Welcome to the second installment of this maybe-actual-but-who-really-knows weekly feature. Today there is nothing about Columbus but that doesn’t mean my fellow CBJ fans shouldn’t look into what’s happening in the rest of the league. What’ll we do this year during the playoffs otherwise? Anyway, here’s some stuff about two players who have last names starting with the letter V.
Regression is in Full Effect for Thomas Vanek But It’s Probably Ending Soon. Do you remember all of a month ago when Thomas Vanek was the front-runner in early season Hart Trophy talk? It was pretty well deserved at the time. For the month of January, Vanek was scoring at a blistering 2.5 point per game pace. Even Sid Crosby probably raised an eyebrow at that kind of production. But it was never going to last. During that crazy month, Vanek had a shooting percentage of 21.4%.
Now Thomas Vanek is sometimes overlooked when stars like Stamkos, Malkin and Crosby grab the headlines (and Vanek doesn’t quite fit in the fat contract he got via offer sheet from Edmonton), but he had the 8th most goals in the NHL between the last two lockouts. He’s a really solid sniper that helps drive the Buffalo offense. He is also absolutely not a 20% shooter. Nobody is. His previous season-highs were 18.1% and 19%, but his career average is a 15.2%.
Lo and behold, things started correcting around Mr. Vanek. After his insane January, a much more pedestrian shooting percentage of 13.3% followed in February. This meant his “cold” stretch between 2/10 and 2/28 (9 games with only one goal scored) was really not all that unexpected; his numbers simply had to correct at some point. Now, his season shooting percentage (as of this writing) is at a more sensible 16.4%.
Variance is to be expected out of shooters, especially when puck luck is such a large factor in the act of goal scoring. But streaks are not necessarily an unexpected or profoundly descriptive way of talking about hockey players. Over at PPP, there’s a great post on Phil Kessel’s cold start and how a streak is actually just something probable within the normal distribution. Taken as a whole, Vanek’s season looks inflated, but “normal” for shooting percentage. His increased number of shots are driving his higher scoring.
Conveniently for Vanek, the luck part of his game might be coming back to a more normal place. With that overall shooting percentage now at a reasonable range and his PDO at 1003. He’s only somewhat luckier than average now so the percentage of shots going in shouldn’t be expected to slip that much further. If he can maintain his higher-than-average shooting rate (that’s a different thing altogether, but is possible within a single year), Vanek may very continue on with his year-to-date scoring clip soon. It just won’t be like that whole January outburst.
It’s Too Early To Pass Serious Judgement on Tomas Vokoun. When the Penguins picked up Tomas Vokoun, it looked like a brilliant move. Their primary netminder is actually sub-par in the NHL, despite his first overall draft status. So to find the criminally-underrated Vokoun for cheap as a backup? On paper, it was a great way to pave the path to the Cup, giving Fleury less ice time and putting a real starting-caliber goalie in the crease.
But after 5 straight games with a save percentage below .900, Vokoun looks like crap, right? Let’s check out some graphs to aid our visualization of his performance. We’ll take the average of the game save percentages for Vokoun since the 2007-2008 season (his first with Florida) through this year, and then one column for his career average. That is, we’re not looking at the overall season save %, we’re looking at the average of game-to-game performance, what an “average night” would be for the goalie.
Boy, he looks pretty bad this year. And it’s not surprising, thanks to a sub-par .908 even strength save % and very poor PK SV% of .780. If either one improves, it’s probable that he doesn’t keep such an abysmal overall save percentage.
But what would happen if we put error bars of 1 standard deviation from each season on this same data?
One caveat here: we moved the scale on the right so we could see the top and bottom of the error bars a bit better. So we’re not quite seeing the same perspective, but that’s kind of the point. His performance this year is bad, but not quite so dire as it seems when considered in the context of sample size and standard deviation.
Let’s be clear: Vokoun is not having a good year on average. That even strength save percentage of .908 is not impressive, and a far cry from his .927 last year or the .919 the year before. But it is very early (in terms of data set size) and there’s a lot of variance in his game-to-game performance. One or two good games would thrust Vokoun’s numbers right back toward his career norms. If his special teams save percentages can be shored up (that aforementioned PK one at .780 is the biggie, way down from his past performances), it will absolutely bring him back to the land of average goalies.
Be upset at Vokoun so far if you want, but remember that 9 games played does not begin to represent a full season. A handful of games below .900 is normal for any season from the very good goalie.