Is Low Latency Emulation Cheating?

[Mark_MSX]

So a while ago I posted an article about a RetroArch configuration I have been working on specifically designed for shmups called ShmupArch. The thing that makes ShmupArch interesting is that it is able to achieve only 1 frame of input lag on the shmups that it is capable of running, given the right setup.

So this posted ended up getting a great deal of feedback and stuff, and part of that feedback was the question about whether or not reducing input lag via emulation is a form of cheating, since people playing on original hardware may experience more lag than people using ShmupArch. I think this is an interesting topic, so I wrote an article about it expressing my thoughts.

http://electricunderground.io/is-low-la ... -cheating/

This is pretty much how I view the matter and thought it would be worthwhile to share my outlook since I can see this topic becoming a sticking point in the future. Hopefully my article doesn't upset too many of you out there, that certainly is not my intention. If you don't agree with my perspective, I understand, but hope you can consider where I am coming from.

Cheers!

[poptart]

It's a debate only those who have wr tier competitive scores should be having as it's completely irrelevant for everyone else, i don't personally care about it at all because im yet to see this lagless elevate anyones performance much beyond if at all the level they are currently at anyway to any degree that couldn't be attributed to placebo and other factors as there's much more to playing than just lag.

[donluca]

The point of emulation is to get as close to the original board as possible as to make the difference between emulation and PCB negligible.

If emulation goes further then it's not emulation anymore, it's an improvement and yes, that's cheating.

If the original board has large amount of lag you either live with it or play another game if you want to play for high scores or make new world records.

[poptart]

or make new world records.

so like i said it only applies to less than 1% of the player base, the level of skill required for wr tier play is such that having to readjust to a frame or 2 of lag for the sake of MUH AUTHENTICITAH PCB sperg fucks is going to not be an issue.

Anyone less than wr level its a completely pointless discussion and you should play what you want to ignoring the dadshits with candy cabs who are always garbo players anyway.

[clippa]

No, it's not cheating.
Input lag is the arch nemesis of the gamer and should be stamped out wherever we see it.

If there was a tournament then it'd be only fair that everyone was playing on identical hardware but if I was in charge it'd be the hardware that gave the least lag, not the amount of lag that was most authentic to the original.

You're not honouring the devs vision here.
Nobody who made a shmup ever said "turn up the lag a little, this doesn't feel right" it was always a compromise, an unfortunate side effect of the hardware of the time that gamers in this day and age should no longer have to put up with.

[mycophobia]

if you're submitting a score to a leaderboard or otherwise competing for something that demands at least the level of lag you would get on a real pcb, then yes, it's cheating. If you're not, then it's not. Simple as that

e: I mean I guess you could make the case that it's a cheat in the sense that infinite health or something would be a cheat, but who really cares how you play the game as long as you're not trying to pass off your play as something it's not? maybe one should note "lag free" in replays and stuff but I wouldn't hold, say, a Ketsui 2-ALL done on a lag-free setup in any less esteem than a 2-ALL done on real hardware (assuming there even is baked-in lag on that game).

[clippa]

What about people on pc. Some games come with vsync on by default, if the game feels laggy to me and I force off vsync and crank up the refresh rate, and then submit my score to the scoreboard, would you consider that cheating?

[mycophobia]

What about people on pc. Some games come with vsync on by default, if the game feels laggy to me and I force off vsync and crank up the refresh rate, and then submit my score to the scoreboard, would you consider that cheating?

The relevant question is would the person running the leaderboard consider it cheating. Are there specific rules regarding vsync and refresh rates, etc and if not, why not ask for clarification. and so on

[OmegaFlareX]

people playing on original hardware may experience more lag than people using ShmupArch

I don't think this is true. There's always going to be some input lag when emulating with a PC due to windows, drivers, LED monitors, USB polling etc. Using RetroArch with properly configged latency settings feels pretty close to real hardware but unless the current computer architecture changes, I doubt it'll ever be faster than real hardware + CRT.

[Durandal]

If low latency emulation for competitive submissions is considered cheating, then the only fair measure in response would be to set-up an official guideline so submissions are only accepted if they were played on this kind of machine with this kind of monitor and this kind of controller using this type of port to ensure that every entrant has to deal with the same amount of input lag. Which may not be very realistic

Less input lag does not necessarily have to mean better scores/times/performance, but if a method is already widely available to eliminate input lag for a game then there's no reason why everyone shouldn't try submitting new entries with the input lag elimination method enabled. Which is honestly more preferable than having to deal with the input lag for honor's sake

[neorichieb1971]

Fine tuning equipment should be encouraged and if you know how to do this to get an advantage thats a plus, but its not cheating.

Auto fire is more of a cheat then low latency lag.

[zakk]

This community has long ago accepted all sorts of deviations from 'original hardware' that I'm not sure how this is any different. No one was writing articles about shmupmame even though it sucked (but it was the best we had at the time). It's not like we have any way of even knowing if current emulation is actually 'correct' other than some people that happen to own PCBs play it and go 'well, it feels the same to me'.

For fuck's sake this entire forum once ran a competition on a completely broken version of Omega Fighter and NO ONE NOTICED FOR A DECADE. I think there's more of an argument against RetroArch from that angle; it uses old versions of many of the emulators and in some cases there have been obvious improvements to accuracy (and maybe some non-obvious). Then again it's not like this community has standards in that regard either (see previous Omega Fighter).

I'm pretty sure the only reason this run-ahead thing is even a debate (other than you like writing articles for practice) is because when it came about a few 'prominent' emulator devs shit on it and used arguments like 'competitive advantage' and 'modifying the developers original vision' to try and convince everyone this was a terrible idea. I mean maybe that matters for some Twin Galaxies level stuff but this community has long since abandoned that sort of philosophy.

[DietSoap]

Not "good" enough for WR or Twin Galaxies, but good enough for the farm. imo.

Auto fire is more of a cheat then low latency lag.

fucking delete this

[theclaw]

people playing on original hardware may experience more lag than people using ShmupArch

I don't think this is true. There's always going to be some input lag when emulating with a PC due to windows, drivers, LED monitors, USB polling etc. Using RetroArch with properly configged latency settings feels pretty close to real hardware but unless the current computer architecture changes, I doubt it'll ever be faster than real hardware + CRT.

I'd swear games on newer hardware are inherently less responsive. Even games like Tetris Attack that may seem trivial to run today, lose a certain "magic" I remember when they're emulated.

[louisg]

Wait, how would an emulator be less lag than the original PCB (assuming that it's an 80s/90s PCB and isn't going through a bunch of OS layers)? I'm legit interested, because my understanding is that old games generally worked on a pretty hard-wired game loop where it'd alternate input with video refresh. Wouldn't you have to alter the game code to do something like, say, check for input more often than once a frame?

[mycophobia]

some games have baked in lag. like battle garegga has 3 frames natively iirc, which you can get around using this new feature

[louisg]

some games have baked in lag. like battle garegga has 3 frames natively iirc, which you can get around using this new feature

Huh, how does it fix it? Does it patch the ROM?

[Meseki]

some games have baked in lag. like battle garegga has 3 frames natively iirc, which you can get around using this new feature

Huh, how does it fix it? Does it patch the ROM?

If I understand what I've heard correctly, it also stores a previous game state (from a few frames before what you see) and any input is applied to the previous game state and then the previous game state is copied to replace the current one, and multiple frames are played out in one visible frame to be on time with what you see.

Basically, your input is applied to the past and then the game state is corrected to account for such, to decrease/remove any built-in input lag; it's the same kind of thing that fighting games with rollback netcode do to mitigate the effects of latency.

[louisg]

some games have baked in lag. like battle garegga has 3 frames natively iirc, which you can get around using this new feature

Huh, how does it fix it? Does it patch the ROM?

If I understand what I've heard correctly, it also stores a previous game state (from a few frames before what you see) and any input is applied to the previous game state and then the previous game state is copied to replace the current one, and multiple frames are played out in one visible frame to be on time with what you see.

Basically, your input is applied to the past and then the game state is corrected to account for such, to decrease/remove any built-in input lag; it's the same kind of thing that fighting games with rollback netcode do to mitigate the effects of latency.

Whoa, that's seriously impressive! I always forget that we have so much memory and cpu power that we can do these things so transparently to the user. So it rolls back a couple frames, applies the input, then runs a couple frames really fast to show you the game state now.