Statistically Analyzing Hockey's One-Timer

Through my work so far at this site, I have tried to paint something of a picture of the foundational pillars of power play success, though there are many more to come. I have so far talked about the emergence of the 1-3-1 formation and its ability to create passing triangles. I have analyzed the importance of placing players on their off-wings to generate maximum shooting efficiency. The third pillar, something that has catalyzed the Washington Capitals rise to the top of the league, has been their use of the one-timer. Fourteen of the team’s 35 non-rebound 5-on-4 goals this season have come off one-timers, and everybody knows about the Ovechkin spot. So what about the one-timer? It is a shot that hasn’t historically been analyzed because it is not classified as its own type of shot in the NHL’s play-by-play files. Fortunately, Ryan Stimson and his team have been tracking one-timers for the last couple of years in their passing files. With my project, I have also been tracking them as a separate category, so that I have a couple of different datasets to analyze.

Steven Stamkos last Wednesday scored on a well-executed one-timer against Jonathan Bernier. Even though he's quite a distance from the net, his shot speed and placement, as well as the quick pass from Stralman and subsequent change in angle, led to the result.

But first, conceptually, why are one-timers so dangerous? Well goaltenders have become so sound positionally in this day and age — with most well over six feet tall and able to cover the majority of the net — that making them move has become a key to goal scoring. One-timers accomplish that feat to the extreme, as goalies are forced to move laterally without the ability to reset and square up before the puck is shot. If a one-timer is well executed, the player should essentially be shooting at half or more of an empty net. But does the data suggest that one-timers are as effective as we might think?

The first dataset to analyze is Stimson’s passing data from last season. Unfortunately, power play data wasn’t kept, so this is all at 5-on-5. The same principles should hold, though, and the sample size is luckily decently large.

There’s a lot to unpack here, so let’s dive in. I compared one-timers to all other manner of shot. It is important to note that for Stimson’s data, pass-tips were classified as tips and not one-timers, so it is my belief that this leads to a lower one-timer shooting percentage than one would otherwise observe. But still, one-timers overall appear to be the superior shot. The overall one-timer versus non one-timer result is statistically significant.

Why is the split not more extreme, if the ultimate goal is to get goalies moving? Well one-timers can be very difficult to pull off. Just from watching power plays, I can tell you that Alex Ovechkin is the exception rather than the rule when it comes to success on them. Taking a hard pass and timing a shot perfectly, when the puck may be rolling or on end, while also aiming at a small net, particularly from far distances, can be very difficult. That is why, as you can see, the on-net percentage for one-timers is lower than other shots. Generally, if you get a one-timer on net, there’s a good chance you can catch a goalie out of position or find a seam. But you’re more likely to miss the net entirely. Fans should show some patience when it comes to players on their teams who can’t seem to replicate Ovechkin’s success from past years.

Next, it’s clear that what one looks for as the ideal shot is a scoring chance one-timer that crosses the Royal Road (an imaginary line from the center of the goal out to the top of the faceoff circles/ringette line). In other words, you want to get the goalie moving laterally as much as possible, while also shooting from close in — giving yourself a larger net to shoot at, and allowing the goalie less time to recover.

Perimeter one-timers however appear to be less effective than other kinds of shots from the outside. This makes sense considering how tough it can be to get a one-timer on net, and how small the net appears from the blue-line; there isn’t much room for error on a shot like that. Without building a distance model, it's tough to say at what distance or angle from the net one should simply not attempt a one-timer, and instead cradle the puck and assess, but that model will come. The dataset simply isn't large enough yet.

The table above, as I mentioned, features 5-on-5 tracked data from last season, which grants a decent sample size for binary analysis. Unfortunately, our 5-on-4 data is limited, since it comes only from this season, so I will update it as both the Passing Project and Special Teams Projects progress. But I still took a peak at what Stimson and his team’s data suggests in terms of one-timer effectiveness on the power play, keeping in mind that he counts pass-tips as non one-timers.

The 530 one-timers that they have tracked this season on the power play haven’t painted a pretty picture, at 3.8 percent. The reason, though, why I point out about the pass-tips, is because they are the highest percentage one-timers. By looking at my own data, I isolated for the difference, and you can see for yourself.

I also was curious about the difference between one-timers for teams set up in formation and teams either on the rush or playing pucks off the boards and the like (Note: The table says 5v5 but of course it's 5v4). I will detail more about my formation tracking method in a future article, but these are approximations, so take them with a grain of salt. It is interesting that, to this point, one-timers taken out of formation are the most successful.

Even though this goal was scored shorthanded, one can see why one-timers on the rush have hte potential to go in at a higher rate than when defenders are blocking lanes.

If I combine the Passing Project data with my own (I didn’t bother worrying about the few potential duplicates), the differences between one-timer and non one-timer shots on the power play appears to disappear, with both hitting the back of the net about six percent of the time. 

Since I know I’m going to get the question, yes, the Capitals are affecting this data because of their prowess. This year, the Caps at 5-on-4 have scored on 9.0 percent of their one-timers, compared to 6.7 percent for the other five teams I track. On non one-timers, the Caps are at 7.0 percent, with the rest at 5.1 percent. Ovechkin, though, is not the driver behind that high percentage, as he has only scored on two of 96 one-timer attempts this year, for a lowly 2.1 percent. The rest of the team is shooting at 13.5 percent on their attempts, with Oshie and Williams, the two slot men with the best shooting angle, leading the way with four such goals each. So this isn't a one-man spike, the Capitals are generating high percentage one-timers, something other teams should be able to figure out at least to an extent. Not every team has a Nicklas Backstrom, but passes like the one featured in the picture below don't require superstar-level vision or touch, just the right setup, quick movement and passing with intention.

T.J. Oshie has been a huge weapon for the Capitals this year, a team that has always looked to feed its slot-man for one-timers as a great second option after Ovechkin.

So what are my takeaways from this data? One-timers are the most dangerous non-tip shots in the NHL, at least until I see a bigger sample of data disproving that. It appears, though, that even more so than feeding a Steven Stamkos for the bomb, finding a one-timer on the rush, and not settling there for an average wrist shot, can make a huge difference. As can feeding your team’s bumper for one-timers in the slot. And finally, I encourage teams to practice one-timers more, and for players to take it upon themselves to spend time perfecting them. It’s a difficult thing to do, because you need a passer with the patience to feed you hundreds of passes in their own time — or have a passing machine that could do it — but it’s worth it. Improving those percentages, being able to hit the largely empty nets, can make all the difference.