Monster Hunter Wilds patch 4: what it fixed, and what it didn't

The patch notes for Ver.1.042.00.03 list five items under performance: texture streaming efficiency, shader compilation, multiplayer sync overhead, VRAM usage, and a catch-all "overall stability improvements".

We ran the game before and after the patch to see which of the five are real.

How we tested

Same machine (i5-13600K, RTX 4070, 32GB), same save, same quest: Wildspire Waste, a four-player hunt of one large monster, from departure to rewards, 30 minutes end to end. High preset, frame generation on, DLSS Quality.

Five runs, median taken. We captured frame times in PresentMon and looked at two numbers: the median and the 1% low.

Results: throughput up, hitching unchanged

Metric	Before	After	Change
Average fps	52	67	+29%
Median frame time	19.2 ms	15.0 ms	-22%
1% low	41 fps	43 fps	+5%
0.1% low	28 fps	29 fps	+4%

The patch notes did not exaggerate. Median frame time really did fall 22%, and everyday traversal feels noticeably smoother.

The 1% low went from 41 to 43. The worst one percent of frames barely changed.

That is why forum posts after the update still said it stutters. Those players are not describing average frame rate. They mean the hitch when you cross into a new zone, or when a large monster winds up its biggest move. The patch improved steady-state throughput and left the tail alone.

Where the tail comes from

We captured call-stack signatures for the slow frames. Three sources dominated.

Texture streaming. For the first three to five seconds after leaving camp for a new area, frame times repeatedly hit 40 to 60 ms. The patch claims better streaming efficiency, and that window did get shorter, from about six seconds to four. The peak did not drop.

Shader compilation. The first time a given effect appears, such as a monster entering rage for the first time, you get one or two long frames. This patch does not fix it, because it is an engine-level precompilation policy. The good news is that it happens once; a second run of the same quest is clean.

Four-player sync. A spike when a teammate joins or leaves the scene. The patch claims an optimisation. Our numbers land inside the error bars.

Three settings that actually help

The parts the patch cannot fix, you can.

Drop textures from High to Medium. This alone lifted the 1% low from 43 to 49, the largest single gain we found. Distant terrain gets slightly soft, but during a hunt you are watching the monster, not the horizon.

Turn volumetric fog off. Median frame time drops another 2 ms, and the "stutter when entering fog" pattern mostly goes away.

Turn frame generation off. This one is counterintuitive: frame generation inflates average frame rate without reducing input latency, and below a 50 fps base it makes the controls feel worse. If you are running high-difficulty quests, turn it off.

One more thing: after loading in, stand in town for a minute and let shader compilation finish. It saves the two long frames on your first hunt.

Verdict

This patch does most of what the notes promised. The median improvement is real and noticeable.

The gap is in what "optimisation" means here. The studio raised the average; players complain about the tail. A 2 fps gain on the 1% low says the target and the pain point do not fully overlap.

If the next patch pulls down the peak cost of texture streaming, the game is genuinely fixed.