A stuck pixel is worth trying to fix, and a dead one is not. That is the whole decision, and it takes about thirty seconds to make: a stuck pixel is still receiving power and jammed showing one colour, so cycling colours through it at speed can shake it loose. A dead pixel receives nothing, and no software on earth will bring it back. This page covers the methods that genuinely work, how long each deserves, the ones that will damage your screen, and the point at which to stop.
On this page
Do this before anything else, because attempting a fix on a dead pixel is time spent on something that cannot work.
Display a full black screen, then a full white one. A dot visible on black but not white is stuck — still powered, jammed on, worth attempting. A dot visible on white but not black is dead — no power, nothing to work with. A dot that is white on both is a hot pixel, which behaves like a stuck one and responds to the same treatment.

Clean the screen first. Dust accounts for a large share of reported faults and wipes straight off — the safe way to clean a screen avoids the pressure that creates real ones. And if the dot moves or vanishes after a reboot or a cable swap, it was a driver or cable artefact, not a pixel. Dead versus stuck pixels covers the rest of the distinguishing signs.
This is the method to start with and, for most people, the only one worth using. Rapidly cycling colours through the stuck pixel forces its sub-pixel to switch on and off many times a second, and that repeated exercise can unstick a transistor that has seized. It is entirely non-invasive, carries no risk whatsoever to the panel, and costs nothing but time.

It does not always work, and nobody honest will tell you otherwise. But a reasonable share of genuinely stuck pixels do come back, and since the method risks nothing, there is no argument for skipping it.
An hour is the minimum worth bothering with. Several hours is better. Some pixels only respond after being left overnight, so if the first attempt does nothing, a full night is the reasonable second attempt.
The single most common reason the method “fails” is that the screen went to sleep. Turn off automatic sleep, screen lock, screensaver and any battery-saver display timeout before you start, and if it is a laptop, leave it plugged in.
If a full night produces no change at all, stop. At that point the sub-pixel has almost certainly failed outright rather than seized — the transistor has stopped working rather than frozen in one position. It is dead in everything but name, more hours will not help, and the question becomes a warranty one.
Sometimes, and it is genuinely risky. Gentle, targeted pressure can realign the liquid crystal in a cell that has seized. Too much pressure creates new dead pixels or a permanent mark, and that damage is not recoverable. Treat it as a last resort, only after the flashing method has had a full night.
Never attempt this on an OLED panel. The structure sits close to the surface and damages far more easily.
If it has not worked after two or three careful attempts, stop. Escalating the force is how a single stuck pixel becomes a cluster of dead ones plus a pressure mark, and you will have turned a minor annoyance into a panel that genuinely needs replacing.
Several popular suggestions are actively harmful:
The underlying point: pressure damage is a real cause of dead pixels, not a theoretical one. What causes dead pixels covers how much of it is self-inflicted.
If the flashing pattern has run overnight with no change, the most likely explanation is that the sub-pixel is not stuck at all — the transistor has failed completely rather than seized in one position. Nothing in software will recover that.
Two other outcomes are worth recognising because they mean something different:
In all three cases the honest answer is that the panel is faulty rather than the pixel, and the next step is a claim rather than another night of flashing colours.
Most manufacturers publish a pixel policy with a threshold: a number of failed pixels before the panel counts as defective. Stuck pixels are frequently counted differently from dead ones, and a bright pixel in the centre of the screen is usually weighted more heavily than one near an edge. A single stuck pixel commonly falls below the threshold on consumer panels, while professional and premium displays tend to have stricter terms — occasionally a zero-defect guarantee that one pixel will trigger.
Before contacting anyone:
Timing matters more than any of that. A stuck pixel caused by a manufacturing fault is present from the first power-up, and most return windows and pixel policies run for a short period after purchase. Test any new screen the day it arrives. Testing a monitor covers what to document, and we publish the results people report to us by manufacturer if you want to know how common this is on your brand first.
You cannot prevent a manufacturing defect, which is the largest single cause. The stress-related causes are more within your control:
One thing that does not help: running a colour-cycling pattern preventively on a healthy screen. It has no protective effect at all and simply adds hours of use. Run it when you have a stuck pixel, not before.