I feel its extremely important to any fighter nowadays and that a lot of company’s don’t seem understand this. It’s even more essential to games being released in countries without an arcade scene
I’m making this thread after being disappointed with netcode after netcode after knowing it can be done better.
TLDR at the bottom
SF4 did it well in 09. Blazblue did it even better in 09. Seems like everything since then has been downhill. Tekken 6 online still isn’t playable post patch. SF3:OE. UMvC3 still isn’t on par. I didn’t play MK9 online but I’ve read nothing positive about its netplay. KoF 13 was just released with a not so acceptable netcode. Though I’m seriously hoping they can patch it! Because its really close to good.
When a game doesn’t have a good netcode its hard to have a scene for it here in the U.S. For example. Tekken is my 1st love when it comes to fighters. I was excited about T6 literally for years. I’d finally be able to play with people. Of course that dream was crushed when I got it. A game I’d anticapated for so long ended up collecting dust very quickly. Without a strong netcode there was no chance to take my game to the next level or anything close. Combo’s & punishes couldn’t be performed & you’re not even playing the real game anymore with the netcode destroying the meta.
Even if a game isn’t that popular (every game not made by capcom in the U.S. Lol) it can have a competitive scene that lasts if the netcode is good. Look at Blazblue. It’s a much smaller scene but it still lasts because of the players have access to a way to improve.
The tournaments are affected too not just in turnout but by the winners. I remember in the beginning of SF4 when you only needed to know 2 things to determine who would win. 1. Is daigo there? 2. Is JWong there? Lately you really have no idea who will win the next SF4 tourney. Players like Wolfkrone & Latiff are largely a product of SF4’s solid online play.
TLDR: Why are netcodes getting worse instead of better? All it does it hurt the game’s scene. Can’t see why any company would release a game without its netcode being at least as good as SF4’s. WIth the large amount of new fighters incoming I’m scared for us online warrior’s who would like to become offline warriors.
didnt ponder already explained that the excuse of using 3d as a reason of why a rollback netcode couldnt be used is pure bullcrap, also the dbz game made by namco has very detailed 3d models and use GGPO
2nd BB netcode its good, not the greatest, but its acually good, dunno why you put it as an example
Notice that you ignored the word “heavy”.
Rollback net code is heavy already and the current consoles are more limited in power/memory etc. than you might think. So the game needs to be made with that in mind, but those games aren’t. So your console can’t handle both running the game AND a rollback netcode.
Arcsys games for me (and for all the other locals) are unplayable even vs someone across the street so I can’t really give credit to anything less than GGPO at this point, since that’s the only solid experience I’ve had until now.
In Japan they can also play PS2 SNK games with little lag but mentioning that doesn’t get us anywhere…
Rollback net code proved to be the best route for the future of online fighting games and we should not compromise on anything inferior.
I always thought that online play wasn’t so critically needed but maybe I need to rethink. Isn’t one of the reason FPS are very popular due to consistently playable online play?
Right now I must admit netcode for most game aren’t good but at the same time most people don’t have their set to be optimized ( so I take bad netcode with grain of salt). I think we really need to address the issue of people not optimizing their set ups.
1.) As stated before, depending on the title some methods of netcode have to be considered when designing/building the title early on. You don’t just throw rollback-based netcode into a 3D title that is juggling so many non-trivial assets on rapidfire. You can also inherently make a FG more netplay-friendly by slowing the pace down(not really preferred) or have mechanics like Blazblue 5f input repeat to help smooth things out, which is probably something better decided early on in development rather than later.
2.) “Good netcode” is extremely subjective because it plays second fiddle to the actual issue behind lag; connections and local network infrastructure. There are plenty of people who will feel even games with decent netcode have garbage netcode for this reason. For Americans, the actual answer is to get the cable companies off their asses and upgrade the US infrastructure instead of milking everyone for easy $$ on whatever we’ve got(yeah, good luck with that…). That’s also why Japanese developers likely don’t care much about it as much like tataki said.
3.) AFAIK code that solves the actual problem (turning 200ms+ ping matches into something playable) is something that would be costly and time-consuming to develop; pretty much impossible for smaller groups/publishers, and naturally never 100% reliable(because again, the issue is largely the connection itself). So it’s often an outsourced task handled by some other company rather than the publisher/developer themselves making a shot at it.
4.) GGPO is the silver bullet for competitive players here, but I speculate publishers want to avoid exposing the average player to what rollback-based netcode can do for extremely laggy connections, which make the game look like it’s broken or poorly made. That’s probably one reason why Capcom hasn’t used it for one of their major 2.5D titles yet. I think everyone is used to the “online play experience may vary” disclaimer and have had shitty online matches in some form, but explaining to a reviewer that the other guy teleported halfway across the screen or he never actually won the match because the netcode is for competitive players would be… well, IDK about you, but* I* wouldn’t want to be the guy selling that.
IIRC Skullgirls is going to make a better attempt at managing GGPO settings from match-to-match than 3S:OE did, so seeing how that title turns out on the netcode front might persuade more FG devs to consider using it. Hopefully, anyway.
Netcodes aren’t exactly getting worse its just that Japanese developers either suck at making them or that they only care that Japan’s godly infrastructure can run it even though they have world markets that they want to sell to. These companies use these super ancient ass netcodes that are just plain obsolete in the face of rollback netcodes. I hate how guys like Katsuhiro Harada of Namco say things like GGPO isnt a “magic solution” like its an excuse to put out absolute shit netcode. Ponder already debunked the 3D thing years ago. It comes down to having the game’s source code and how well you code around the stages and junk so it doesn’t have to run off the CPU and burn the computer to hell. If don’t know how to utilize it fully you are going to crazy rollbacks and shit is gonna heat up. And even if you cant import a foreign netcode like GGPO into the game then try to make a solid homebased rollback netcode for the game so there can more control over how it works.
Rollback netcodes are the only thing that take into the account that not every country has a great infrastructure and tries to work on keeping the input delay to a minimum regardless of the distance. The only Company that seems to understand that is Arc System Works. They even thanked Ponder for the idea of a good netcode to use way back before BB:CT came out on console. So their games and the ones that they pick up like Arcana Heart 3, (I think they picked up Chaos Code too but I cant remember) are the only ones even worth caring about online wise. Far as I’m concerned and have experienced with, the fighting games that don’t at least use a rollback netcode have and will have shitty online play if you don’t live in countries like South Korea and Japan. Fuck SF4, fuck both Marvel 3s, fuck KOF 13, fuck Tekken 6 and fuck MK9. Playing online shouldn’t be “playable” it should be “excellent”
Good Netcode is crazy important because that’s what keep games relevant after they come out. KOF 13 is in danger of dying out quickly because the online is so bad, which hurts and scares the hell out of some people because they dont want a game that they really like to die out because no one can play each other outside of majors. Plus online play is the future of fighting games. These Japanese companies need to step into the 21st century and stop using netcode from the 90s
From my experience GGPO does it’s job in 3SOE. It’s just that the matchmaking is so horrible that you can’t find the people you are supposed to play with in the 1st place. But when I played locals from my friends list it was as great as playing them on ponder’s PC client and I’m expecting SkullGirls to be the same.
Bull, SF4 has the bare fucking minimum acceptable netcode for a fighting game. To call it “well” is a fucking travesty with the resources Capcom has. To compliment it is insulting and continuing to get use “par for the course” in AAA budget games. It is UNACCEPTABLE that Capcom cannot do better. They should be going above and beyond when indie titles and small dev studios are kicking the shit out of their games right now.
I’ve not believed the ‘3d = slow netcode’ excuse ever. The “heavy” graphics often refer to the fillrate it takes for the effects stuff to blend and MVC3 has a lot of it.
It’s a GPU. It can take it. What netcode doesn’t like is the 60hz sync refresh, and considering the PS3/360 are oversplurged in CPU power and THREADS I find the bad netcode/3d graphics correlation extremely wrong.
All netcode has to do is push boxes, positions, moves as fast as it can do so in sync. It doesn’t push polygons, bones, textures, facial grimaces, titties, etc. all that is irrelevant to packets. It would push the same stuff a 2D fighter would do, and ironically the 2D fighters are still technically 3D, using the GPU to render a scene, chars in front, etc. Just with a lot less polygons and no bone transform, that’s all.
You’re missing the forest for the trees there. Of course all of that information isn’t sent across during netcode, the problem is that there is so much information being moved in and out of client-side memory during the normal course of gameplay, that having to set aside an enormous additional portion to accommodate potential rollbacks isn’t feasible with the limited RAM a console has to work with.
There are plenty of glitch videos that demonstrate what happens in MvC3/UMvC3 when the game becomes starved for RAM.
I also think developers want to figure out their own netcodes via trial and error. GGPO is awesome right now, but I don’t think they want to be giving Cannon royalties for an outdated product years from now.
Totally OT but it always bugs me, why do people constantly say only capcom does well over here? Especially over several games, Namco has been the dominant developer of fighting games for the entire last decade.
I guess you could say I’m part of the problem, since I am just happy for having online play in the first place. It helps that I moved to stuff like Blazblue and MvC3 (Which has abyssmal netcode against some connections, yet I never really cared) from single-player games and Smash Bros.
I do think bad infrastructure isn’t going to change in the near-future, so game companies need to work around the problem instead of working as if it didn’t exist.
It’s because SRK is dedicated to Street Fighter, as well as other Capcom games, so the majority of discussion is around them. It helps that this site is thriving.
Still, we barely heard about Tekken at all here. The Tekken fanbase on SRK is smaller, but it also isn’t **visible. **Indie games like Skullgirls and even old classics like Third Strike have much more talkative fans in the general fighting game discussion and on the front page.