Stimson: How passing relates to shooting percentage and close situations

Ryan Stimson has been tracking passes for HP and In Lou We Trust. Follow him on Twitter @RK_stimp

Last time, I took a look at the relationship we’re seeing develop between a team’s shot generation efficiency (proportion of shot attempts generated from passes that results in shots on goal, SAGE) and the number of goals they score during 5v5 play. One of the reasons that relationship exists is because teams are more likely to score a goal from a shot that is the result of a pass versus one that is created in another fashion.

Below is a chart showing shooting percentage data for the New Jersey Devils, the New York Rangers, the New York Islanders, the Florida Panthers, and the Chicago Blackhawks. The blue line represents their 5v5 shooting percentage taken from The orange line above that represents the shooting percentage generated from completed passes. Again, this is early in the season, but for every team we’re tracking data on, teams are consistently shooting at a higher percentage from successful primary passes.

(click to enlarge)

SH graph

If this pattern holds true, then it stands to reason the generating shots from completed passes will be of vital importance going forward. Teams that may track data similar to this in-house may be more inclined to find a bottom-six forward that is a decent distributor of the puck and can contribute through the passing game rather than just focus on a “lunch pail” guy to dump the puck in and not make mistakes.

There is a second reason why SAGE may correlate strongly with the number of goals a team scores: how often teams win. I’m keeping track of every passing metric we have and how often a team wins that category as well as how often they win each game. In the below chart, you will see the eight metrics most closely associated with teams winning through 100 games decided in regulation. The ninth metric is the most successful shot-based metric: which team controls shot possession during 5v5 close score situations.

Now, recent (and excellent) work done by Micah McCurdy (@IneffectiveMath on Twitter), revealed that there may be little value in “Close” situation metrics as we have come to know and use them (within a goal in the first and second periods and tied in the third). Again, it is still early in our research, but we’re finding a contrast between passing metrics in close versus all 5v5 situations.

 Win percentage

The two metrics most closely associated with teams winning are the same: SAGE during all 5v5 situations and SAGE during close situations. The following metrics read as follows: OZ SG Poss. Close, which team has the greater share of shots generated from passes in the offensive zone minus scoring chances in close situations; SG Poss. Close, which team has the greater share of shots generated in total during close situations; SC SAGE Close, which teams are most efficient at generating scoring chances during close situations; A2 SAGE, which teams are most efficient at generating shots from two or more consecutive passes during close situations; SC SG Poss. Close, which team has the greater share of scoring chance shots generated in close situations; and, finally, A2 SG Poss. Close, which team has the greater number of shots generated from two or more consecutive passes.

These were our only metrics that correlated with teams winning games above 60%. I’ll update as we increase our sample size every fifty to one hundred games, but the more these patterns start to emerge, the more we should want to collect as much data as possible league-wide. Once we have an online database we will begin recording the time of each event, so that passing metrics can be sorted by various score situations and delved into even further.

4 thoughts on “Stimson: How passing relates to shooting percentage and close situations

  1. Hey Ryan, I think it would be better to look at 5 on 5 goal differential instead of winning percentage. Winning can also come as a result of special teams success, or luck in the shootout/overtime.

    Just some food for thought.

    • That’s probably right. I’ll probably compare the passing metrics to goals, winning, shots, Corsi, etc. just to get the full picture. But, goals scored and goal differential are probably right.

  2. Pingback: Estimating Standard Deviation in On-ice Shooting Percentage Talent -

  3. Pingback: Possession, Shooting Percentage and Outlier Teams -

Leave a Reply

Your email address will not be published. Required fields are marked *