clock menu more-arrow no yes

Filed under:

Game theory: Rethinking the shootout, and the value of a one-trick pony

New, comments

The surprising lesson an ex-Hab can teach us about excellence

Reinhold Matay-USA TODAY Sports

When you look at the best guys in the league at the shootout, they all have three or four really good moves, but they all start out exactly the same way. Look at T.J. Oshie. He’ll pick the puck up and take a similar route toward the net every time. He’ll stick handle it in a similar way every time. And then when he gets to a certain spot, he has four moves that branch off from there. You don’t really know what he’s going to do until he does it. So it forces the goalie to make a reactionary save, which is much more difficult. - Jonathan Quick (via Players' Tribune)

The passage above was published earlier this week on The Players' Tribune. I thought Jonathan Quick and the ghostwriter the site employed did an excellent job breaking down some of the NHL's best scorers, but I don't totally agree with the Kings' goaltender on one point, and I'll show you why shortly.

Why a one-trick pony is a good thing to be

I fear not the man who has practiced 10,000 kicks once, but the man who has practiced one kick 10,000 times. - Bruce Lee

Since 2009, T.J. Oshie has scored on a remarkable 54.7% of his shootout attempts (29/53), so Quick is correct in saying the American is pretty tough to deal with one-on-one.

Ask a knowledgeable NHL fan to name the other best shootout scorers in the league, and s/he might bring up Pavel Datsyuk, Patrick Kane or Jussi Jokinen. But if we let the results do the talking, we get a somewhat different picture.

Select Players (Since 2009) SO Attempts Goal%
P.A. Parenteau 31 48.4%
David Desharnais 37 45.9%
Radim Vrbata 67 43.3%
Pavel Datsyuk 55 38.2%
Patrick Kane 69 37.7%
Jussi Jokinen 47 27.7%
Alex Semin 26 26.9%

Strange, isn't it?

It gets stranger still when you look at how he goes about scoring those goals. You can do a Youtube search for "Parenteau Shootout Goal," but let me save you a little time with this collage:

These are five goals, scored in a five-year span, off the exact same move, which goes like this:

Datsyuk has three solid breakaway dekes, plus a highlight-reel-inducing change-up. Patrick Kane is one of the most talented stickhandlers in the history of hockey. But Parenteau has consistently out-performed both of those Stanley Cup-winning All Stars by having just one scoring move. He'll never be T.J. Oshie, but he has been getting almost equivalent results despite having far fewer tools to work with.

It defies logic.

Or maybe not.

From skills comp to flipping coins

The idea for this investigation actually stemmed from a conversation I had with a former NHL goalie earlier this year. Originally, I had asked this player whether he used video as a learning tool for improving his technique. He answered no, but then told me that late in his career, he began using video extensively to study opposing skaters in the shootout.

At one point, he told me he was able to know exactly a shooter was going to go with his shot, even before he crossed the offensive blueline. And that's because while players try to "mix things up" in order to maximize their chances for success, the stress of the moment actually causes them to default to their subconscious patterns. So if you come in wide and fast, you might instinctively shoot far side. Down the middle and slow, that's a backhand deke. And that's how this goalie was able to out-smart plenty of great scorers (Henrik Zetterberg was one notable example he mentioned).

About two-thirds of shootout attempts are stopped, and it would seem that a smart, studious goaltender would have even better odds against a shooter flying by the seat of his pants.

But there's something the shooter can do to shift a 33/66 losing bet, into an even 50/50 chance at a goal: tell the goalie exactly what he's going to do beforehand.

This sounds patently crazy. But it's true. We only need to look a bit further down the list I came up with earlier to see the best example of a One-Trick Shooter:  Radim Vrbata.

Like Parenteau, Vrbata has shot at over 40% in the past 6 years by having one basic move. But unlike Parenteau, who takes about six shootout attempts per year, Vrbata has averaged 11 per season since 2009 without losing in efficiency.

Here he is making former Habs goalie Peter Budaj look bad in 2007:

And here he is again with the exact same run-up and deke four years later, in 2011.

Now, Vrbata was one of the only snipers on the Coyotes; a lock to be one of Dave Tippett's three shootout choices. So opposing goalies would have had plenty of opportunities to study video and identify Vrbata's tendency. It's just common sense.

Just to show how thick you would have to be to miss it, here's another goal where Vrbata goes forehand-backhand-back of the net (2009):

So if you were a goalie facing the Czech in a shootout, you'd be pretty sure that he is going with his bread-and-butter. So you cheat to take away his best shot, and camp out deep in your net on the backhand side. And that would be a big mistake.

Going back to the idea that 2/3 of shootout attempts are stopped. If you are a goalie, the odds are going to be naturally in your favor. If you square up on the top of your crease and play the shooter straight-up, he could shoot the puck right on you, miss wide or over the net, or lose the puck on a deke.

But if you already know what to expect and act on that information, then you've effectively stopped playing the game - you're now playing game theory. You guess right and it's an easy save; you guess wrong and it's a goal.

It's now a 50/50 coin-flip, and that's how guys like Parenteau, Vrbata and Brad Boyes (the most underrated player in the history of the shootout) end up doing so much better than the Kanes and Datsyuks of the world. They know they won't score every time. They won't even score most of the time. But by letting the prepared goalie psych himself out one out of ten times, they can open up a gap that can be sustained for many years.

And here we get to the crux of the issue: Jonathan Quick believes making a reaction save is tough, and he's totally right, but what seems to be even tougher for an NHL goalie is trying to out-guess his opponent one way or the other, instead of relying on his years of reactionary practice.

I'll leave you with this 2013 clip of Vrbata with a change-up. Roberto Luongo never knew what hit him.

Jack Han is the Video & Analytics Coordinator for the McGill Martlet Hockey team. He also writes occasionally about the NHL for Habs Eyes on the Prize. You can find him on Twitter or on the ice at McConnell Arena.