunately we can't compare them directly to the U.S. teams - but it doesn't really matter since none of the Canadian teams are going to win anything important.

With all that out of the way, let's move on to the numbers.

Here is the SONA P/C for 2013:

And the 2013 CSONA:

Analysis

• Your 2013 SONA champion: Buffalo! They moved way up this year, after placing 11th in the previous SONA index. At first I was surprised to find out that 1 out of every 83 people that live near First Niagara Center are sex offenders, but then I thought about it for a bit and remembered a couple of things we learned about that area in the past year, and now I'm not surprised anymore. Congratulations, Buffalo!
  
• St. Louis and Nashville remain high on the rankings this year. Tampa Bay made a HUGE jump. Detroit dropped significantly, which fits right in with how crappy they're going to be this year.
  
• Vancouver and Edmonton swapped places on the CSONA rankings, but other than that there were no changes north of the border.
  
• Boy I really wish we had better numbers for Canada, since it certainly stands to reason that Vancouver, which has a greater population density than Winnipeg, would have the highest SONA P/C of all the Canadian teams, and might even place near the top of the overall rankings. That would likely bolster the predictive validity of SONA, and, more importantly, would also be funny.
• If SONA turns out to be any sort of predictor of team performance this year, Minnesota is in for a rough ride. If Wild fans hope to see their team make some noise in the playoffs this season, it's clear what they need to do.
• Columbus in 5th place certainly seems like a misfire for SONA, but I'll withhold judgment. After all, last season none of the top 8 SONA teams missed the playoffs.

So there you have it everyone: the 2013 SONA index. Will it turn out to be as remarkably good a predictor of team performance as it was last year? Only time will tell.

Probably not though.