Sports Stack Exchange is a question and answer site for participants in team and individual sport activities. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

It seems like a lot of shots hit the crossbar in hockey. Is this just my mind playing tricks on me, or do more shots hit the crossbar than might be expected for a bar a few inches wide set above a goalie's shoulders.

First, what percent of shots hit the crossbar in the NHL?

And is it higher than would be expected?

The "expected" rate, I'd imagine, is the same rate as any other space on goal of equal size to the crossbar.

share|improve this question
1  
what would be expected? – Nicholas V. May 5 '14 at 12:15
    
@NicholasV. Good question. I added a little clarification. – samthebrand May 5 '14 at 14:24
6  
@SamtheBrand I cannot find anything out there officially on this topic. The tough thing is that a shot off the post or the cross bar in hockey is not technically a shot on goal. There isn't an official stat that keeps track, so aside from an independent study I think it would be tough to find/calculate. Here is an article that talks about it. – diggers3 May 5 '14 at 18:20
    
I suspect that more shots are directed at the top part of the net by shooters since most goalies use the butterfly technique. This technique (or style) blocks the bottom part of the net. So two things: 1) goalies make more saves on low shots thus preventing the puck from hitting the lower parts of the posts; and 2) shooters know that the upper portion of the net is more open so they choose to shoot high more frequently. This is just a hypothesis. Without running an independent study as diggers3 suggests, it's not possible to determine an answer. – John Cullen May 7 '14 at 17:21
    
@JohnCullen The easiest places to score are usually low on the stick hand side and high opposite the stick hand; up the middle either high or low and high on stick hand side are medium level; and low on blocker hand is hardest to score (as the stick covers the low on opposite side normally, so the least movement is needed to block a shot there). Butterfly technique still gives up a fairly high number of five-hole goals if you don't time it perfectly (as if you go to early, you end up giving up a high goal). – Joe Oct 15 '14 at 21:55

The NHL tracks missed shots that hit the crossbar. http://www.nhl.com/stats/team?reportType=season&report=shottype&season=20152016&gameType=2&sort=missedShotsHitCrossbar&aggregate=0

What percent of shots hit the crossbar in the NHL?

In the 2015-2016 NHL regular season, 240 missed shots hit the crossbar out of 28,517 missed shots. That is 0.842%.

And is it higher than would be expected?

According to your crude measure of "it seems like a lot of shots hit the crossbar in hockey," then it is much less than expected. Also, these findings are for missed shots. Perhaps there is a percentage of shots that hit the crossbar and went in for a goal, but that is the "missing" subset.


There aren't any stats that I know of that track shots that enter "any other space on goal of equal size to the crossbar." However, let's try:

  • There were 6,565 goals during the 2015-2016 NHL regular season.
  • A NHL regulation goal opening is 72" wide and 48" high(1)
  • A NHL regulation crossbar is 2 3/8" in circumference(1)
  • Dividing 48" by 2 3/8", we get ~20 (more precisely 20.211) such segments that fits the criteria of "any other space on goal of equal size to the crossbar."
  • Dividing 6,565 by 20 (assuming an equalized probability), the number of goals that would enter "any [segment] (out of 20) on goal of equal size to the crossbar" is 328.25 (more precisely 324.8).

Thus, 240 missed shots due to the crossbar compared to 324.8 goals on any given segment of equal size to the crossbar means the number of missed shots is lower than expected (the same rate as any other space on goal of equal size to the crossbar).

share|improve this answer

Shots that hit either the post or the crossbar are not considered Shots on Goal. Currently, they are quantified as a Missed Shot. These shots are included with shots that miss the net entirely; as such, they will appear in both Fenwick and Corsi totals (note that Corsi is just Fenwick plus blocked shots).

There's no way to extract these data currently.

However, we can estimate the relative frequency with which shots hit the crossbar versus shots that hit either post.

The net is 6 feet wide by 4 feet high. Since there are two posts, this means that (roughly) the crossbar is 6 linear feet and the posts are 8 linear feet.

Naively, then, one would expect that the crossbar is hit at a 3:4 ratio to the posts. However, while shooting, one typically picks one side of the net to shoot at. If you miss your shot while trying to pick a side, you're probably not going to hit the post on the other side of the goal, unless you're a very bad shooter. So really, the crossbar might get hit at a rate of about 3:2 compared to the post, or about 50% more shots hit the crossbar than hit the post.

This is just a naive approximation.

Practically speaking, shooters have very small windows with which to work. Goalies cover a lot of area quickly in today's NHL, and so shooters are often gunning for extremes. The posts provide a visual reference of your target. When the goalie is moving and changing angles, the posts provide a fixed frame of reference with strong visual contrast. Many shooters try to aim for the inside of the posts with their shots, because it's usually the farthest you can target from the goalie and still hope to score.

Probably more shots miss the net entirely than hit a post, and probably close to as many shots that hit the crossbar and go out (quantified as a miss) hit the crossbar and go in (quantified as a goal and a SOG).

When weighing this approximation against all total shots, I'd say hitting the crossbar is more rare than hitting any other similarly-sized window of the net, but without very accurate puck telemetry data, it's impossible to conclude that.

share|improve this answer
    
I think this is a good comment, but doesn't answer the question with anything other than extremely vague speculation. – Joe Oct 15 '14 at 21:50
    
@Joe Nested in my answer is the core of a Bayesian estimation (without explicitly going into the details). Since the data aren't available, this is literally the most accurate possible answer to this question. Since the OP's question involves a Selection Bias, a back-of-the-envelope estimation is often sufficient to negate that bias. For instance, suppose you buy a red Toyota, and you start seeing more red Toyotas. You don't need to have access to Toyota's production records to make an estimation to prove that you're experiencing selection bias. – Arkamis Oct 15 '14 at 23:20
    
And, all other commentary notwithstanding, I provided a correct answer to the question in my third sentence. Everything below that is there simply to address the perceived bias. – Arkamis Oct 15 '14 at 23:23
    
I'm well familiar with Bayesian estimations, but I don't agree that this is a good answer to the question. The question doesn't have a good answer (unless the statistics are out there, but I don't see them either), which to me means it shouldn't be answered. – Joe Oct 16 '14 at 2:24
    
@joe The answer "the information you're looking for is not available" is a perfectly reasonable answer on every other StackExchange site, when it is the only correct answer. If you'd like to provide a better answer to this five month old question, please feel free to do so, otherwise what I've offered is reflective of the current state of the art of hockey analytics. If you don't like the reality that the data don't exist, then vote to close the question rather than let it persist in perpetuity. – Arkamis Oct 16 '14 at 2:30

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.