Union and Blue: Weekly Stat Round Up 11/15


Nov 13, 2015; Pittsburgh, PA, USA; Columbus Blue Jackets center

William Karlsson

(25) skates with the puck as Pittsburgh Penguins defenseman

Ian Cole

(28) defends during the third period at the CONSOL Energy Center. The Blue Jackets won 2-1. Mandatory Credit: Charles LeClaire-USA TODAY Sports

Welcome back to the Union and Blue Weekly Stat Round Up. Last week we covered how the season had been going up to that point, in a word, poorly. The team sat at 4-11 and the advanced stats indicated that was right about where they deserved to be. Fortunately this past week included two wins in just three games and pushing their record up to 6-12.

The question remains, has the process changed? Or is this improvement based purely on puck luck?

Going back to the tried and true method of process evaluation, shot differential, we see that this week went basically the exact opposite way that probability would dictate. The Jackets were out shot in both victories by a combined 8 shots (57-49), while strangely maintaining superior possession in the defeat (45-30). In total, the shot differential for the past week was +7, or +2.3 per game, much better than their season long -0.4, which last week sat at -0.9.

It’s things like this that make shot differential look questionable in the eyes of more traditional fans, when the results diverge so far from what the process suggests it’s natural to start questioning whether or not we’re tracking the process correctly. But in truth this is simply a reflection of small sample size. Shot differential takes a long time to even out, taking a week by week nap shot can be interesting but not necessarily predictive of the future. This same concept is represented in more simplistic statistics, like goals. If a player scores a goal on opening night then he is on pace to score 82 that season, two goals and hes got 164, a hat trick and I think you get the idea. People accept this and forgive it as small sample size, is it to far a stretch to do so with more advanced statistics like shot differential?

All of this indicates that while the Jackets were likely out played this past weak, the hockey gods were with them and statistical variance broke their way. This is the polar opposite to the first 10 games, where there were several examples of the exact opposite happening. Nonetheless if they’re going to turn this season around and get back into the playoffs they’ll need to see some improvement in the way the team controls possession, relying on variance is never a good way to go.

An interesting side story this past week was the appearance of Kevin Connauton in all three games, averaging about 15:00 minutes TOI. I wrote about Connauton earlier this season, comparing how he was very effective producing goals last season, but appears to be struggling earliy. Noting that he spent nearly all of his TOI playing with James Wisniewski. I asked if Connauton’s success last season should be credited to Wisniewski’s passing skills. Well we’re twelve games in and some interesting results have been returned.

As you can see he’s still getting shots on net, though the rate at which those shots become goals has cratered. It remains to be seen whether this sample is a fair reflection on Connauton with the team struggling as a whole. The conclusion to be made from these numbers, if Connauton is not paired with a gifted puck mover like Wisniewski, you can expect a decline in his play. Another interesting fact is his penalties per game have doubling since last season, one has to wonder if the lack of discipline is directly related to his questionable shot selection.

So there’s the last week in in review, it was overall a positive week from a results standpoint winning 2 of three and moving out of last place in goal differential. Unfortunately the process appears to remain the same and it’s doubtful that any of this predicts a significant improvement down the line. As of today playoff odds stand at 4.2%, nearly double what they were last week.