• Hey everyone, staff have documented a list of banned content and subject matter that we feel are not consistent with site values, and don't make sense to host discussion of on Famiboards. This list (and the relevant reasoning per item) is viewable here.
  • Furukawa Speaks! We discuss the announcement of the Nintendo Switch Successor and our June Direct Predictions on the new episode of the Famiboards Discussion Club! Check it out here!

StarTopic Future Nintendo Hardware & Technology Speculation & Discussion |ST| (New Staff Post, Please read)

I’m assuming most of these changes are things that we would never notice?
As in changes to the store/login/cloud services? Yes. Changes to the way that games handle their online? Maybe? LiC has called me out for using too general a term once already - I'll say that any game that is mostly peer-to-peer can't be substantially improved with patching the game, because updating a server doesn't change what happens once players join a race/match/whatever.
 
As in changes to the store/login/cloud services? Yes. Changes to the way that games handle their online? Maybe? LiC has called me out for using too general a term once already - I'll say that any game that is mostly peer-to-peer can't be substantially improved with patching the game, because updating a server doesn't change what happens once players join a race/match/whatever.
I’m relatively impressed with online at least when it comes to Mario Kart. The fact I can play in a lobby with only Japanese players when I’m in the US and the performance is decent is surprising to me.

I would expect the rubber banding to be much worse but it feels like it’s gotten better over the years. I remember Smash being really bad but I never used Ethernet and the wifi card in the Switch is already pretty bad so I’m sure that didn’t help.
 
Whether PM is 30fps, 60fps, or even 240fps, it ultimately does not matter. I’m also just in general fed up with the 30 vs 60 frames per second debate because most of the time, they never seem to ask a very simple question:

How responsive is the game?

I’ve played 30fps games with awesome responsiveness, and 60fps games that are quite laggy.

I remember playing the demo for Kirby Forgotten Land, and was amazed at how responsive, and tight the controls felt. I actually mistakened it as a 60fps game at times because it felt so nice to play. But I also know Nintendo are pretty good at this responsiveness sort of thing where some other studios fail at this, or use terms like “weighty,” and “immersive“ to justify laggy controls. Killzone 2 was notorious for this back in the day, and while I got over it then, I think it’s a stupid excuse now, and it’s a dumb buzzword to throw around for some developers' shitty programming. A higher frame rate can help mask some of those shortcomings, while a 30fps cap would expose it.

Another example is Uncharted Trilogy on PS4, which all play at 60fps. It certainly feels better to play than PS3, but because Naughty Dog wanted a more “weighty” feel, the controls don’t always feel the greatest, and aiming can sometimes feel a bit of, or…laggy.

By comparison, I played all of Doom Eternal on Switch, and it’s one of the most responsive shooters I’ve ever played at 30fps. Despite the lower frame rate, it feels rather tight. And this on top that I’ve played Eternal at 60fps on Steam Deck, and my Desktop.

Finally, whether it is 30fps, or 60fps, your brain will adapt to it, and rather quickly really. And within ten mins of play, you'll be engaged enough in the experience to not give a shit. But I’ve also grown up as a gamer playing tons of N64 titles at less than 30fps, and in some cases, damn near 10fps, and I’m just happy to be playing. My brain adapts, and I forget about it...provided the controls are responsive. Mario 64 on N64, despite running at 20fps, feels rather responsive too, and you sometimes forget it’s running as low as it is. Doesn’t hurt that Mario 64 is also still a joy to play nearly 30 years later. 🤷‍♂️
 
Do we actually know the timing for action commands? A developer on here suggested tying mechanics to frames is outdated (or something like that) and it's possible the game instead polls the controller at a higher rate, making it possible the timing is the same. I don't know much about it though.
The main thing was Good Vibes Gaming thinking that the timings seemed different when they were watching the Nintendo rep playing the game pre-release. Not outright confirmation though, and I doubt people have datamined the frame data so far.

Don’t see why that would matter. If the normal guard action commands take 6 frames at 60 FPS in the original, make them 3 at 30 FPS. If the superguards take 2 frames ar 60 FPS, make them 1 at 30 FPS, etc.
Superguards in the original are 3 frames at 60fps, which is not cleanly divided by 2.
 
I don't think this is right. What causes the problems in Smash is the fact that it uses bad and outdated delay-based netcode instead of modern rollback netcode. You can go hop on any number of p2p fighting games with much better netcode than Smash and feel the difference immediately
It's not feasible to use any other form of netcode for Smash but delay-based lockstep with the features and gameplay that Smash supports, which go way beyond other fighting games. I know gamers think they know better than Bandai Namco's engineers, but they don't, and it was even tried during the development of Ultimate, and it didn't work.

It's much easier to do it in other fighting games that consist entirely of two characters on a flat plane doing nothing but shuffling back and forth and jumping. In those games, you also don't get to see the negative outcomes of those other kinds of netcode as much either, since they sell 2 million copies if they're lucky, and are played mostly by an older demographic with good wired Internet connections. But if you play a match with a bad connection, then sure, the game doesn't slow down, but the characters start teleporting around instead, and already completed actions get undone and you find out you're already dead from a hit that happened half a second ago.

Also note that Smash is a great example of people complaining about "Nintendo's netcode" being a red flag, because Nintendo doesn't develop Smash.

Nintendo standard libraries for managing those peer to peer connections are, indeed, of very old vintage
This is a perfect example of casual misinformation. You're thinking of NEX, which was Nintendo's old system for managing lobbies and matchmaking, which was licensed and developed out of Quazal Rendez-Vous, a library that has been around since 2003 or 2004. At some point somebody found a string mentioning Windows 98 or something in an NEX binary, and that became shorthand for people to laugh about how old and bad it was. I think you know as well as I do, that's simply an unserious avenue of criticism for a piece of software. To illustrate that nicely, Rendez-Vous's initial release was a few years after the 2000 release of Havok Physics, a library that Nintendo similarly adopted and built on and still uses today, perhaps most notably as the foundation for all the crazy things you can do in Tears of the Kingdom.

But that has nothing to do with netcode, since NEX doesn't handle peer-to-peer gameplay. Nintendo has an in-house library that does that, called Pia, which debuted somewhere between the launch of the 3DS in February 2011 and MK8's release in May 2014. People don't have the "Windows 98" meme to fall back on for this one, so the only criticisms you see of it are "the netcode is bad" as a catch-all devoid of knowledge or merit. Pia works on exactly the same principles that the entire rest of the industry uses for netcode. There's nothing remarkable about it, and nothing lacking. It does what it's supposed to do, and when people have good internet connections, it works great. When they have bad Internet connections, it works badly.

On top of that, Nintendo isn't even using NEX anymore. They created an in-house replacement called NPLN which has picked up all new development and features and is now the recommended library for lobbies and matchmaking. So to the extent that the "vintage" NEX was hampering online play in Nintendo games -- which would have been in the realm of leaderboards, lobby features, and matchmaking methods -- that issue has been addressed.
 
Last edited:
If I may, this is part why I think the idea of paid patches is a non-starter. Not only do these games not seem to have any DRM checks on their resolution, the resolution cap's 4K support is baked in. In some games it would be more work to charge for a patch than simply provide them. Beyond that, of course, charging people to use sections of code they already own a copy of is, in my view, unethical, like on-disk DLC or manipulative in-app purchases.
So you think Nintendo would release patches for free. Also theoretically would it be as easy to do this for older switch titles like breath of the wild.
 
0
I've seen stuff like the Smorg battle lag on Twitter. There's also examples like the Yoshi stampede from Nintendo's own trailers which cause the game to lag. Don't know if I believe DF after Nintendo's own trailers lag lol
do those fights have a ton of enemies on screen? the frame drop that DF recorded had that situation. bandwidth overload induced frame drops wouldn't be unusual
 
If people see a full on remake quality enhancement as "disappointing" beyond a cut to framerate, then their expectations are unrealistic
It's easy to say something isn't disappointing if you pair it with a "beyond the part that is disappointing".
60FPS for Paper Mario? Really guys?
Like its father before it. It is pretty rare I see a game and think "In this case, I'd prefer choppier animation."
I know it's sort of a long shot, but is there a chance the engine is handling its resolution in terms of percentages instead of a hard-coded value? Like right now at a 1080p/docked output, it internally renders at 83% so 900p, while a 720p/portable output, it renders at 88% so 640p. Maybe if it recognizes a higher value from the OS, say 2160p/docked, it could adjust its target to 83% so 1800p, while 1080p/portable could adjust to 88% so 960p? Sort of like an automatic upgrade without needing a dedicated patch?

Truthfully I really don't think this is it, but it's something that came to mind.
I think this would be kind of a silly way to handle Switch games, since no particularly good reason Switch 2 modes should use an equal percentage as Switch 1 modes. But FWIW, Meta Quest games do something like this for different generations of hardware. Quest 1, 2, and 3 have different default rendering resolutions, and if a Quest 1 game is set to use 80% of that default, so will it on Quest 2 or 3 unless it gets a patch giving it more specific instructions for different behavior on different hardware.
Sure, for some games like TTYD you can easily have it run at 4K60 out of the box on T239, but for more demanding games like TotK, Doom Eternal, and MK1, I imagine that more work is needed to get them over the hump, probably through DLSS. That's the kind of thing you can charge for.
If TTYD is running at a similar frame rate and resolution as BOTW/TOTK on Switch, not sure why it would be much easier for it to hit 4K60.
Reason for TTYD being 60 FPS: The graphics allowed this framebudget

Reason for the TTYD remake to be 30 FPS: The graphics did not allow this framebudget

If Nintendo thought performance was more interesting than graphics in this case, do you know what would have happened?
I would agree with them.

But in this case it seems more like bad decisions made for Origami King that we're stuck with now, rather than bad decisions specifically made for TTYD.
TTYD HD getting a 60fps patch on Switch 2 is ultimately pretty meaningless in my eyes because the frame data is already irrevocably changed. Gamers actually don't give a shit about game logic being correct, because they'll praise the hell out of the Mario 64/Ocarina of Time/Majora's Mask recomps even though raising the framerate there does jack shit. All the graphics and animations are simply interpolated.
Just because animations aren't made with infinite steps per second doesn't make a frame rate increase any less valid for those games than a resolution increase. In both cases it's taking something simple and presenting it in clearer fashion. If SM64 Mario's walk animation only had three key frames, rendering the game at 3fps wouldn't be the logical choice.
 
We know where that Zelda demo ended up: Breath of the Wild, using the processing used in that room alone they made it show us an entire world with 'just' a Wii U - and then taking it on the go, the first first party game confirmed and shown for Nintendo Switch. We know with reasonable certainty of the tech demo for their next system, Breath of the Wild, 'U-King-O', where they show how they've gotten that world to load instantly, render at sixty frames per second, and a full 4K resolution, without sacrificing on portability to do so.

Hopefully, all going well, we'll see that in action, and hopefully, all going well, all current owners of Breath of the Wild will have the chance to experience it on the new system by just updating their existing copy.

Another tech demo for this device was the Unreal Engine 5 Matrix demo, which is notable because that was one of those jaw dropping demos on PS5 when it came out, and Nintendo appears poised to launch a device capable of running that same software, that happens to also be portable.

I'm pretty happy with that, even if I do hope they're made public soon.

...oh, right. I guess TECHNICALLY we do have tech demos lol.

But I guess I meant those that are actually, you know, released to the public.
 
do those fights have a ton of enemies on screen? the frame drop that DF recorded had that situation. bandwidth overload induced frame drops wouldn't be unusual
Yes, the Yoshi stampede move summons a ton of yoshis and the Smorg battle is a creature made of like hundreds of the little gas cloud dudes.
 
0
TTYD HD getting a 60fps patch on Switch 2 is ultimately pretty meaningless in my eyes because the frame data is already irrevocably changed. Gamers actually don't give a shit about game logic being correct, because they'll praise the hell out of the Mario 64/Ocarina of Time/Majora's Mask recomps even though raising the framerate there does jack shit. All the graphics and animations are simply interpolated. That in particular is actually a case of placebo without anything backing it up, and then those same people will argue that TTYD NEEDS 60fps? Nah man, people will adjust to the new timings.

Really though, I'm disappointed it is indeed 30fps, but that's because they inevitably made changes to make Action Commands easier. So people arguing 30fps is harder are ultimately wrong - they should be upset that it'll make the game EASIER due to the changes with the frame data. A 60fps patch would still carry over that new frame data and therefore be pretty different from the original regardless.

Hey, so, disabled person here, but changing the window on the timing of this, actually helps people like me. It doesn't make things easier, it makes things farer. Please understand that.
 
It's not feasible to use any other form of netcode for Smash but delay-based lockstep with the features and gameplay that Smash supports, which go way beyond other fighting games. I know gamers think they know better than Bandai Namco's engineers, but they don't, and it was even tried during the development of Ultimate, and it didn't work.

It's much easier to do it in other fighting games that consist entirely of two characters on a flat plane doing nothing but shuffling back and forth and jumping. In those games, you also don't get to see the negative outcomes of those other kinds of netcode as much either, since they sell 2 million copies if they're lucky, and are played mostly by an older demographic with good wired Internet connections. But if you play a match with a bad connection, then sure, the game doesn't slow down, but the characters start teleporting around instead, and already completed actions get undone and you find out you're already dead from a hit that happened half a second ago.

Also note that Smash is a great example of people complaining about "Nintendo's netcode" being a red flag, because Nintendo doesn't develop Smash.


This is a perfect example of casual misinformation. You're thinking of NEX, which was Nintendo's old system for managing lobbies and matchmaking, which was licensed and developed out of Quazal Rendez-Vous, a library that has been around since 2003 or 2004. At some point somebody found a string mentioning Windows 98 or something in an NEX binary, and that became shorthand for people to laugh about how old and bad it was. I think you know as well as I do, that's simply an unserious avenue of criticism for a piece of software. To illustrate that nicely, Rendez-Vous's initial release was a few years after the 2000 release of Havok Psychics, a library that Nintendo similarly adopted and built on and still uses today, perhaps most notably as the foundation for all the crazy things you can do in Tears of the Kingdom.

But that has nothing to do with netcode, since NEX doesn't handle peer-to-peer gameplay. Nintendo has an in-house library that does that, called Pia, which debuted somewhere between the launch of the 3DS in February 2011 and MK8's release in May 2014. People don't have the "Windows 98" meme to fall back on for this one, so the only criticisms you see of it are "the netcode is bad" as a catch-all devoid of knowledge or merit. Pia works on exactly the same principles that the entire rest of the industry uses for netcode. There's nothing remarkable about it, and nothing lacking. It does what it's supposed to do, and when people have good internet connections, it works great. When they have bad Internet connections, it works badly.

On top of that, Nintendo isn't even using NEX anymore. They created an in-house replacement called NPLN which has picked up all new development and features and is now the recommended library for lobbies and matchmaking. So to the extent that the "vintage" NEX was hampering online play in Nintendo games -- which would have been in the realm of leaderboards, lobby features, and matchmaking methods -- that issue has been addressed.

So like why does rollback work in NASB2 on Switch.

I feel like "rollback is simply impossible in four player platform fighters" does need to deal with a four player platform fighter on the Switch with rollback.
 
It's not feasible to use any other form of netcode for Smash but delay-based lockstep with the features and gameplay that Smash supports, which go way beyond other fighting games. I know gamers think they know better than Bandai Namco's engineers, but they don't, and it was even tried during the development of Ultimate, and it didn't work.

It's much easier to do it in other fighting games that consist entirely of two characters on a flat plane doing nothing but shuffling back and forth and jumping. In those games, you also don't get to see the negative outcomes of those other kinds of netcode as much either, since they sell 2 million copies if they're lucky, and are played mostly by an older demographic with good wired Internet connections. But if you play a match with a bad connection, then sure, the game doesn't slow down, but the characters start teleporting around instead, and already completed actions get undone and you find out you're already dead from a hit that happened half a second ago.
Yeah, I think the best case for rollback netcode in the next Smash would for 1v1s only. Maybe not even with items. During COVID, they did update 1v1s specifically to use I think a 60 hz tickrate (not sure if that's the right terminology) instead of the 30 that it was for all modes.
 
Yeah, I think the best case for rollback netcode in the next Smash would for 1v1s only. Maybe not even with items. During COVID, they did update 1v1s specifically to use I think a 60 hz tickrate (not sure if that's the right terminology) instead of the 30 that it was for all modes.

But there are platform fighters with rollback on the Switch?
 
"Nintendo's netcode" (a combination minsnomer/buzzword) has nothing to do with online architecture, because as is often criticized, it is almost always peer-to-peer. The only factor that practically matters then is the reliability of the connections between players, and that is what actually sucks, and causes the problems in games like Smash and Splatoon.

The amount of casual misinformation on this subject is immense and a pet peeve of mine. But is it even an appropriate topic for this thread? It doesn't have anything to do with hardware.
The next round of shipment data can't get here soon enough.
 
This is a perfect example of casual misinformation. You're thinking of NEX, which was Nintendo's old system for managing lobbies and matchmaking, which was licensed and developed out of Quazal Rendez-Vous, a library that has been around since 2003 or 2004.
I wasn't thinking of NEX, I was thinking of Pia, but you're right that I was conflating Pia, which I knew was newer, with NEX's development timeline. I stand thoroughly corrected.
 
There are two things that frustrate people (including me) about online in Nintendo games:
  1. Players' bad Internet connections.
  2. Nintendo's philosophical design decisions.
I already covered the first one in my previous post. It's not a matter of unskilled programmers who are trying and failing to make good netcode. That's just not the reality on the ground, and is a really silly thing to say about a company that publishes games from over a dozen different development teams, both internal and external, which span the globe and have different backgrounds. If all of that were actually true, it would suggest that there's something about doing a project for Nintendo that leads to bad online programming, and that's, er, ridiculous. "Nintendo netcode" is, in fact, fine. When you have a good connection with another player, it works as well as anything else.

The second one also has nothing to do with "netcode" in and of itself. Here are some examples:
  • Nintendo doesn't aggressively filter out poor connections in matchmaking, because they want those players to be able to play, even if that degrades the experience for the other players they have to connect with, and they want quick matchmaking times.
    • This is also why "P2P vs. dedicated servers" is a red herring in this discourse. It's not about the technical solution, it's about the philosophy of wanting everyone to have an equal experience, even if degraded, rather than some players have a flawless experience and others essentially not being able to play or compete at all (which is what I'd choose if I could, since I do have a stable Internet connection, and like everyone on the planet I find it really annoying when I think someone else's lag is affecting the match). If Nintendo did use dedicated servers, that philosophy would still be there.
  • In at least one case where Nintendo did try to do the above -- using strict region- and ping-based matchmaking in Splatoon 2 which prevented players in different regions like the Americas, Europe, and Japan from being matched together -- it was regarded negatively and was undone for Splatoon 3.
    • You can argue whether specifically region segregation was a good choice for Splatoon, especially when players within the same region will often still have poor connections, but it was a decision made to try to address the issues poor connections cause, and the response from US and EU players was that they'd rather have laggy Japanese players than not have them at all.
  • Nintendo still releases games with extremely dumb limitations on online play such as whether and how you can play with friends. Probably the number one recent example is Super Mario Maker 2. It did ultimately get online play with friends, which is a positive example of Nintendo responding to feedback, but it's kind of unfathomable that it wasn't included from the start.
    • I can't even attempt to explain some of these decisions, like the one for SMM2, but the point is, it has nothing to do with "netcode" and related subjects. It's a design decision.
  • The lack of on-system parties, voice chat, etc. is purely and simply a design decision. It's not something where they don't know how to make it work.

The next round of shipment data can't get here soon enough.
India data could be out any day now. Vietnam we can start anticipating around June 4th.
 
India data could be out any day now. Vietnam we can start anticipating around June 4th.
Curious, what are you hoping to see with India data? I figured the T239 validation testing at nvidia India have pretty much ran its course.
 
We know where that Zelda demo ended up: Breath of the Wild, using the processing used in that room alone they made it show us an entire world with 'just' a Wii U - and then taking it on the go, the first first party game confirmed and shown for Nintendo Switch. We know with reasonable certainty of the tech demo for their next system, Breath of the Wild, 'U-King-O', where they show how they've gotten that world to load instantly, render at sixty frames per second, and a full 4K resolution, without sacrificing on portability to do so.
No, the tech-demo didn’t end up as Botw, they’re two separate projects. It did sadly end up as nothing. The closest thing we’ve got is Twilight Princess HD.
 
Other than those extra security features, would it be sensible to assume the practical difference in A78C cores is that they save at least some energy thanks to larger caches vs. A78?
 
In Nintendo's June Direct, will they hint at any games coming to their next-generation console?Although the console itself won't be showcased.

Not likely. You're more likely, at this point, to see a third party announcing Switch 2 games on their own at this point.
 
Not likely. You're more likely, at this point, to see a third party announcing Switch 2 games on their own at this point.
For example, announcing that Prime 4(or some one and third party game) is coming to the Switch, and also to the next-generation console.
 
In Nintendo's June Direct, will they hint at any games coming to their next-generation console?Although the console itself won't be showcased.

Probably won't see anything exclusive to Switch 2 announced. But look forward to loads of cross gen games. Metroid Prime 4 is now in its final visual polish alongside framerate and frame pacing tests and fixes. So I'd bet that's going to get shown. This would make sense because it'll likely be a showcase for Switch 2 "Look, at it run at 4K/60!"

If there's any time to show an OOT remake it's now, again a good cross platform game to showcase for next year's launch.

But that's all I currently suspect, I'm hoping Mario Kart 10, and a new Mario will be at least two games we find out about when the console is announced as exclusives to the system at launch. If i was to hamper a guess, launching with Mario Kart 10 in March and a new Mario in June. Just to give some space in-between and Mario Kart usually sells even harder than even a mainline Mario, so it makes sense IMO to launch with a hard hitting software seller first and continue momentum with a mainline Mario a few months later.

Anyway having an ADHD waffle. I'll stop.
 
In Nintendo's June Direct, will they mention at any games coming to their next-generation console?Although the console itself won't be showcased.
It makes no sense to announce games for a console that hasn't been announced yet. They specifically mentioned there will be no news about the next gen console in the June direct, only that they will have something to share within this fiscal year, which ends on March 31st 2025. So we're looking at an announcement between July 2024 and March 2025, pretty wide window if you ask me.
 
Skyrim dropping on Switch in 2017 is nowhere near GTA VI releasing on Switch 2 day one or even if the port came years down the line.

GTA VI on Switch 2 would be up there with the most impressive ports of all time-- bookmark this if it happens. If that drops on Switch 2, there should be almost no excuse for other major 3rd parties to drop on Switch 2 barring some logistical reasons and resource constraints. But on the tech side? There would be little to no reasons. Now... I am not expecting GTA VI despite whatever rumors pop up. I will believe it when I see it.

To further this discussion; Witcher 3 being on Switch was far more impressive than Skyrim. Skyrim was a PS3/360 game that didn't need much VRAM on top of the Switch having better tech than PS3/360. So it being on Switch was a no brainer. That's why people are mentioning Elden Ring and BG3, more so Elden Ring since BG3 wasn't released on last-gen machines but I think there's a solid chance we get it because Divinity OS2 came to Switch and it has the RAM to do it. Even though the Switch is a couple of generations ahead, Elden Ring was made with PS4/XB1 in mind, so this is why people bring it up as the Skyrim of Switch 2. Skyrim was made with PS3/360 in scope. GTA VI doesn't fall anywhere in this equation and is less likely to happen.

The expectation should be PS4-level games with some PS5-level games, expecting them all to drop on Switch 2 would be a mistake. I am certain it will get more modern-day AAA games than the Switch for sure though. Like I said, bookmark this, if GTA VI releases on Switch 2. I will fly to Rockstar headquarters and shake Mr. Rockstar's hand.
People including tom henderson heard that switch 2 will have basically same support of 3p aaa as ps and xbox, it should tells everything about hardware imo
 
Other than those extra security features, would it be sensible to assume the practical difference in A78C cores is that they save at least some energy thanks to larger caches vs. A78?
There's that, yes.
Also inter core communication should be better/more efficient between cores #0-3 and #4-7.
If it were two clusters of A78, going from one cluster to the other involves exiting out the starting cluster, go through the interconnect/fabric, then enter into the destination cluster. All 8 being within the same cluster cuts a lot of that time/energy out.

It's kind of annoying that I can't find tests of multiple clusters of ARM cores (not counting the big server chips). That said, if you want an example of the concept over in x86 space, just look at the difference between Zen 2 and Zen 3. In addition to expanding the core to push IPC up, Zen 3 also changed their cluster size from 4 cores to 8 cores. And as we know now, Zen 3 was a pretty nice uplift over 2.
 
It's not feasible to use any other form of netcode for Smash but delay-based lockstep with the features and gameplay that Smash supports, which go way beyond other fighting games. I know gamers think they know better than Bandai Namco's engineers, but they don't, and it was even tried during the development of Ultimate, and it didn't work.

It's much easier to do it in other fighting games that consist entirely of two characters on a flat plane doing nothing but shuffling back and forth and jumping. In those games, you also don't get to see the negative outcomes of those other kinds of netcode as much either, since they sell 2 million copies if they're lucky, and are played mostly by an older demographic with good wired Internet connections. But if you play a match with a bad connection, then sure, the game doesn't slow down, but the characters start teleporting around instead, and already completed actions get undone and you find out you're already dead from a hit that happened half a second ago.
I think there's something being missed here because:
A. There are Smash clones that have rollback net code.
B. There are other fighting games that have rollback, that specifically do not support it on Switch.

That would naturally lead one to believe that it's a problem with Nintendo's network, the Switch hardware, or both.
 
do those fights have a ton of enemies on screen? the frame drop that DF recorded had that situation. bandwidth overload induced frame drops wouldn't be unusual
Yeah, those both involve large crowds. The original game really liked throwing around triple digit numbers of entities on screen at once, presumably as a way to show off the GCN's power. Even just the normal battle theater gets pretty big at higher levels. On Switch, the asset quality compromises from doing that are seemingly mostly gone, but if DF's analysis is any indication, the framerate can be a little less than perfect during some of those bits.
 
I've seen stuff like the Smorg battle lag on Twitter. There's also examples like the Yoshi stampede from Nintendo's own trailers which cause the game to lag. Don't know if I believe DF after Nintendo's own trailers lag lol
Usually Nintendo trailer aren't the most consistent in terms of quality, sometimes they suffer of bad frame pacing or artifacts because of lower bitrates.
For example Side Order trailer had many lag spikes, especially during the "fake loading" with the matrix like code in the background, but in the end I never encountered that type of lags.
 
I think there's something being missed here because:
A. There are Smash clones that have rollback net code.
B. There are other fighting games that have rollback, that specifically do not support it on Switch.

That would naturally lead one to believe that it's a problem with Nintendo's network, the Switch hardware, or both.
The various platform networks play a much smaller role in the overall online gameplay experience than most seem to think. The networks are mostly there to establish identity, and provide certain common pieces of functionality that need or benefit from being handled in a game agnostic way, like friends lists, invites, voice chat (sometimes), etc. The actual gameplay parts mostly come down to game-specific implementation details.

The specific phenomenon of games lacking rollback specifically on Switch is broadly believed to be a CPU bottleneck issue, which seems to be a plausible enough theory.
 
The various platform networks play a much smaller role in the overall online gameplay experience than most seem to think. The networks are mostly there to establish identity, and provide certain common pieces of functionality that need or benefit from being handled in a game agnostic way, like friends lists, invites, voice chat (sometimes), etc. The actual gameplay parts mostly come down to game-specific implementation details.

The specific phenomenon of games lacking rollback specifically on Switch is broadly believed to be a CPU bottleneck issue, which seems to be a plausible enough theory.
That would be my natural assumption as well based on my basic understanding of how rollback works. It makes the most sense that it could be a CPU issue.

Which then circling back to the original notion on whether Smash could have rollback, it seems like it should be very feasible on Switch 2.
 
That would be my natural assumption as well based on my basic understanding of how rollback works. It makes the most sense that it could be a CPU issue.

Which then circling back to the original notion on whether Smash could have rollback, it seems like it should be very feasible on Switch 2.
Most fighting games with rollback are way simpler to synchronize than Smash.
 
A. There are Smash clones that have rollback net code.
There are no games that do what Smash does except Smash. Nothing even close. The "Smash clones" that exist are about on the level of Smash 64.

That would naturally lead one to believe that it's a problem with Nintendo's network, the Switch hardware, or both.
There's no such thing as "Nintendo's network." Smash is P2P. All fighting games are P2P.

Which then circling back to the original notion on whether Smash could have rollback, it seems like it should be very feasible on Switch 2.
It can't and it won't.
 
There are no games that do what Smash does except Smash. Nothing even close. The "Smash clones" that exist are about on the level of Smash 64.
I mean I'm not an expert on other Smash like games out there, but Multiversus and Nickelodeon 2 appear to be sufficiently complex.

Which, beyond the above, I think you're looking at it pretty narrowly when it could certainly be implemented for specific modes(i.e. 1v1), if needed.
 

Hopefully, this won't affect their leak business.

ign?
tenor.gif
 
It's not feasible to use any other form of netcode for Smash but delay-based lockstep with the features and gameplay that Smash supports, which go way beyond other fighting games. I know gamers think they know better than Bandai Namco's engineers, but they don't, and it was even tried during the development of Ultimate, and it didn't work.

It's much easier to do it in other fighting games that consist entirely of two characters on a flat plane doing nothing but shuffling back and forth and jumping. In those games, you also don't get to see the negative outcomes of those other kinds of netcode as much either, since they sell 2 million copies if they're lucky, and are played mostly by an older demographic with good wired Internet connections. But if you play a match with a bad connection, then sure, the game doesn't slow down, but the characters start teleporting around instead, and already completed actions get undone and you find out you're already dead from a hit that happened half a second ago.

Also note that Smash is a great example of people complaining about "Nintendo's netcode" being a red flag, because Nintendo doesn't develop Smash.


This is a perfect example of casual misinformation. You're thinking of NEX, which was Nintendo's old system for managing lobbies and matchmaking, which was licensed and developed out of Quazal Rendez-Vous, a library that has been around since 2003 or 2004. At some point somebody found a string mentioning Windows 98 or something in an NEX binary, and that became shorthand for people to laugh about how old and bad it was. I think you know as well as I do, that's simply an unserious avenue of criticism for a piece of software. To illustrate that nicely, Rendez-Vous's initial release was a few years after the 2000 release of Havok Physics, a library that Nintendo similarly adopted and built on and still uses today, perhaps most notably as the foundation for all the crazy things you can do in Tears of the Kingdom.

But that has nothing to do with netcode, since NEX doesn't handle peer-to-peer gameplay. Nintendo has an in-house library that does that, called Pia, which debuted somewhere between the launch of the 3DS in February 2011 and MK8's release in May 2014. People don't have the "Windows 98" meme to fall back on for this one, so the only criticisms you see of it are "the netcode is bad" as a catch-all devoid of knowledge or merit. Pia works on exactly the same principles that the entire rest of the industry uses for netcode. There's nothing remarkable about it, and nothing lacking. It does what it's supposed to do, and when people have good internet connections, it works great. When they have bad Internet connections, it works badly.

On top of that, Nintendo isn't even using NEX anymore. They created an in-house replacement called NPLN which has picked up all new development and features and is now the recommended library for lobbies and matchmaking. So to the extent that the "vintage" NEX was hampering online play in Nintendo games -- which would have been in the realm of leaderboards, lobby features, and matchmaking methods -- that issue has been addressed.
So what is it exactly? I played Smash Bros locally and i still had a freeze or two in the game. Is it the network being capped or what? I do feel like the wi-fi chip plays a huge part in it as well.
 
0
It can't and it won't.
I got into an argument with someone and I was saying that Smash likely has to send not only characters data frame, but stage hazard frame, and item generation. Someone claims that's not true and that there is a seed that generates when an item show and the seed is sent to all players at the start of the match. How true is that?

Edit: so the argument was that seeding a random generated item spawn is done and only button input is done.

Edit: I am thinking from a algorithm perspective. The guy made sense initially. Items could generate at the same time if you made random generated algorithm and send the same seed out.

But what about boxed items. Not necessarily the items itself but what about where the items land When you open the box or barrel.

What about trophies. The data has to be sent to all system to know information as who to target and where is the location of the trophy assist.
 
Last edited:
It's much easier to do it in other fighting games that consist entirely of two characters on a flat plane doing nothing but shuffling back and forth and jumping. In those games, you also don't get to see the negative outcomes of those other kinds of netcode as much either, since they sell 2 million copies if they're lucky, and are played mostly by an older demographic with good wired Internet connections. But if you play a match with a bad connection, then sure, the game doesn't slow down, but the characters start teleporting around instead, and already completed actions get undone and you find out you're already dead from a hit that happened half a second ago.

Also note that Smash is a great example of people complaining about "Nintendo's netcode" being a red flag, because Nintendo doesn't develop Smash.


This is a perfect example of casual misinformation. You're thinking of NEX, which was Nintendo's old system for managing lobbies and matchmaking, which was licensed and developed out of Quazal Rendez-Vous, a library that has been around since 2003 or 2004. At some point somebody found a string mentioning Windows 98 or something in an NEX binary, and that became shorthand for people to laugh about how old and bad it was. I think you know as well as I do, that's simply an unserious avenue of criticism for a piece of software. To illustrate that nicely, Rendez-Vous's initial release was a few years after the 2000 release of Havok Physics, a library that Nintendo similarly adopted and built on and still uses today, perhaps most notably as the foundation for all the crazy things you can do in Tears of the Kingdom.

But that has nothing to do with netcode, since NEX doesn't handle peer-to-peer gameplay. Nintendo has an in-house library that does that, called Pia, which debuted somewhere between the launch of the 3DS in February 2011 and MK8's release in May 2014. People don't have the "Windows 98" meme to fall back on for this one, so the only criticisms you see of it are "the netcode is bad" as a catch-all devoid of knowledge or merit. Pia works on exactly the same principles that the entire rest of the industry uses for netcode. There's nothing remarkable about it, and nothing lacking. It does what it's supposed to do, and when people have good internet connections, it works great. When they have bad Internet connections, it works badly.

On top of that, Nintendo isn't even using NEX anymore. They created an in-house replacement called NPLN which has picked up all new development and features and is now the recommended library for lobbies and matchmaking. So to the extent that the "vintage" NEX was hampering online play in Nintendo games -- which would have been in the realm of leaderboards, lobby features, and matchmaking methods -- that issue has been addressed.
So are you saying that Nintendo can't/won't do anything to improve online play on Switch 2? It will be the same much worse online functionality compared to their console rivals in the future as well?
 
So are you saying that Nintendo can't/won't do anything to improve online play on Switch 2? It will be the same much worse online functionality compared to their console rivals in the future as well?
Man there's so many factor to this. How far are you and your peers are from your router. Are you connected to an ethernet. How far are you from other players.
 
0
Please read this new, consolidated staff post before posting.

Furthermore, according to this follow-up post, all off-topic chat will be moderated.
Last edited by a moderator:


Back
Top Bottom