Does 4:3 stretched change your sensitivity in CS2?

No. Your sensitivity, your cm/360 and the hand movement needed to reach any target are identical at 4:3 stretched and 16:9 native. What stretched changes is how far things travel across your monitor: on a 16:9 panel, horizontal movement on screen is 1.33 times faster at the same rotation, vertical movement is unchanged, and targets are 1.33 times wider. Because targets grow by the same factor as the movement, nothing about your aim in mouse terms has changed. Keep your sensitivity.

Why the 360 does not move

CS2's sensitivity is a rotation, not a screen distance. Valve's console variable list gives m_yaw a default of 0.022 and describes it as the "Mouse yaw factor": every count your mouse reports turns you 0.022 degrees multiplied by your sensitivity. Neither variable has a resolution term, so the distance for a full turn is the same at any resolution or aspect ratio:

That is the same cm/360 formula this site uses everywhere, with the Source family constant from the constants page. If you measure your 360 with tape on 16:9 and again on 4:3 stretched, you get the same distance.

What does change: the projection

CS2 scales its view the way the Source engine always has. The Valve Developer Community says aspect ratios "are achieved by expanding the horizontal field of vision, not by cropping top and bottom portions of the image", and gives the formula with 4:3 as the standard ratio. Wikipedia calls this Hor+: "the vertical FOV is fixed, while the horizontal FOV is expandable depending on the aspect ratio". csdb.gg puts CS2's base vertical field of view at 73.74 degrees, which is exactly 90 degrees horizontal at 4:3.

From there the screen speed follows. At the crosshair, pixels per radian of rotation equal half the image width divided by tan(horizontal FOV / 2). Native 16:9 on a 1920 x 1080 panel gives 960 / 1.333 = 720. A 4:3 render stretched to the same 1920 pixels gives 960 / 1.000 = 960. Vertically both are 540 / 0.75 = 720, because the vertical field of view never changed.

Black bars land on 1.00x because the 4:3 image is drawn 1440 pixels wide, and 720 / tan(45°) is 720 pixels per radian, the same as native. csdb.gg lists 86.07 degrees for 5:4; the formula gives 86.30, and we use the derived value.

Why your aim does not change

Take a target 30 degrees to the right of your crosshair, at 800 DPI and sensitivity 1.0.

The target is 138 pixels further away on screen, and you reach it with exactly the same 4.33 cm. The target is also 33% wider, so the hand movement that puts the crosshair on it, and the error you can afford while doing it, are both unchanged. Flicking, tracking and micro-adjusting are all the same task in mouse terms. Only the visual feedback runs faster sideways.

That is why both popular answers are half right. "Nothing changes, the crosshair moves the same pixels" is wrong about the pixels: the crosshair sweeps 33% more screen per degree horizontally. "Stretched raises your sensitivity, so lower it" is wrong about the sensitivity: the rotation per count is untouched. The 25% figure that also circulates, including on csdb.gg's calculator, which says models appear "~25% wider", is the share of screen width a 4:3 image gives up (1440 of 1920 pixels), not the stretch, which is 1920 / 1440 = 1.333.

Should you change your sensitivity?

No. This is the 0% case from our monitor distance coefficient framework: the engine has already kept your rotation per count constant, and your angle memory is worth more than your screen memory.

If you still want stretched to look like 16:9 when you move, the arithmetic is to multiply your sensitivity by 0.75. At 800 DPI that takes sens 1.0 to 0.75 and your 360 from 51.95 cm to 69.27 cm, so every flick you have trained is now a third too short. It also slows your vertical movement by a quarter, which stretched never changed in the first place. A single sensitivity number cannot correct one axis without the other. The only axis-specific lever is m_yaw, which Valve's list carries without the cheat flag; 0.0165 would be the horizontal-only match, and it would still lengthen your 360 by a third. We do not recommend either.

Where this stops applying

• Other games. The table assumes CS2's fixed vertical field of view. Games that scale field of view another way give other numbers, and we have not sourced Valorant's behaviour, so this page does not cover it.
• Other panels. The stretch factor is panel aspect divided by rendered aspect. On a 16:10 monitor, 4:3 stretched is 1.6 / 1.333 = 1.2x, not 1.33x.
• Scaling that does not stretch. If your GPU or monitor is set to preserve aspect ratio, you are in the black bars row whatever the game's menu says. The 4:3 image keeps native proportions and native screen speed.
• Feel is real. Faster horizontal motion on screen can change how tracking feels for the first sessions after a switch. That is a visual adaptation, and changing the sensitivity to cancel it trades it for a motor one.