Deadlock FPS Drops and Stuttering: What to Check After an Update

The hideout runs beautifully. You enter the city, turn the camera, and the smoothness disappears. You lower the graphics, then lower them again. The game keeps looking worse without becoming much easier to play in a fight.

Players on different hardware have reported performance drops after City Never Sleeps, including large changes depending on camera direction. As of October 2, 2026, there is no verified universal fix. There is, however, a useful way to find which adjustments actually help your machine.

Start with the FPS and frame-time display, find a repeatable scene, and change one setting at a time. If an old performance config has survived several major patches, test that separately too.

Search Settings for HUD Telemetry. Enable Show frame time and FPS, plus Show ping and Show network problems for comparison. Use continuous display while troubleshooting where that option is available.

FPS describes frame frequency. Frame time describes how long a frame takes; the built-in display reports the worst recent frame time. An average of 120 FPS therefore does not prove that an unpleasant hitch never happened.

At a steady 60 FPS, a frame takes roughly 16.7 milliseconds. At 120 FPS, it takes about 8.3 milliseconds. Those are arithmetic examples, not targets for your hardware. A sharp spike when the game hitches gives you something useful to investigate.

If the picture remains smooth but players jump around while network indicators worsen, you are looking at a different problem. Lower shadow quality cannot repair missing network packets. Both issues can occur together, so watch the readings during the actual interruption.

For an initial comparison, use Explore NYC and find a location where performance noticeably drops. Stand still, face a dense part of the city, then turn toward another view. Remember the position, direction and result.

Now you have something to compare. Change a setting and return to the same spot. “Yesterday I had 180 FPS and today I have 90” says much less without matched conditions: locations, match stages and effects introduce too many differences.

An empty-map test cannot represent a fight involving several players. It helps you reject adjustments that clearly do nothing. Take promising settings into a normal match afterward, especially the situations that made you investigate performance in the first place.

Use the following as a starting point for comparison, not a universal best-settings preset. Write down your existing values or take screenshots before changing them.

You do not need to leave every option at its minimum. If shadows look worse and the result barely changes, restore them and move on. This gives you a configuration you understand, including what each performance gain costs visually.

Be particularly careful with internal resolution. Making an enemy harder to distinguish is a poor trade for a tiny improvement. Inspect camera movement, fine detail and targets against the streets, not just a stationary screenshot.

Lower resolution reduces pixel-rendering work. A frame also depends on other work, including scene preparation, game logic and processing objects. Making the picture smaller does not guarantee a proportional increase in frame rate.

If a substantial reduction in internal resolution barely changes the same scene, stop making the image progressively blurrier. That result does not identify the exact component limiting performance. It does show that this particular adjustment is doing very little.

The same caution applies to “just switch to Vulkan.” If you compare graphics APIs, record the original choice, restart after switching, and use the same scene. There is no universal winner to promise across all systems.

A low overall CPU percentage also does not establish that the processor is irrelevant. An average across every core cannot tell you whether the work needed for the next frame is finishing in time.

If you installed mods or copied launch commands from a guide, compare the game with and without those additions. Save your settings and back up modified files first, then temporarily disable the specific modifications and custom launch options you recognize.

Avoid deleting game folders or the entire userdata directory at random. Adding another performance config on top of the first makes it even harder to identify what is affecting the result.

If you suspect damaged files, use Steam’s verification feature: Library → Deadlock → Properties → Installed Files → Verify integrity of game files. It checks the installation; it does not promise to remove every extra mod file or fix an engine performance problem.

If the client is preloading the map and compiling shaders, let that process finish. Clearing shader caches every time performance drops, without a specific reason, is not a useful maintenance habit. Some of that work will need to happen again afterward.

Compare the original and adjusted settings in the same scene more than once. Look for a consistent result rather than the highest number that briefly appeared. Close a heavy background task if it was running during only one comparison.

Then keep the chosen configuration for a match. If the empty city becomes smoother but late fights still struggle, you have improved one workload without resolving another. That is useful information; it just is not a complete fix.

If frame rate is stable but fights remain difficult to follow, separately adjust the new HUD and health bars. Hiding names or calming indicators can improve readability. That does not establish a particular FPS increase.

Revisit your problem scene after subsequent updates. Our guide to reading patch notes helps identify which changes are relevant. One unusually smooth match does not prove a problem has vanished, but you do not need to rebuild your configuration from scratch every time either.