Shots versus shot attempts

A puck can go into the net, be stopped by the goalie, miss the net or be blocked before it gets there. Shot charts can count different stages of that activity. Read the full category before shortening it to “shots.”

For this lesson, shots on goal include goals and saved shots. All attempts add missed and blocked attempts. Unblocked attempts add misses to shots on goal but leave out blocks.

These are nested counts, not separate piles you can freely add. A shots-on-goal total already contains goals; adding the goal count again would count those attempts twice.

Some reports call all attempts Corsi and unblocked attempts Fenwick. The labels can be useful, but the plain definition still matters: which event categories entered the count, and which game situation was selected?

Our invented Team A has 20 shots on goal, 12 misses and 18 blocked attempts. Its total is 50 attempts; its unblocked count is 32. Team B has 25 on goal, ten misses and five blocks: 40 attempts and 35 unblocked.

A leads total attempts 50 to 40. B leads shots on goal 25 to 20 and unblocked attempts 35 to 32. The ordering changes because you are selecting different categories from the same offensive activity.

A’s larger block count contributes much of its total-attempt lead. That is useful composition to show, but it does not establish why those attempts were blocked or whether every other chance was equally dangerous.

If A and B are the two sides of this same invented game state, A’s attempt share is 50 ÷ 90 = 55.6%. Its shots-on-goal share is 20 ÷ 45 = 44.4%. The same team leads one share and trails the other.

Neither percentage is broken. One compares all included attempts; the other compares only shots on goal. Keep those names beside the numbers so the reader can explain the disagreement.

For real data, match time and manpower on both sides. Five-on-five attempts for and all-situations attempts against cannot be treated as one coherent share. Compatible scope comes before the division.

A blocked long-range attempt and an unblocked chance near the net are different opportunities. A raw attempt count does not assign a scoring likelihood to each one. Expected-goal models add context under their own documented rules.

If a team leads attempts but trails a chance-quality measure, inspect the composition rather than choosing the row that matches your first impression. The different views can explain why the offense looked busy without creating the stronger modeled chances.

A helpful note for our game is: “A attempted more shots overall, with many more blocks; B led both unblocked attempts and shots on goal.” The labels make the mechanism readable without pretending one count explains the whole game.