Sample Sizes, Results-Based Analysis, and Eric Nystrom

Updated: January 10, 2018 at 7:13 pm by Josh L.

Source: Wikipedia Commons

Sample size issues have been known to cause crazy amounts of unwarranted hysteria. The 2011/12 season of Eric Nystrom should be Exhibit A in the coming years whenever a player comes out throwing bulls eyes to start a season out of nowhere.

There is usually a reason the career bottom six player with ten goals by Thanksgiving startles you. In your mind you know it shouldn’t happen, but it has. Yet, you look up again and said player is getting more ice time, playing against increasingly more difficult competition, and given more defensive responsibility. All of this after said player was given away by one of the worst teams in hockey to help the team you follow get above the cap floor after the New York Rangers ditched Sean Avery (which, I realize, makes very little sense). Fans had to attempt to reconcile what they saw on the ice with what they intuitively knew about the player.

This is the story of how Eric Nystrom confused an entire fan base for several months while proving how fickle small sample sizes and the results-based analysis springing from them can be.

The Rise of Nyzerman

Eric Nystrom made his Stars debut on the 18th of October after inexplicably finding his way to Texas. The bankrupt Stars carefully straddled the cap floor last year. Around two million of their 2011 cap space was dedicated to “the Sean Avery please leave” tax. When the Rangers got tired of his schtick the Stars dropped below the cap floor. Recently waived Eric Nystrom was Joe Nieuwendyk’s solution to the cap problem.

Nystrom wasn’t intended to be anything more than a bottom six minutes eater which, oddly enough, is pretty similar to what he was for most of his career after being a premium draft pick. After a week he began scoring though. The week of November 4th-11th Nystrom scored in four straight games. He went another week without scoring. In the seven games past the cold streak Nystrom got on the board with five goals and two assists.

Seems fishy, no?

It wasn’t fishy enough. Nystrom became an immediate fan favorite. It was hard not to root for the guy. The goals he was scoring were ridiculous. Every other one seemed to feature Nystrom jumping off the half wall into the slot, humiliating the left defenseman, and firing a low hard wrister for a goal. Example one is here. The problem is that expectations rose based on the limited sample.

Fans began clamoring for Nystrom to get more minutes. The Stars were struggling to score goals, and, hey, this guy has been putting them in the net. It’s a common, and logical, reaction to the situation. Nystrom never had the peripheral statistics to support that he was a brand new player though, and it caught up to him like it does for everyone eventually.

Through New Years Eve Nystrom had 12 goals. His previous career high for a season had been 11. He scored those 12 on just 59 shots. Nystrom was leading the NHL in shooting percentage when the calendar changed over at 20.33%. At one point he was as high as 25%.

The crash was anything but pretty. In the second half of the season Nystrom scored three goals on 43 shots. The 7% clip he shot the remainder of the year is much more in line with his career numbers, and much more representative of the player he actually is, but expectations had been raised by that point.

After the hot start Nystrom’s minutes continued to increase. The graph of Nystrom’s minutes per game is below, and as you can see there was a consistent upward trend until his 48th game of the season.

As the season moved forward he was established on a de facto checking line with Vernon Fiddler and Radek Dvorak. They skated together for just about the entire season. The Stars lack of depth undoubtedly played into the fact that they continued to be used as a checking line, but Dallas did have some other options so they don’t get to completely wiggle off of the hook. Their fourth line was particularly effective in transition and in their own end, even if they couldn’t do much offensively.  The checking line remained in tact though. After the hot streak the Dvorak, Fiddler, Nystrom trio had an aggregate 47% Fenwick% as a line which is in Columbus and Calgary territory.

The early season results more or less established the trio, and it would have taken something drastic to alter their role significantly. Results-based analysis failed the Stars.

The Moral of This Story

Drawing conclusions based on small sample sizes is very rarely going to be beneficial, and should be avoided at all costs. This is only the most recent example. It wouldn’t take long to find several others, and you can probably think of a few just off of the top of your head.

Eric Nystrom is a solid NHL depth player. His game is even fun to watch at times. It was easy for some fans to get sucked in by the hype and make bold proclaimations about his future. I vaguely remember someone guaranteeing me that he would score 30 goals at one point (though in fairness it’s been a while, and I can’t actually find proof that it happened). We see how that worked out. Nystrom’s situation isn’t unique. It will come up again, and very possibly in 2013. There’s nothing wrong with being a fan, but if you expect an outlier situation such as Nystrom’s to continue you are guaranteed to be disappointed.

Keep your eyes open and don’t get carried away by the hype.

Follow @JoshL1220 and @NHLNumbers on Twitter