Match your sens, keep it, and if you are new, go lower

Sensitivity is the one setting that touches every single shot you take. It decides how far your hand moves to turn your view, and so it decides which muscles do the aiming, how much your hand's natural wobble shows up on screen, and whether the distances you have practised still mean anything.

The advice in this post fits in one line: match your in-game sensitivity, keep it, and if you are new, go lower. Here is the evidence for each part, from our players' data and from independent research, including where that evidence is thinner than the internet suggests.

1. Match it, because most players do not

Of the 4,685 mouse players whose sensitivity we could see, 62% never changed it from the default. Some of them happen to play at that speed, but most are training their hand to travel distances their real game will never ask for.

The fix is simple, but compare cm/360, not the sensitivity number. cm/360 is how far your mouse travels for a full 360° turn, and it is the same physical distance whatever game you are in. Games turn by different amounts per mouse count: Counter-Strike's yaw is 0.022° per count (it is in Valve's own source code), and Valorant's is 0.07° per count (measured by the community, since Riot does not publish it).^1,2 So Valorant sens ≈ CS2 sens ÷ 3.18. Our sensitivity converter does this for 7 games.

2. If you are new, go lower

Here is the part that surprised us. We took players in their first 10 runs on Online Aim Trainer who had set their own sensitivity: 1,682 of them, all starting at the same difficulty. Then we grouped them by the cm/360 they played at.

Raw accuracy tells the same story: 61.8% for new players under 20 cm/360, against 71.4% at 55–80 cm. And it holds for experienced players too. Among players with more than 50 runs, those on very high sensitivity played at an average difficulty level of about 24, while players between 30 and 80 cm played at 35 to 38 (difficulty rises as players improve). That second comparison rests on smaller groups (25 to 60 players each), so treat it as a hint rather than a measurement.

The research agrees, with a sweet spot

The most direct study comes from NVIDIA's research group. They had experienced FPS players do 500 aiming trials at each of six sensitivities, from 5 to 160 cm/360. Performance was best between 20 and 80 cm/360. The trade-off they describe is exactly the one in our chart: "low sensitivities reduce in-game view rotation speed while high sensitivities result in reduced motion precision (increased corrective motions)." Hit rate was highest at the lowest sensitivities, and lower sensitivity helped most on small and distant targets.³

Classic research on desktop mice found the same two-sided limit. High mouse gain "increased overshooting, indicating an issue with muscle control accuracy," while very low gain hurt speed "largely because of increased clutching" (lifting and re-placing the mouse).⁴

And the pros sit right in that window. Across the 670 Valorant pros tracked by ProSettings.net in October 2026, the median eDPI is 230, about 57 cm/360, and the middle half of all pros sits between roughly 47 and 71 cm. In CS2 (902 pros) the median is 800 eDPI, about 52 cm/360.^5,6

Why lower helps

Every hand has a tiny amount of unavoidable wobble. The desktop-mouse research puts the precision floor of the hand at around 0.2 mm.⁴ Sensitivity multiplies that wobble into crosshair movement: at high sens, a fraction of a millimetre of jitter is enough to slide off a head at range. At lower sens, the same jitter moves your crosshair less.

Lower sensitivity also changes which muscles do the work. Motion-tracking studies show that as gain rises, people shift from using the whole arm to just the hand and fingers.⁴ Bigger, stronger muscles produce force with less variability: in lab measurements, "stronger more proximal joints in the human arm have lower levels of motor noise than weaker more distal joints."⁷ That is the honest version of "use your arm": not that arm aiming is magic, but that spreading the movement over steadier muscles and shrinking the effect of jitter both point the same way.

Just do not overdo it. Past about 80 cm/360 our new players got worse again, and the research explains why: you run out of mousepad and have to lift the mouse mid-fight.

What that looks like in your game

Here is the in-game sensitivity for common cm/360 values, at 800 DPI. (At 1600 DPI, halve it; at 400, double it.)

3. Then keep it

Motor-learning research is clear that learning a new hand-to-screen mapping interferes with the one you already have. In a well-known experiment, learning one cursor rotation and then, five minutes later, the opposite one wiped out the benefit of the first training when people were tested the next day.⁸ A three-lab study found the second mapping "fully interferes" with the first even when learned 24 hours later.⁹ In the NVIDIA study, the researchers threw away the first half of every session (250 trials) as adaptation time each time they changed sensitivity.³

To be fair to the other side: a plain change of speed is one of the easiest mappings to adapt to. It generalizes across distances and directions far better than a rotation does.¹⁰ In our own data we could not find a measurable penalty on the run right after someone changed their sensitivity. So changing sensitivity once will not ruin your aim. What hurts is never settling: every change means your hand has to relearn the distance to every common angle, and those distances are what you are training.

How to pick yours

For pro examples and recommended ranges by role, see what is a good Valorant sensitivity.