Over the past couple weeks, I decided to dive back into Ultimate and ended up reducing the game's native console input latency by 3f (~50%)! ~6.5f (Left) -> ~3.4f (right) native on console Huge thanks to @SSBM_Arte for testing and providing these graphs. Details in the thread!

i hung this snow shovel in the modern art museum and nobody noticed

Shout out to the buffalo smash player who was in front of me on the train to penn station today your timbs were lit

Again, I'll be going into more details in a full write up after LMBM, so I won't go into details now, but most of the reduction on the game side, and a small fraction* of it is from custom USB drivers. *if you are on mayflash, this USB driver saves you another frame of lag

I guess I should have been more clear that the "console input latency" referred to in that post is about total latency from pressing the button to seeing the result, not console side latency.

Lots of confusion about whether I removed "console lag" or "game lag". A while ago I was incorrect in saying that the console responsible. ~5.5f of input delay are caused by the game. This removes 2.8 of those, and the new consistency brings the average to 3.0f improvement

The most insanely skilled modder and single biggest contributor to SSBU QoL does it again : ) Some extra info: this was measured on a cheap 60Hz portable monitor, definitely not the fastest, so don't mind the absolute values too much, we care about relative. While I don't have a

@SSBM_Arte Oh and I suppose some quick FAQs that might come up: - Wi-Fi safe (tested) - Emulator(s) can go down one more frame, I had to add it back for console

As a final note, the numbers above are based on Arte's portable monitor, which is not the best 😅. The absolute measurements in these graphs are probably above average, but what's important is the relative difference.

This has been a huge undertaking! I look forward to releasing it later this month when I get back from LMBM - there are still a couple kinks to work out with the new USB driver. There will be a full technical write-up posted with the release for those interested 🙂

There is another key difference in the graphs: input variance. The best variance you can possibly have is 1f (~16.7ms). My changes reduce Ultimate's base variance of ~26ms to ~17ms, removing 95% of the possible "missed polls". Spot the difference in the data outside the red bars.

The graphs shown were measured with the help of Arte and his latency testing setup. They measure the latency between an "A" press and the second frame of Pikachu jab, so they are adjusted by one frame's length to indicate when the game received the input.