Steam Battlefield 6's loudest PC complaint is that the mouse feels choppy while a controller feels smooth, and the fix everybody repeats is to cut your polling rate. Nobody checks what it does. A game reads the mouse once per frame — so a mouse that reports slower than your display refreshes leaves frames with nothing in them at all.
Move your mouse across this strip, steadily, for three seconds.
Nothing leaves this tab The strip reads pointer events and the display's own frame clock. No pointer position, no movement and no timing is sent anywhere.
A report rate is a period and a refresh rate is a period. Divide one by the other and you have how many times the mouse speaks per frame. Below one, the mouse cannot fill every frame — and the frames it misses are frames where the game turns your view by nothing, because a mouse reports movement rather than position.
| Display | 125 Hz | 500 Hz | 1000 Hz |
|---|---|---|---|
| 60 fps16.7 ms frame | 2.1× | 8.3× | 16.7× |
| 120 fps8.3 ms frame | 1.0× | 4.2× | 8.3× |
| 144 fps6.9 ms frame | 0.87× 13% empty | 3.5× | 6.9× |
| 240 fps4.2 ms frame | 0.52× 48% empty | 2.1× | 4.2× |
That leaves about 13% of frames with no new mouse data in them. The advice is repeated as though it were free. It is not free at this refresh rate, and the number has been sitting in plain arithmetic the whole time.
The table above is true for everybody. This one is true for you: your own sample, batched the way a slower mouse would have sent it, then cut into your own display's frames. The movement is preserved exactly — a slower mouse does not lose distance, it delivers it in fewer, larger packets — so the only thing changing between rows is when it arrives.
Move the mouse across the strip at the top of the page and this fills in.
It is the number on the box and the one every guide argues about. On its own it decides almost nothing: past the point where every frame gets a report, more reports per frame do not give the game anything it can act on until the next frame arrives.
A 1000 Hz mouse whose reports land at 1 ms, then 0.4, then 3.1 is delivering its rate and still arriving unevenly. This is the one people feel and the one the rate number hides — which is why the reading beside the bench shows the spread, not just the average.
The game does not see your mouse. It sees whatever arrived since its last read, once per frame. That is the only place rate and jitter turn into something on screen, and it is the only one of the three that the popular fix changes for the worse.
A thumbstick reports a position, so a frame that reads it late still reads the correct value. A mouse reports movement since the last report, so the game must add up whatever arrived inside that frame — and a frame that received nothing turns your view by nothing.
A late read of a stick is still a correct read.
So the game has to add up the packets that arrived inside the frame. Two frames that received different numbers of packets turn your view by different amounts even when your hand moved at a constant speed — and that difference is the texture people describe as choppy.
A missed read of a mouse is movement that never happened.
What this does not settle Whether controller aim assist gives an advantage in Battlefield 6 is a separate argument and this page takes no side in it. The point here is narrower and checkable: the smoothness difference has a mechanical explanation that exists before aim assist is switched on.
Read from this site's own verified Battlefield 6 entry, checked 7 September 2026. These are listed because they sit on the input path, not because they are "best" — a number that is right for one hand and one mousepad is not a recommendation for yours.
Removes Windows acceleration and smoothing from the chain.
Keeps every scope on the same distance-per-360 relationship; 178 approximates 0% monitor distance.
Equivalent to roughly 30–40 cm/360 at 800 DPI; convert your CS2 or Valorant number instead of guessing.
Lower for aircraft so a full mouse swipe is not a full stall.
Lets a fast swipe finish a 180° turn even when your pad runs out.
Battlefield 6 picks crouch or slide from your speed, so one key does both jobs.
Q as a hold key fights with strafing; move it to a thumb button.
Full binding list Every default on keyboard, pad, vehicle, aircraft and helicopter is on the Battlefield 6 controls page, with its sources and the date each was checked.
Every list of fixes for this problem is a list of everything at once, which is why nobody who follows one knows which part worked. Each step here names the measurement that confirms it, so a change you cannot see is a change you can put back.
Sample the strip above and write down your report rate, your empty-frame share and your jitter. Without a before, every after is an opinion.
Windows pointer precision and the game's own raw input setting are two separate switches on the same path. The settings checker reads what the browser can see of the first.
And read the table above for your refresh rate before you do. If your empty-frame share is already zero, dropping the rate can only take it away from you.
Same strip, same hand speed, same tab. A change that moves none of the three moved nothing you can see either.
Pointer events in this tab, timestamped by the browser, and the display's own frame clock read from the animation callback. The rate, median interval and jitter come from the same analyser the site's polling-rate test uses, so the two pages cannot disagree about your mouse. Lower rates are modelled by batching your real reports, which preserves the distance exactly and changes only its arrival pattern.
Nothing is uploaded. The sample lives in this tab and ends with it.
A browser is not the game. It can be given coalesced events, it is subject to the tab's own scheduling, and it sees nothing of the engine, the driver or the USB stack. So this page never says what Battlefield 6 does with your input — only what your input looks like before anything receives it. The engine-level stutter people describe may well survive every change on this page.
Three consistent runs beat one dramatic one.
One measurement caveat worth stating plainly: a browser may hand a page one merged event in place of several. Where that happens the observed rate reads low through no fault of the mouse — which is why the bench reports the interval spread beside the rate, and why a rate far under your driver setting is a reason to repeat the test rather than to change a setting.
Part of it has a mechanical explanation that has nothing to do with aim assist. A thumbstick reports a position, so a frame that reads it late still reads the correct value. A mouse reports how far it moved since the last report, so the game has to add up whatever packets landed inside that frame — and two frames holding different numbers of packets turn your view by different amounts even when your hand moved at a constant speed. The bench on this page draws that unevenness from your own stream. It cannot tell you how much of the remaining difference is the engine-level problem players have reported, because a browser cannot see inside a game.
Check what it would cost you first, because it is not free. A game reads the mouse once per frame. At 144 fps a frame is 6.9 ms and a 125 Hz mouse reports every 8 ms, so the report cannot keep up with the frame and roughly one frame in eight arrives with no new mouse data at all. At 60 fps a frame is 16.7 ms and the same mouse fills every one, which is probably why the advice worked for whoever first posted it. The table on this page has the arithmetic for every common combination, and the bench runs it on your actual stream.
There is no single answer and anyone giving you one is not measuring your machine. The useful rule is narrower: your report period should be comfortably shorter than your frame, so every frame has something in it. Past that point, extra reports per frame give the game nothing it can act on until the next frame arrives. Measure your empty-frame share on this page; if it is already zero, lowering the rate can only take that away.
Players have widely described it as engine-level, and the developers have acknowledged the feedback. This page does not adjudicate that. What it can do is separate the part that is yours from the part that is not: if your own stream shows heavy interval jitter or empty frames before the game is involved, that part is fixable on your desk. If your stream is clean and the game still feels choppy, no setting on your end is going to explain it.
Raw input removes Windows acceleration and smoothing from the chain, which is a different problem from the one this page measures — acceleration changes how far your view turns for a given hand movement, while per-frame delivery changes how evenly that turn is spread. Both are worth having right, and they are worth changing one at a time so you know which one did anything. The verified Battlefield 6 settings list is on /controls/battlefield-6/.
It keeps every optic on the same relationship between hand distance and degrees turned, so switching scopes does not silently change your sensitivity. 178 approximates the zero-monitor-distance match on a 16:9 display. It is a separate axis from everything on this page: it decides how far your view goes, not how evenly the movement arrives. The full verified settings list, with sources and check dates, is on /controls/battlefield-6/.
Browsers may hand a page one merged event in place of several, and the tab is also subject to its own scheduling. The bench here unpacks coalesced events where the browser offers them, which recovers most of it, but a reading well under your driver setting is a reason to repeat the test in another browser rather than to change a setting. That is also why the reading shows the interval spread beside the rate — a stable distribution says more than one peak number.
Battlefield 6 ships more than 220 configurable slots across infantry, ground vehicles, aircraft, helicopters, gunner seats and the UI. The full verified list for keyboard and controller, including the vehicle layouts most people never open, is on /controls/battlefield-6/ with the date each was checked.
Report rate, interval jitter, empty frames. They take three seconds to collect and they turn the loudest argument on the Battlefield forums into something you can check on your own desk.
The input path has more than one place to go wrong. These measure the rest of it.
I answer from this site only — the 28 browser tests, the controls library for 46 games, the sensitivity converters and the two games. I can tell you which page to open, how a tool works and what its result means.
I will not guess. If something is not on the site, I will say so.
Answers come from this site's pages.