NBA 2K27 · jump shot timing lab

The green window,
measured to the millisecond.

Everything here is free. No account needed to browse the jump shots.
Measured, not guessed

We show the raw shots

Every millisecond we tested, with the greens over the attempts. Not a smoothed curve, not a star rating. If a number here is thin, you can see that it is thin and decide for yourself.

Which way you miss

Every shot is logged as green, too early, or too late, so early and late error rates can be compared at the same offset from the peak hold.

Where we actually are
stated plainly, because a database that overstates itself is worth nothing
Every row: 6'4, 90 3PT, no badge, Superstar, open, catch-and-shoot, shot meter off, hot zones off.
JumperBuildRoomPeakShots
Quentin Grimes / Oscar Robertson + Sam Hauser · Release6'7 · 200lb · 93 3PT50ms (505-555ms)86.7% at 530ms185
2K27 early access is August 26

Day one, every jumper resets and every number on every site goes stale. We start testing immediately. When the list is live, you get the 2K27 windows as they land, before they are public.

We'll email you when 2K27 windows land.

Jump Shots

Vote Which jump shot should we measure next? Add yours or vote for one already on the board.

Any
Base Release 1 Release 2 Min ht Under Min mid Min 3PT 3PT Speed Window 50%+ window 50%+ avg Peak Window avg Ht Spd Def Win

Don't see the jumper you're after? Ask for it and we'll measure it on the rig.

Request a jumpshot
← Back to jump shots

Window width
Green window
Peak hold
Green window curve
Green rate at each hold. Shaded band is the 95% interval.
5ms steps
Loadout
Outcome by hold time
Green Too early Too late

Badge Ladder

one jumper, one build, every tier of one badge

Everything else held fixed: build, difficulty, contest, shot type.
Where the window sits at each tier
440ms530ms620ms
TierWindow Range Room Peak Room gained
All tiers on one axis
the same curves stacked, so a real difference separates and a fake one does not

Curve at each tier
same axes on every cell, so the shift is real and not a scaling trick
Delta table each tier against the baseline you picked above
TierRoomΔ room PeakΔ peak EarliestΔ timing nVerdictMeasured
Δ timing negative means the window arrives earlier, so you release sooner. A tier is only called a change when the 95% bands do not overlap.
← Back to jump shots

Compare

Overlaid green window curves
440ms530ms620ms
← Back to jump shots

Standards

Every number on this site is a count of real shots taken under settings that do not change between runs. Nothing here is voted on, averaged from clips, or estimated from feel. Where the evidence is thin we say so on the row rather than rounding it away.

Held identical on every run
Shot meter
off
Boosts
off
Hot zones
off
Difficulty
Superstar
Contest
open
Shot type
catch and shoot
Stamina
full
Change any one of these and it becomes a different row, never a footnote on this one.
When a window gets published
  • A millisecond is only called confirmed once it has been repeated enough times that the result stops moving.
  • Anything short of that publishes as still testing and is drawn faded, not hidden.
  • A row carries its own shot count, so you can weigh it yourself.
  • A dash means we have not tested it. A zero means we tested it and it was zero.
What we will not do
  • Publish a rate without its shot count.
  • Smooth a curve until the raw counts disappear.
  • Carry a number from one game into the next and call it current.
  • Call a jumper "best" without naming the build.
  • Replace a count with a slogan.
Update log newest first