3PT Shooter Analysis

How many threes should an NBA team take, how accurate must a shooter be, and does being "cold" matter? NBA.com data, 2014-15 to 2025-26.

Open the shot explorer View the code and write-up

The shot explorer is a Buckets-style player view: court map of where a player shoots and how well compared with the league, shot frequency and FG% by distance, left vs right splits, home/away, win/loss and opponent filters, and two-player comparison. Pick any player-season since 2014-15.

Shot explorer screenshot — James Harden, 2015-16.

Findings

Break-even 3P% vs league 3P% — The average three has caught up to the average two: break-even rose from 32.3% to 36.6% while league 3P% stayed near 35 to 36.6%.

Adjusted 3PA curve — No clear optimal number of team threes: the adjusted curve is flat or mildly convex, never a hill.

Corner vs above the break — Corner threes are about 3 points better than above-the-break threes.

Cold streak result — Cold streaks do not predict the next three, and there is no hot hand.

Full explanation of data, methods and conclusions: see PROJECT_OVERVIEW.md in the repository.