I tend to disagree because if you give the mvc3 net code to ANY other game, no one would play that shit online because its so bad. Input delay fucks everything up regardless of the game being played. Take away the CoD examples and use counter strike 1.6 from 2000, Counter-Strike (2000) – 4.2 Million, not that much steam behind the game. In fact the only reason I played it was because someone told me it ran incredibly well online so I tried it out.
the FPS scene does well online because for 1, they have a really good net code and 2, their tournament scene can be accessed through the internet. Not only is playing online for that game enjoyable when everyone is playing properly, people run leagues through the internet making it that much more competitive.
Its not the sales approach, its about what content the game has IMO. You also have to remember that FPS games weren’t popular @ one point, fighters ruled pretty hard. Its just when the internet became part of the successful equation for a popular game, fighting game developers said fuck it we have arcades and the FPS people honed in on the net play aspect of games. That was the turning point and its why the FPS market is 100x ahead of the fighting game market.
I personally don’t think Capcom pays enough attention to the future, only to the now. Otherwise after cvs2\HF on xbox 1, they would of realized that the FPS market was making a killing by having successful online components.
First off, input delay is in all fighting games, not just MvC3, and for a reason. You cant use the same prediction techniques as other games, and there is a lot of additional hitbox and collision data, on top of the fact that the player clients are hosting the servers directly.
As mentioned before, CS vs any game without a dedicated server is an unfair comparison. CS runs on dedicated servers. That’s why the netcode seems so great. Any game with a dedicated server that’s hosted on a machine that’s created to be a server seems to have great netcode. But try hosting a CS server on a home PC and invite your friends who do not live locally and you will see things arent so incredible. To even compare to a 1v1 fighting game you will probably need about 8 players on your server. CS even makes the console-versions of CoD netcode look like complete crap for the same reason. Your talking about a dedicated server.
If it were possible to host your own dedicated server for fighting games (and NOT host it on your home PC) it would run great as well. The problem with that is fighting games are not very prevalent on PC’s, and hosting a dedicated server for a 1v1 gamewould need at least 16x as many servers as FPS games, and is therefore not cost efficient.
FPS games use the same netcode as everything other game. Saying they have “better netcode” is a foolish uneducated comment. Like most non-fighting games, they can just use all the prediction techniques that are not available in fighting games.
Their movement is only fast in “forward” directions making movement prediction easy. The servers and client assume you kept moving that direction for the duration between packets, and if you are in a different spot the animation is blended to that point.
Automatic weapons shoot many raycasts at a time which naturally makes hit prediction easy.
Accurate weapons such as Snipers have prediction codes that detect if you would have had a head shot within a short period of time of your shot (not at the exact time of the shot) which is what actually determines headshots.
Weapons have built-in recoil which makes inconsistencies with netcode less noticable.
FPS games do not have startup frames, active frames, and recovery frames the same as fighting games, and therefore do not require the same amount of accuracy, which means most communication is sent unreliably - which means it takes less bandwidth.
Unreliable packets use less bandwidth but do not have to “confirm” that they reached the other players. Things like FPS movement, bullet effects, etc, are all sent unreliably, because if a packet or 2 is lost here and there it won’t be noticable. Reliable communication takes more bandwidth, and is reserved for anything that “needs” to reach the other players. Basically it “hit-confirms” that the packets have been sent, which uses more bandwidth, and has to resend the packets if they are not recieved.The only things sent reliable in FPS games are hit confirmations/deaths.
In fighting games, everything has to be sent reliably. Imagine it was sent unreliably and every once in awhile an attack (either a hyper, a projectile, or a move in the middle of a combo) did not go thru? It would be game-breaking. Imagine the packets that say you are jumping never went thru? It would be game-breaking.
Which means everything in fighting games is sent reliably, which naturally increases the usage of bandwidth.
This is on top of the fact that collision, physics, and hitboxes are a lot more complex in fighting games - physics affect characters differently depending on size and jumping styles, everyone has their own unique collision and hitboxes. This is even more data being sent reliably, which again uses much more bandwidth.
All non-RayCasting effects (anythign that’s not a bullet - which the hit is determined by effectively shooting a ray out from your gun and see if it collides with anything) in FPS games, such as rockets and grenades, are NOT required to be frame-accurate the same way as fighting game attacks need to be, and they are “AoE” abilities which make perfect accuracy of the netcode less important.
These are all methods of prediction or bandwidth saving techniques that can all not be used in fighting games. FPS netcode is not better, fighting games are just far, far more strenuous on bandwidth, do not have dedicated servers as PC games do, and do not have as many options for accurate prediction to hide the latency between players.
Which is why the only way to “hide” the lag between players is how GGPO does it - by intentionally adding frame delay and using a P2P infrastructure so that rollbacks can happen on both sides (rather than server-client directly) and so that the frame delay “hides” the fact that a rollback actually happened. But these are not natural things for game engines to use, and the engine has to be designed for it (particuarly in the rendering and input department, as well as the server structure has to be made for P2P rather than server-client from the ground up).
You can’t just say “Hey that sounds awesome, I want to add it in my game engine with existing rendering, input, and network infrastructure, with all of our tools based around how those things work)”. Game engines have a set of tools designed for the developers to make their own game - think like the toolsets to the Elder Scrolls games or StarCrafts editor. All those would be broken. This is assuming none of the business issues I mentioned in the other posts existed.
^^^^
It’s hard to buy all your point because several 2d titles like blazblue / sf2 hdr / mvsc2 online edition( what ever it’s called ) run a lot better over capcom modern fighting game titles , specially mvsc3 . If you want bad 2d online fighting game , just take Kof12 , this game is real mess online .
All these games are on psn or microsoft servers and some of them completely out perform other game in term of online stability / netcode and online components.
Kinda similar with ggpo on pc. I do not think pounder has the best server ever to run his program and it’s run pretty damn well.
I still think capcom butchered the crap out of sf4/mvsc3/tvc online mode because the online experience is not something important to increase the sells .
Edit : It’s again relate to the priority and sadly, netcode and online component are extremely low in capcom priority .
Name a 3D fighting game with exceptional online. And that means not only is it lagless but it has a decent amount of features in them, spectator mode and lobbies included.
The game is unplayable if you constantly have to switch between offline and online
If you play mostly online you get used to the timing and get to play smooth games against good connections.
I aint trying to say that it is anywhere as good as offline play, but there are really good players online to practice with, and is wrong to think that playing a couple thousand matches online wont help you improve your game knowledge a lot…
It does help for people who don’t have an offline community.
You know what, lets just change the term, lets just say that MVC3’s online is adjustable :lovin:
I know there are legitimate issues with it, but I like the game and I want to get some matches in.
Im hype for UMVC3, because SSF4’s netcode was improved a lot over Vanilla SF4, so at this point the experience cannot do anything but improve.
Yeah if you don’t wiff punish, use combo beyond abcd, xfactor through block strings, etc. Its the worst online net code i’v ever played…reminds me of the DC 56k connection.
lol just to pro capcom fan boyish bias comment , did you ever play something a little more advance over the basic stuff ?? Not only you will constitutively drop any moderate/advance combo , a lot of unsafe moves are safe on block because of the extra delay input create by the bad netcode and shit load of cross-mix are unblockable on reaction online . —> that is the reason people said it like playing under-water …
Both netcode of s/sf4 and mvsc3 are the same . It’s built by the same people and both suck . The only difference between s/sf4 and mvsc3 are
some better online component for ping display and the ability to kick people out of the room ( better matchmaking too )
The flow of the gameplay is super slow in sf4 , unlike the fast spacing in mvsc3 .
Does this mean ssf4 is more playable ? hell no , it’s fucking atrocious , just like mvsc3 .
that is one of the reason people want ggpo or something better for modern titles
***Before posting irrelevant comment , can you please take the effort to read the whole thread ? ***this is the kind of message that can create a lot of conflits in this thread
I agree with what you said, but dude, chill out. The guy admits that there are “legitimate issues” with the game’s netcode. He just says that it’s adjustable, that’s all. Nowhere does he praise Capcom for doing a wonderful job with MvC3’s online.
I’m not gonna say SFIV is great online, but marvel 3 is absolutely positively significantly worse, I agree with everything everyone says about mvc3 being terrible online here, but I think it’s a huge exaggeration to lump SFIV into that category, at least I can play SFIV with local friends with next to no lag, I can’t even do that with mvc3, it’s not even playable with friends in the same city.
If SFIV and MVC3 use the same netcode, and I don’t think they do, but if they do, MVC3 is one hell of alot less optimized.
EDIT: to clarify, not saying SFIV is great online, it too would benefit from ggpo, just saying if UMVC3 was even comparable to super SFIV, I might actually play it online.
SF4 and MvC3’s netcode is mostly the same (same engine and were done by the same people). This has been confirmed by Seth.
Primary difference is MvC3 moves faster (so slowdowns can be more noticable) and if your using assists you use slightly higher bandwidth. Which, for some people, could be a big deal, considering the game uses just as much outgoing traffic as incoming, and in general internet connections in the US are optimized for download speeds and screw the upload speed (that’s where the input lag comes from that makes it seem like the opponent can react faster than you can). .
In most cases, if SF4 runs fine for you, MvC3 will too, unless your upload happens to be just enough for SF4 but falls short of MvC3. But if SF4 was unacceptable or barely acceptable, then MvC3 is going to be worse. Needless to say, that’s why it’s no surprise if you have issues with MvC netcode in general, it definitely won’t work with ur local friends - your most likely on the same ISP with the same upload speeds which means both players slowing things down.
Most fighting games need about 120-125+ kilobytes per second in a speed test ~1000 miles away to have acceptable performance. (please don’t mix this up with kilobits - most tests and advertizements tell you bits not bytes… 100kilobytes is approximately 900kilobits if memory serves.). To find out your speed go to dslreports.com and use one of the speed tests on the tools page, and use a location that’s around 1000 miles. Then to find bytes instead of bits, check the detailed text information.
I don’t mean to be diffifcult here but I don’t buy that, do you have a source of seth confirming this? Infact this part of this interview kinda contradicts the idea that both games have the same netcode.
EDIT: the ad messes up the time in the URL, just set it to 15:00 for good measure.
my connection is great… there is no reason beyond the game itself for MVC3 to suck locally.
seriously, I don’t care if the netcode is the same and MVC3’s data transfer is just sooo much higher that it makes it perform badly, or if the netcode is just totally different and inferior, either way the games are not equal.
my only point is that SFIV is passable when compared to other recent online fighters, my experiences with super SFIV are in the same league as mvc2 or even blazblue, MVC3 however… is like the fighters on the original X-box live, or something terrible like kaillera (before p2p)
and comon, if join dates mean something to you guys, look at mine, lag is not masked by the pace of the game, mvc3 doesn’t just feel super delayed because it’s fast, it’s super delayed and SFIV is not, I can tell the difference.
