Interesting project about mouse/ gamepad latency

[Chief Blur Buster]

Given the unique position Blur Busters is in -- and deep interest by many esports readers -- I'm willing to redirect a small bit of Blur Busters funds towards new tests for a new Blur Busters article.

Bonus if a group can define a "test outline" for peer review (including by spacediver too) -- I'd love to review the test outline so that I can chime-in onto known variables and weaknesses.

Inquire within: [email protected]

[ad8e]

Thanks, do0om. Marwan is busy atm and is focusing on his manual tracking idea. I'm supposed to be doing math so I can't really justify writing wavelet detection code right now (finding the reaction point in real-time from 1000 Hz mouse input), but I have the rest of the framework ready, including the input, display, and logging. That's enough for a click-based test but I think my slow-drag-to-jerk idea is important to have in there. It will be a few weeks until I can find an excuse to write code. But if someone else is able to supply detection code that finds pulses quickly and without false positives, taking variable mouse CPI into account, I can plug it in to my framework and we'll hand it off to do0om.

[Chief Blur Buster]

The door remains open year-round. It's the very rare researcher that is also simultaneously understanding of the esports situation. Put some science into the hyperbole.

What I know is I am seeing some shockingly uncanny stuff going on with my real eyes in person, even if it's only <1% of the players that's doing Guiness-quality stuff right in front of my eyes such as ultraprecise mouse pointer targetting in those target-practice apps that esports atheletes use.

It's orders of magnitude beyond those researchers who just pull students off the campus to try to do some reaction time tests but I roll my eyes at these street-league tests masquerading as a ResearchGate-like paper. They miss so much stuff that Blur Busters knows exists. (cue eye roll here)

So you researchers/scientists -- who google and accidentally find this thread months later -- door is open.

Some Blur Busters work is in multiple papers including this NVIDIA one (page 2) and this pursuit camera one coauthored with NIST.gov, and more coming in future. Being known as that "Refresh Rate Einstein" website, we'd love to help out where we understand certain system-side variables (button-to-photons) that many researchers are unaware exists.

[ad8e]

To correct a small misconception you have, ResearchGate is not a real journal. It's more like a shady back-alley dealer that visits the real journals and tries to make photocopies of everything. I don't know of anything dishonorable they've done, but it does appear pretty spammy and unpleasant to an academic. They've tried to spiff up their image recently.

It's probable that some researchers who specialize in neuroscience/actuation are already aware of the whole picture, and the stuff we're doing right now is a small part of what they already know. Reaction time is only tangentially related to my main bodies of knowledge, and I actually have a profound disinterest in reaction time. Even with my paltry interest, I believe I have a complete understanding of everything Marwan is doing, what he really should be doing, and what the final results and interpretation will be, before doing any tests. If I'm able to make these connections as a casual academic observer, a researcher who specializes in a more closely-related field should have a far more complete insight. The papers Marwan has cited are pretty bad, though they are valuable, so I'm not sure of the quality of the direct specialists in reaction time. But there's sufficient overlap with bigger and better fields that some really smart people probably know everything. On the other hand, they probably won't pay attention to this project unless solicited directly.

I don't mind, though, that those researchers aren't here. Pushing this along is fun, and if a real expert showed up and spoiled the ending by leveraging her vast and powerful years of experience, that would be anti-climactic. I'll try talking to someone only after all the tests are done.

[Chief Blur Buster]

Understood.

There are certainly better sources than ResearchGate -- mainly quote it like a Google (good & bad results there) or like a Wikipedia (good & bad articles there) because many people in this part of the world reference it knowing the limitations of the site. Even some of the world's best papers also end up as a copy on ResearchGate too. It's way more often read than many most-prestigious-but-paywalled journals (an unfortunate situation).

I talk to many researchers offline as well giving them brainstorms of what to study, in an openly collaborative manner. Manufacturers too. I have come up with many ideas since 2012 that I've given out for free to many others, that sometimes conflicts with newer ideas brought up more privately. One example is the "strobed variable refresh rate" algorithm ideas posted in year 2013 posted as a free Blur Busters idea, plus others have posted algorithms. Today, the ASUS TUF series gaming monitors with ELMB-Sync variable refresh strobing *seems* to be taking these ideas providing a much-less-flickery experience in variable refresh rate strobing than the ULMB+GSYNC hack. It's great that manufacturers are learning from Blur Busters ideas which we often write about freely, and we have increased reputation partially because of all our open writings over the years.

The rising tide of knowledge lifts all boats, whether it's milliseconds, or MPRT, or strobing, or motion-smoothing, etc. Sometimes ideas are proprietary but some ideas are just basic science -- that was already posted/shared/discussed by Blur Busters (and its fans) -- long before someone else came up with the idea.

[Chief Blur Buster]

I am pretty sure that a lot of people are reading the ongoing discussion in this thread.

This thread has almost 2000 views now.

Briefly looking at the logs, the lurker:poster ratio appears pretty big so I have to agree with this assessment. Long-time forum members suddenly chimed in too (having logged into their forum accounts for the first time in a while) -- such as Sparky, mello, and others -- which supports this lurker assessment.

Invariably, this grows over time, as people google/search for stuff (e.g. reaction time research and then narrowing down to esports) and sometimes those leads to these kinds of forum threads because we're such a unique google topic.

Since 2013, I know more than half a dozen people who have started companies (or new department, or new line of business) partially because they've been inspired by Blur Busters years before their business. Or because they lurked a Blur Busters Forum thread, or article, etc. It's wonderful stuff that we're an accidental incubator of "Better Than 60Hz" technology and we've generally got great reputation in some industries because of that.

Even moreso, several hundred companies/organizations have been influenced by Blur Busters. Good public example is the NIST.gov / NOKIA / Keltek researchers contacting me about the pursuit camera. Many respect the public research that Blur Busters has done. (Even big companies with lots of patent, such as NVIDIA, too!)

I bet at least 1 person from one medium/bigger size companies are paying attention to this thread, because of vested esports interests -- who knows? For example, this forums site also has visits from NVIDIA/AMD employees that aren't allowed to reply but lurk these forums anyway.

Not all of our research is perfect but we're such an idea factory around here, that reveals a lot of unturned stones.

[ad8e]

One last thing about manual tracking - while I don't see the potential for Marwan's "manual tracking reflex" to exist, Rejhon's idea that continuous reactions may lower reaction time doesn't have the same story barrier against it. It's pretty easy to find a plausible reason why it might happen, although I would strongly recommend to revise this hypothesis to "reaction tests happening in quick succession will be faster, since the earlier tests prime the later tests". (The other direction, that successive reactions may increase reaction time, is not interesting even if true, because it's too plausible and doesn't generalize well to different situations.) Testing this would be better done using the usual luminance-change stimuli rather than moving targets, and you wouldn't need a "continuous" reaction test, rather a second Yes/No test happening after the first one. This second reaction would be at a fixed timepoint after the first one. For initial experiment design, time between the two reaction tests would be pretty short to capture the effect, but not too short, because of neuronal refractory period story reasons. Some guesses for a good initial separation between the tests would be 70 ms - 400 ms.

[Sparky]

I have seen (in person) an esports player do ultra-precise mouse movements similar to this:
https://twitter.com/slasher/status/1148 ... 09761?s=21
(Just one random impressive example)

Though that example is memorization and timing accuracy, not reaction, so you expect a skilled player to be within 5ms of the mark(similar to the timing consistency of a good drummer). There's a reason some Osu! players get those two key keyboards to avoid the matrix scanning.

[1000WATT]

I have seen (in person) an esports player do ultra-precise mouse movements similar to this:
https://twitter.com/slasher/status/1148 ... 09761?s=21
(Just one random impressive example)

Though that example is memorization and timing accuracy, not reaction, so you expect a skilled player to be within 5ms of the mark(similar to the timing consistency of a good drummer). There's a reason some Osu! players get those two key keyboards to avoid the matrix scanning.

Of course, a person who is not familiar with OSU! can make a lot of wrong conclusions.
I downloaded this guy's replay. And I will try to show you. What an experienced player sees.
https://youtu.be/LkkORxRv2uQ
Even if you remove the point 1-2 and take into account clause 3-4 it's ~ 287 ms
An experienced player sees both 1 and 2.
And after 10 repetitions of the song, remember the location exactly, of each note.

[1000WATT]

Offset + is also available in the game +\- 300ms