6:41 PM. A beat writer posts the starting lineup. The starting center is out — back spasms, nothing announced earlier.
His backup is $3,700. He's averaged 15 minutes a night all season and he's about to play 33.
By 6:44 he's in forty percent of the lineups being built. By 7:00 he's the highest-owned player on the slate. And he's still correct.
That's NBA DFS in a paragraph. Not a subtle analytical edge — a structural one, sitting in the open, available to anybody who's looking at the right screen at the right time.
It exists because of a simple mismatch. DraftKings sets salaries in advance, based on what a player's role was. The role changes an hour before tip. The salary doesn't move.
WHY MINUTES
PREDICT EVERYTHING
A given player, in a given role, produces at a reasonably stable rate per minute. His shooting percentage bounces around, his rebound totals bounce around, but across a full workload those swings mostly wash out. What doesn't wash out is how long he was on the floor.
So fantasy output is close to linear in playing time. Double the minutes and you roughly double the expected score.
Efficiency is noise. Opportunity is signal.
Nothing in baseball or football is this reliable. A hitter gets four plate appearances and can do nothing with all of them through no fault of his own. A receiver runs a perfect route eighteen times and catches three passes because the quarterback looked elsewhere. Those sports hand you opportunity and then let randomness eat it.
Basketball hands you 33 minutes and 33 minutes is 33 minutes. The randomness sits on top of a much larger predictable base. Which is why the single most useful question you can ask about an NBA player is not "is he good" but "how long is he playing tonight."
SALARIES ARE SET
ON OLD INFORMATION
DraftKings prices a slate before the news that determines the slate. That's not a flaw in their process; it's unavoidable. Contests have to be live before lineups are confirmed.
The consequence is that on most nights, at least one and often several players are priced as bench players while playing starter minutes. They're not underrated. They're not sleepers. They're just priced against a role that no longer exists.
Finding them isn't analysis. It's attention. The analysis comes next, in figuring out which ones are actually worth what the news implies.
MINUTES AND USAGE
ARE DIFFERENT THINGS
Here's where most people leave money on the table, and it's the one genuinely non-obvious idea in this article.
When a starter sits, two separate things get redistributed, and they don't go to the same place.
The minutes usually go to the direct backup at that position. That's a rotation decision, it's mechanical, and it's easy to predict.
The usage — the shots, the possessions that end with him — usually doesn't. Offensive role doesn't follow position; it follows hierarchy. When the team's second-leading scorer sits, his shots don't go to the guy who took his roster spot in the starting five. They go to whoever is already the next creator in line, who might be a guard when the injured player was a forward.
WHO GETS WHAT
Minutes only. A defensive specialist steps into the starting lineup, plays 30 minutes, takes six shots, and scores 14 DraftKings points. He was $3,400 and he returned fine value, but he didn't win anybody a tournament. This is the most common outcome and it's the one people overpay in ownership for.
Usage only. A teammate's role expands without his minutes changing much. He was already playing 32; now he's taking 19 shots instead of 14. Often the better play, and almost always the cheaper one in ownership terms, because there's no obvious announcement attached to him.
Both. This is the slate-breaker. A player who steps into the starting lineup and inherits offensive responsibility. Thirty-four minutes at a usage rate he's never had before, priced at $4,000. These are the games where somebody returns six times value and wins a Milly Maker.
Practical test before you roster the obvious beneficiary: when this player has previously played 30+ minutes, what did he actually do with them? If the answer is 22 points and 9 rebounds, you have a real play. If the answer is 8 points on 5 shots, you have a body in a jersey playing 30 minutes, and the field is about to make him 40% owned anyway.
THE SIX WAYS
MINUTES MOVE
THE INJURY SCRATCH
The classic. Announced anywhere from the afternoon injury report to fifteen minutes before tip. The later it lands, the more valuable it is, because fewer entrants have absorbed it and ownership hasn't caught up.
REST AND LOAD MANAGEMENT
Often a healthy player sitting a back-to-back, especially the road half. Frequently announced late, sometimes only when the starting lineup posts. A team playing its second game in two nights is always worth a look regardless of what the injury report says.
THE ROTATION CHANGE
The quiet one. No injury, no announcement — a coach simply decides someone has earned more run, and it shows up over two or three games rather than overnight. There's no news alert for this, which means salaries lag it for a week or more. Checking recent minutes trends rather than season averages is how you find them, and it's the least crowded edge in the sport.
BLOWOUTS AND GARBAGE TIME
Predictable in aggregate from the spread even though you can't know which games break open. Deep bench players on heavy favorites carry a real tail: 12 minutes of fourth-quarter run at minimum salary is a fine outcome for a punt slot.
FOUL TROUBLE
Pure noise, and it cuts both ways. You can't predict it, but you should know it's a risk on any center you're paying up for, and it's the reason a foul-prone big is a worse Captain than his projection suggests.
MARCH AND APRIL
The end of the season is the best stretch of the year for this entire approach. Teams out of the race rest starters and hand 35 minutes to players who barely played in November. Teams locked into seeding do the same. Salaries chase these role changes all the way to the finish and never quite catch up.
THE VALUE PLAY
IS ALSO THE CHALK
Everyone reads the same injury report. So the obvious beneficiary of the biggest news of the night is frequently the highest-owned player on the slate, sometimes north of 50%.
That doesn't make him wrong. It makes him a different kind of decision depending on what you're playing.
In cash games, take the chalk. He's correct, and the salary he frees up is what makes the rest of your roster affordable. Being clever in a 50/50 is how you lose a 50/50. Our beginner's guide covers why cash construction and tournament construction are genuinely different problems.
In tournaments, look one step past him. If he's on half the field, rostering him gains you nothing relative to the people you're trying to beat — you need him and something they don't have. That something is usually the second-order beneficiary: the teammate whose usage quietly expanded without a headline attached, at a fraction of the ownership. The full framework for that trade is in our leverage and ownership guide.
The useful mental model: news creates value, but news also creates ownership, and those arrive at different speeds. The gap between them — the twenty minutes where the news is out and the field hasn't fully reacted — is worth more than the value itself.
THE NEWS DOESN'T STOP
WHEN THE SLATE LOCKS
NBA tip-offs are staggered from 7:00 through 10:30 Eastern, and DraftKings lets you swap any player whose game hasn't started. So a scratch in a 10:30 game is actionable at 10:05, hours after the contest locked.
Almost nobody uses this properly. Most entrants set a lineup at 6:30 and walk away, which means every late-breaking piece of news is a free gain against a large share of the field.
Build for it deliberately. Put your flexible, lower-salary spots in late games and your locked-in core in early ones. A lineup where all eight players tip at 7:00 has thrown away the option before the night started — and the option costs nothing.
WHAT TO ACTUALLY DO
EVERY NIGHT
Afternoon. Read the injury report and note the questionable tags. Don't build yet. NBA "questionable" is genuinely uncertain in a way NFL questionable isn't, and building around one is building on sand.
Two hours out. Sketch a lineup with the salary flexibility to absorb news. Know which two or three players you'd pivot to if specific scratches happen, so you're not researching at 6:45.
Thirty to sixty minutes out. Confirmed lineups post. This is the moment. Check every team on your slate, not just the ones you have exposure to — a scratch on a team you're avoiding still changes which game is worth attacking.
Before lock. For each piece of news: who gets the minutes, who gets the usage, and are they the same person? Then check what that player has done historically with 30+ minutes before you trust the projection on him.
After lock. Stay at the screen if you have late games. That's where the free money is.
WHY THIS IS HARD
TO DO IN YOUR HEAD
Everything above is directionally right and numerically vague. How many of the 33 vacated minutes actually go to the backup versus getting spread across three players? How much does a usage bump translate into fantasy points for a player who's never carried that load?
Those are estimation problems, and they're the reason projection models exist — a minutes model that responds to depth-chart changes, and a simulation that carries the resulting uncertainty through to a range of outcomes rather than a single number. Our breakdown of how ML projections get built covers the machinery, and the Monte Carlo explainer covers why a range beats a point estimate when the whole situation is new.
None of which replaces looking at the confirmed lineups yourself. A model tells you what a 33-minute role is worth. It can't tell you the role changed until somebody tells it.
ATTENTION
IS THE EDGE
Most edges in DFS require building something — a model, a database, a process. This one mostly requires being at the screen between 6:30 and 7:00 and knowing what to do with what you see.
Minutes drive production. Salaries lag role changes. Usage and minutes get inherited by different people. The obvious beneficiary is chalk, and the player behind him is where tournaments get won. And when the news lands at 10:05 on a late game, you can still act on it while most of the field can't.
That's the whole sport. Everything else is refinement.
Put it to work
Everything in this guide can be set in one place. The NBA DFS optimizer builds DraftKings NBA lineups from minutes-first projections inside game simulations, Classic and Showdown.
Put this guide to work.
The NBA optimizer projects minutes first, simulates every game, and builds your DraftKings lineups inside those sims.
24 hours of full access, no subscription. Or every sport for $39/mo.