Skip to content

Expected points per shot

also: xPts · shot quality

The points a shot is worth on average, from where it was taken and what kind it was, fitted on league history.

Formula
expected points = P(make | location, shot type) x shot value

Each field goal attempt in the play-by-play carries an x/y location. We fit the make rate for that location and shot type on league history, then multiply by the shot's value, two or three. The result is what an average shooter would be expected to score from that spot.

Summed over a player's or team's shots it measures shot selection. Summed against actual points it measures shot making. Those are different skills, and one number each is how you separate them.

The common mistake

Expected points describe an average shooter. A great shooter beats it on purpose, so a positive gap is not always luck, and a negative one is not always a bad shot.

The data behind it

Expected points per shot comes from NBA play-by-play data and NBA shot chart data. Play-by-play for every NBA season since 2003-04: clock, score, players, shot type and location, and win probability from 2018. REST or MCP.

Where to get it

Related

Expected points per shot from the API

500 free credits every month, no card. REST or MCP.

curl -X POST https://mcp.dunkapi.com/v1/get_plays \
  -H "Content-Type: application/json" \
  -H "x-api-key: YOUR_API_KEY" \
  -d '{
  "game_id": "401585601",
  "shots_only": true,
  "limit": 3
}'

get_plays: every parameter and the full response.

Also from DunkAPI: player props history by sportsbook · today's props board · every dataset

Get a free key