Search for basketball lineup tracking and you'll land on stat apps that proudly display plus-minus for every five-man combination you've played. That number looks authoritative — it's right there, calculated to the point. What none of those apps explain is how little that number means on its own, or what you're supposed to do with it once you have it. Ranking lineups by raw plus-minus and trusting the top of the list is how a coach ends up making rotation decisions off noise instead of signal.
Your five best individual players rarely add up to your best five
It feels intuitive to start your most talented players and assume the combination takes care of itself. In practice, two ball-dominant guards who both need the ball in pick-and-roll, or two bigs who both want to operate in the same five feet of the paint, can post a worse point differential together than a lineup with less individual talent but better fit. You can't see that from a stat sheet of individual averages — it only shows up when you track how groups of players perform together, not just how each of them performs alone.
Why a few games of lineup data can lie to you
A five-man unit that's only shared the floor for eleven minutes across two games can show a wild plus-minus in either direction, and it's almost always more about who was on the floor for the other team during that stretch than about your lineup itself. A bench unit that happened to play against the other team's bench will look great. A lineup that got stuck on the floor for a 10-0 run to end a half will look terrible, regardless of how it actually played. Treating either as proof of anything is the fastest way to misread your own team. The fix isn't to ignore the data — it's to wait for enough minutes to accumulate before you act on it, and to look at the context behind the number, not just the number.
What to log alongside the plus-minus, not instead of it
- Minutes played together. A lineup with a big plus-minus over three minutes tells you almost nothing. The same number over forty minutes across a season is worth acting on. Always read the differential next to the sample size that produced it.
- What the opponent was doing on the floor at the same time. A lineup that only ever faced the other team's starters looks worse by default than one that mostly saw bench minutes. Without noting who they played against, you'll credit or blame your own players for matchups that had nothing to do with them.
- The game situation, not just the final number. A lineup that outscored an opponent while protecting a 15-point lead in garbage time isn't the same as one that did it in a tied fourth quarter. Tag which minutes were competitive and which weren't before you compare lineups to each other.
- Why you made the substitution that created the lineup. If a group only existed because of a quick foul-trouble swap, its numbers reflect an emergency lineup, not a rotation you'd choose on purpose. That context matters as much as the result.
Use it to test a hunch, not to replace your eyes
The most useful way to apply lineup data isn't scanning a leaderboard for your best combination — it's checking a specific hunch you already have. If you suspect two guards struggle to share the floor, pull up every stretch they've played together and see if the numbers back up what you're seeing in games. If they do, you've got more than a feeling to bring to a lineup decision. If they don't, that's worth knowing too, before you bench someone over an impression that the data doesn't actually support.
This only works if your lineup data lives next to the rest of what you track on a game — the substitutions that created each group, the opponent, and the score context — instead of sitting in a separate app that only outputs a number. That's what HoopVault's game stats are built for: logging who was on the floor together as part of the same box score you're already keeping, so a lineup's plus-minus comes with the context that makes it mean something. Track lineup combinations alongside your game stats in HoopVault — free to browse, no card required.