-
Ever wanted an RSS feed of all your favorite gaming news sites? Go check out our new Gaming Headlines feed! Read more about it here.
GGPO (rollback netcode) SDK now Open Source under MIT License
- Thread starter Murlin
- Start date
That's interesting stuff for sure, we'll have to see where it goes.
Max's "not invented here" hypothesis was always complete nonsense, as all the major Japanese fighting game developers (Capcom, SNK, Namco, ArcSys) use UE4, a western-made middleware. And a western developer, NRS, decided to develop their own rollback netcode rather than use GGPO.
Max's "not invented here" hypothesis was always complete nonsense, as all the major Japanese fighting game developers (Capcom, SNK, Namco, ArcSys) use UE4, a western-made middleware. And a western developer, NRS, decided to develop their own rollback netcode rather than use GGPO.
Really? Well nevermind then, that's great. I knew MK11 had rollback, but not that it was GGPO.
All that's left is for Japanese FG developers to develop their games from the ground up to allow for the technology like NRS did.
Oh, well nvm then. This is what I thought it was.
Then again, I bring up the matter of whether GGPO is compatible with modern FGs that do not use sprites.
No, it does not. They developed their own rollback netcode and it works very differently from GGPO.
Oh damn you're right. I was sure of that because of the GGPO tease by Boon back in the MKX days before the XL version.
I think it's less sprite based and more whether it's 2D based. 3D just adds much more complexity to the range of possible inputs and states
Yup. I think the main issue would be translation. If French Bread can pull it off then it would be easier for other japanese devs
Isn't that one and the same? Or do you mean on a 2D plane? Because what I'm wondering is if a game like FighterZ, Guilty Gear Xrd, Street Fighter V, or Mortal Kombat 11 could just use GGPO, despite the fact that they're high quality 3D models and effects with their backgrounds but on a 2D plane.
This really isn't that complicated.
- Some sort of GGPO style system can likely be implemented in most games/engines.
- Some are going to take significantly more time, effort, and cost than others.
- More developers should prioritize that time, effort, and cost because it is worth it.
- If they don't, it can be for a shitload of reasons and pretty low on that list is laziness or incompetence.
Well, then there's a significant difference between "GGPO style system" and GGPO itself. A lot of people are seeing this news and acting like GGPO itself can just be glued into any modern FG, when it's likely that developers instead need to create their own rollback systems like NRS with MK11 or Capcom with SFV to accomodate modern, non-sprite based FGs. And in that case, GGPO becoming open source is of no relevance.
Couldn't devs at least study the SDK and documentation to figure out how rollback works? From what I've read in other threads and in Max's video, many Japanese devs don't even know wtf rollback-based netcode is, and once somebody translates the documentation to Japanese, that'll go a long way towards more Japanese devs creating their own solutions that can be more easily integrated into their own engines.
that's great...
but somehow I get the feeling that nobody gives a crap, especially not capcom and big companies
but somehow I get the feeling that nobody gives a crap, especially not capcom and big companies
I haven't watched Max's video, but I'd find it really hard to believe Japan just doesn't know what rollback netcode is. They still make high profile fighting games. ASW has talked about being aware of GGPO in the past. Capcom uses rollback. I think it's more of a matter of them not wanting to do so or put in the resources for its implementation for whatever reason, and less about the technical know-how not being there.
The information GGPO stores, rewinds, and loads doesn't really seem impacted by whether the renderer is 3d or 2d as that's separate from the entire process. The only thing that truly matters is whether the game state is entirely deterministic, which most fighting games already are or could be.
Yes exactly.
The only things that should be sent across the wire constantly are the button inputs that happen on each frame.
Everything else should be rendered independently by each player's machine. Why on earth would "3D vs 2D" complicate any of that?
It's also worth noting that even if GGPO were somehow ill-equipped to handle a 3D game, which I obviously don't believe is the case--storing the position and state of entities in 3D space isn't that drastically different from storing the position and state of an entity in 2D space and as mentioned in the GGPO thread a few days ago, a lot of the "theory" behind what GGPO does has been done in 3D games, FPSs, TPSs, etc. for 15-20 years with client-side prediction and lag compensation dating all the way back to zeroping mods made by the community for Quake and UT (oh, also in RTSs). Even if GGPO were somehow ill-equipped to handle a 3D game, being open source is a great means to solving that by opening itself up to a potentially larger community of developers to improve.
Last edited:
Yep. The mkxl code refactor from what I understand was to make some cpu time leeway to accomodate how rollback works (keeping, predicting and rollbacking game states) which has nothing to do with a 3d renderer in itself.
People are working on translating the documentation to Japanese:
The remaining issue is that it is difficult to convince the bigwigs at ArcSys/Namco/SNK etc. that they should invest time (and/or money) into implementing a different form of netcode when they can simply use what they always have, especially when recently there wasn't even a functioning website for GGPO and the only way to access it was to essentially email a guy.
The remaining issue is that it is difficult to convince the bigwigs at ArcSys/Namco/SNK etc. that they should invest time (and/or money) into implementing a different form of netcode when they can simply use what they always have, especially when recently there wasn't even a functioning website for GGPO and the only way to access it was to essentially email a guy.
For people saying that it's too late for certain games like Guilty Gear 2020, remember that MKX was literally patched to have GGPO midway through its life.
(GGPO) Netcode has nothing to do with what a game looks like or how many dimensions a game uses. The only meaningful difference between 2D and 3D in this case is one extra dimension which will use up a little bit more data in the snapshots/keyframes, but that is negligible.
That wasn't GGPO. And in any case, there's a big difference in terms of budget, know-how, and priority. ArcSys have never once indicated that they've had any serious interest in rollback netcode beyond "we'll look into it".
There's no documentation (currently) in Japanese.
NRS switching to rollback was a big undertaking.
FWIW, Sajam has been working on a big article about netcode and why rollback matters that should be out "in less than 10 days" too.
Speaking of NRS netcode, here's a good (but very lengthy) GDC talk with Michael Stallone on how they changed their netcode to rollback halfway through MKX lifespan. It's a good watch and goes really deep into the whole thing and definitely applies to other games switching over to rollback (and likely actual GGPO).
Michael Stallone is such a treasure. I loved this GDC talk.
well they did also give us a "we're tired of hearing about it"...
Just real talk but the best way to do this is to spread the word about when we do add rollback and support those releases. Our recent Code Mystics releases (KOF 97 and SamSho V Special) on Steam and PS4 being the only real examples for SNK.
They have almost everything that a FGC user could want.
- Steam version
- rollback netcode
- sprite graphics
- affordable price
And last time I looked they both only have around 100-300 reviews on steam. I don't deal with sales numbers now that I am on the dev side, but that doesn't really seem like a runaway success we can use to convince the decision makers.
Despite Code Mystics going in and adding rollback code in response to negative reviews and feedback, KOF 97 wasn't even able to climb back up to a positive review status.
It would make more sense to believe that given Japan's relatively high internet quality, relatively small landmass, and relatively high population density that Japanese developers simply see their net code working perfectly well there and are not overly concerned with optimizations for those overseas with awful internet and large distances between players.
There are many reasons why these things never pick up again (main one being that people have just moved in already) but yes you are right in that we need to talk about it more when they do get it right and put our money where our mouths are.
Ideally these versions would all release on the different platforms at or around the same time, and have rollback netcode from the start.
Those are old games, though. As an enthusiast in the genre, I'm not interested in picking up and playing old games from before my time. Now when we look at modern ones, MK11 is greatly successful and everyone always talks about how good that game's online is, which uses rollback.
(Think I got these right)
It is just a pet peeve of mine, especially as a person who put actual work into those projects.
There is a lot of talk in the FGC about raising each other up, but when it comes to stuff like this the conversation always turns to the bad points and just outright bashing.
Sajam's recent video about netcode spends a lot more time complaining about the negative aspects, when I feel like it would be more effective to try to 'raise up' these bright spots in the industry.
That's a pretty narrow-minded view of the genre, so I don't even know what to say.
If I met a movie buff who claimed to be an enthusiast but refused to watch anything 'before their time' I would probably make a note to never listen to their opinion on movies.
MK11 is successful but there is a lot going on there besides just rollback implementation. They probably have the largest development AND marketing budget in the entire genre and their production values and marketing power are probably the real key to their success sales-wise.
Even just looking at Steam achievement data, only 12% of people seem to have completed the basic online achievements.
(that data might be wonky because the game is blocked here, but it seems to at least show me achievement data)
I still think good netcode is important, but there isn't much data out there to suggest it's a huge draw for the casual audiences and therefore leads to millions of units of sales.
Yes, thank you! I was not going to go that far in my shilling. ;)
I play fighting games very competitively, I'm not in it to be a historian. If I can't get into it like that, where I'm going to actual tournaments with hopes to perform at a high level, then I likely won't be very invested in playing an FG. So when it comes to older games, their scenes are weaker compared to newer games, they look worse (and aesthetics matter very much to me in the genre), and they've also been around for years so it's much harder to get into them on a competitive level. I much rather invest my time in scenes that are active and fresh.
Sure, I don't know what the opportunity cost of rollback netcode implementation would be compared to delay based. Is it actually harder to implement rollback than delay based? But either way, what's basically the only high profile fighting game with great online constantly gets talked about in non-casual circles about how its online is so good, and its online scene is healthy because of it, which attracts more people. This would happen with any new fighting game that had a great online experience.
I mean, I am not asking you to becoming a pro KOF '97 player and take on Xiaohai.
Just pointing out that spending $10 on a title and putting your wallet where your mouth is in relation to rollback netcode support is more effective than complaining on the internet.
If a $10 investment is a bridge too far, then I guess that's that. lol
I have a lot of doubts that supporting an older game that happens to have rollback would be a meaningful message at all when it comes to implementing rollback in a modern game.
I figured. While I have your ear on this topic. Do you know of any awareness around this same idea for single player stuff? For example the run-ahead stuff libretro is messing with (similar idea to GGPO, just doing it local only) for something like metal slug would be pretty great.
You would be wrong. If there was actually a successful campaign pushing these releases to the point where we saw our sales increase by a large margin, that is actionable data people in the company can use to prove there is tangible demand. It wouldn't be some magic bullet, but it's going to send a more meaningful message than bashing our modern games netcode on the internet.
In no reality would supporting any of our rollback releases be considered a negative.
Nope, that is way out of my wheelhouse. I don't specialize in coding and only have a laymens understanding of most of it thanks to some hobbyist stuff I did in high school.
So should I not support the Samurai Shodown NeoGeo Collection if it turns out to have similar netcode issues Street Fighter 30th has and just support Samurai Shodown V Special instead? I am pretty patiently waiting out on the PC version of SamSho and while I think V Special is excellent (love that online especially!), the first two games and the new one are just way more appealing to me.
I am a pretty massive SNK fan so I already have the majority of their releases on Steam regardless of their netcode or port quality. I feel like there is nothing someone like me can do to push SNK to care about rollback on their more recent titles other than say I would play them way longer online.
I am a pretty massive SNK fan so I already have the majority of their releases on Steam regardless of their netcode or port quality. I feel like there is nothing someone like me can do to push SNK to care about rollback on their more recent titles other than say I would play them way longer online.
Well, there's not really anything indicating that the company would look at the large number of sales/support and pin it on the netcode, right? What would make them think it would be that and not any number of other elements? Unless they came out and openly said: "We're closely looking at the sales of these games to consider implementing rollback netcode in future games," I wouldn't just assume that. Especially not when it comes to a Japanese company.
And I certainly won't buy a game I won't play out of charity, simply for the possibility. I'll be real into buying a new game if it has great online.
Correlation is, too. We don't have reason to think that these companies would pin success on the right things. Or that success would even be because of the netcode and not something else. Like MistaTwo already said, there isn't much data to indicate that good netcode even has an impact on casual sales, even in MK11's case.
On the other hand, do you understand that many of us felt like it was a slap in the face for SamSho V Special to be updated on PS4 and release on Steam with rollback, and then have Samurai Shodown 2019 release with delay-based netcode a week and a half later? Different project, different timeline, different developers, different scale, yes, we know. But it was such a bad look especially when we had heard nothing about the netplay before launch. You can at least appreciate how disappointing that was?
I don't generally support buying a game out of some principle but man, Them's Fightin Herds does way too much good stuff to not warrant that.
Hows Punch Planet? i've had Skullgirls for a while and bought/gifted a few copies of TFH.
