Running Net Points — a plainer version
Imagine you sit courtside with a clipboard and credit each player for what they did on every possession. Add up the credits by the end of the night. That's RNP.
Credit: Justin Jacobs (Squared2020)
The math → · Production methodology (pipeline, M1 vs M2, Produced/Allowed) →
The scene
Picture yourself at a basketball game with a clipboard. Not scoring the game — that's the scoreboard's job. You're keeping a different ledger: for every possession that happens, you write down who helped their team and who hurt their team, and by how much.
A point guard pulls up from thirty and swishes it: you write +3 next to his name. His defender, caught under a screen, gets a small negative number — not a whole -3, because five guys were on defense and the breakdown was shared, but some fraction of it.
A missed shot, an offensive rebound, a kick-out to a shooter who drains it: the rebounder earns a partial credit for extending the possession, the passer earns a bigger credit for setting up the easy look, the shooter earns the biggest credit for actually scoring. The four defenders who were on the floor absorb a small share of the points allowed. A bench player chatting with the coach gets nothing — he wasn't part of the possession.
Every possession is a little envelope of credit, opened and split among the ten players on the floor.
You do this for every possession of the game — roughly a hundred of them per team. You add up each player's running total. What you've got on your clipboard at the buzzer is that player's Running Net Points.
Why do it this way
The box score tells you what happened, but not who made it happen. Twenty-two points looks productive whether those points came easy (wide-open transition layups against a scrambled defense) or hard (contested pull-ups late in the shot clock against a set scheme). The box score doesn't know the difference. Assists don't either: a pass to a wide-open corner three counts the same as a pass to a contested mid-range jumper.
Plus-minus — the stat on most broadcasts that says +12 or -8 when a player leaves the game — tries to fix this by watching the score while a player is on the floor. But plus-minus lumps the whole group together. A player surrounded by four great teammates can look great by plus-minus while doing almost nothing. A player dragging a tired bench unit can look terrible by plus-minus even when he's the only reason the margin isn't worse.
RNP is stricter. It asks, possession by possession, what did this specific player do? Not what the lineup did. Not what the team did in aggregate. What the individual did on that one play.
A three-possession example
Say your player starts the night with RNP at zero.
Possession 1. He drains a pull-up three — unassisted. Big credit to him on offense: roughly +3. On the other end, the opponent runs a pick-and-roll that ends in a layup, and your player was the ball handler's primary defender. Partial defensive hit: roughly -0.6. Net for the possession: about +2.4.
Possession 2. He sets an off-ball screen, doesn't touch the ball, and watches a teammate score. No offensive credit — he wasn't involved. On defense, he takes a charge on a baseline drive. The charge counts as a defensive stop, so despite the opponent attempting points, the possession ends with a positive defensive adjustment. Net: roughly +0.4. Running total: +2.8.
Possession 3. He swings a pass to a wide-open corner shooter who buries a three. Assist credit scaled to the value of the three flows to him. Defense: ball went away from him, no rotational responsibility, no adjustment. Net: roughly +1.1. Running total: +3.9 through three possessions.
That's the whole idea. Watch any possession in any game and ask, what did this player contribute to what happened? Sum those answers across the night.
How big a number is "good"
This surprises people. RNP is a net measure, not a counting stat — the negative possessions eat into the positives. Even a good night usually lands between 5 and 12, not a 30-point explosion.
- Less than 0 — net negative; the team was probably worse with this player involved
- 0 to 5 — an ordinary starter's night
- 5 to 10 — above-average; the player made more possessions better than worse
- 10 to 15 — a strong night
- Above 15 — exceptional, usually reserved for true go-to performances
A full season average of 6 or 7 puts a player in genuine All-Star territory. 8 and above is elite. These are not huge numbers; that's the whole point. RNP is miserly with credit by design.
What it isn't
RNP is not a complete measure of basketball value. No single number is. Defensive involvement is inferred from play-by-play tags — the system doesn't see where a player's feet were at the shot release, so small differences between two good defenders may be noise rather than signal. The stat doesn't adjust for opponent strength: 12 RNP against the league's best defense means more than 12 RNP against the worst. And single-game totals are noisier than season aggregates — one clipboard full of a hundred possessions has a lot of randomness in it. Trust the season more than the night.
But it is a more honest attempt to isolate individual impact than anything the box score offers on its own. Who helped their team tonight, possession by possession, shorn of their teammates' brilliance or incompetence. That's what RNP gives you.