Yeah you’re one of those guys that say sf4 “felt” fine.
When it comes to lag, when it comes to netcode, when it comes to something more nuanced than being either unusable or passable just shut the fuck up. Get numbers, get facts, do scientific tests or just shut up until someone does.
I’m not singling you out, I’m not trying to be mean but seriously between the monitor discussions and hearing that people prefer the objectively worse netcode of sf4. Having a smooth picture means nothing when everything is delayed so heavily.
…or maybe you can get numbers, facts, and scientific tests before you tell someone to shut up for giving his own opinion. Did he bashed someone or you in his message? No. Does your reply seem like it? Assuming the choice of dialogue, yes.
If anything, you are just as guilty as him for stating opinions from your past message.
So… where’s your proof that it’s “infinitely” better?
I’m even finding it funny that a mod will like your comment, since there was nothing constructive about what you just wrote. He is stating opinions. Forums are meant to communicate that could diverge into all sorts of aspects of any specific topic in mind. This ranges from facts, opinions, and more… and while certain rules do apply, a thread title such as this is perfect since it’s purely about “discussion.”
You clearly favor rollback over delayed netcode, and there is nothing wrong with it. However, saying that it means nothing seems faulty.
Example: Try and attempt a “Low Short, Low Jab, Low Short xx Super” confirm in 3S with Ken and then try “Low Short, Low Jab, Sweep (delay chain attacks to match 3S timing excerpt)” confirm in SFIV with Ryu. Timing might not be similar, but it’s rather close. Have the same player you know that has some teleports on rollback on 3S Fightcade, and has IV.
What I’ve noticed? I had to not believe what I see for the first frames in rollback for the 3S test that would cue me in to confirm my initial pattern into a super, potentially resulting in a string xx super that’s, in reality, actually blocked. In the SFIV test, while there would be times that I feel I couldn’t react in time, I at least KNOW I can’t convert accurately in time.
Okay, well that’s one issue rollback can provide. We can keep going back and forth and mention how, for example, rollback seems more favorable in terms of execution (that’s unrelated to early hit-confirm’s that are unsafe if miscalculated). Point is, just because delay netcode has delay (surprise), it does not show that it “means nothing”.
I believe it to be fact that the best of GGPO like Skullgirls and Alpha 2 online is definitely better than any delay based netcode. You’d have to be crazy to even argue that.
SFV’s netcode still needs enough work where you could argue what is better in what situation, but I am definitely of the heavy opinion that good SFV connections definitely feel better than IV’s and I would do better cross country now than I would have in IV.
If we had the ability to set our own frames of smoothing, this wouldn’t be an issue AND we’d have consistent inputs. But alas, we can’t
A good implementation of rollback netcode is better than variable netcode in every category. But SF5’s version really isn’t the best implementation and it doesn’t give us any customization. Which doesn’t really help at times in terms of playing on a bad connection.
stupid discussion.
the game uses what it uses and you can’t do a thing about it.
do what you can to have a stable connection, much more important than latency (unless you’re playing people over 150ms or something).
beyond that there’s nothing more you can do other than try to form a group of people you know who have good connections to you and you enjoy playing. matchmaking is always going to be a waste of time in a fighting game like SF. lobbies are where any online fun is had.
The only issue is that means your ceiling becomes your play group. As poor as matchmaking tends to be, at least it gets you out of your comfort zone as it comes to play level. Not everyone has the ability to play strong players all the time. Of course, that’s also irrelevant since unless you’re going to majors who cares if you level up at all I guess.
I always value ranked because you never know who you’re going to get matched up with and it’s always good to get practice vs all levels of play. Some people just don’t care to sit in a lobby all day waiting or just playing the same person for an hour. Ranked gives you that ability to get exposed to a lot of stuff while not having any downtime waiting in a lobby or just running the motions with one person for a long time.
These newer virtual lobbies like the ones in Guilty Gear let you kinda go back and forth between players like at a real arcade or game center, but sometimes I still just wanna just have the next person out in the world come to me in ranked.
MOV said he even found value in playing against CPU at the arcade.
i didn’t say it was some set play group. not that this is different than it actually is anyway. if you want to get very good you’re going to need to travel one way or another at some point.
you can play a good chunk of players no matter where you’re from (except like hawaii and alaska, sorry bob) with at least a decent ping (100ish) and many people with an excellent ping (50ish) if you’re near any population center.
what i mean is matchmaking is always going to be a crapshoot. you’re better off being a part of the community and finding people to play with and all that good stuff. being in a lobby with people is fine if you like the game and people are on mic. so if you’re complaining about matchmaking imo it’s misguided. don’t let an algorithm determine who you should play. make the connections yourself.
English isn’t your first language and that’s fine but I don’t give enough of a shit to single someone out like that.
It was infinitely better because it did not delay inputs, it rolled back the game. Simple fact. That is better at a base level for practice.
Smooth picture means nothing if things are so delayed you are playing an essentially different game (i.e. almost every fighter out there) Rollbacks are annoying (the few times they happen on decent connections) but it’s a minor gripe, not a huge deal dragging down the experience making it work against you in offline play.
It’s not just combo timings, in SF4 I picked Cammy and DP’d like crazy, got away with murder. No one can properly react in that game because by the time they see it and react they’ve already missed the boat.
If your game is rolling back so much that’s probably on you. I played for hours during this beta and got maybe 2 bad matches the whole time while streaming to 3 different sites at 3628kbps. That’s roughly 11mb uploading. I had it set to 3-5 by the way.
If SF5 isn’t working for you I highly recommend you fix your shit.
A stable connection of 100ms between 2 players in SF5/SFA2 (GGPO)/SkullGirls (GGPO)/ KOF2k2 (GGPO) feels offline. A stable connection of 100ms between 2 players in SF4 gives you around 4 to 7 extra frames of input lag.
Thereare exceptions but thats what they are, exceptions: happens rarely.
What I noticed in the beta (if you believe country flags), is that it’s not so much the delay, which is a almost entirely about the distance between the players, that causes problems, but rather the quality of the connection. I had problems even with players in other European countries, notably places like Spain, Portugal and the UK (I’m in Sweden). That’s only like 50 ms away, but the game was still “unplayable”. On the other hand, I had a few decent, but not great, quality game against people in the US, although most were terrible.
This leads me to believe that the most important factor here is the variance in the delay, what’s known as the “jitter”, and possibly packet loss. That’s why a ping/latency indicator isn’t particularly important, unless it’s very specific about updating in real time to indicate changes in the delay. In Counter Strike: Global Offensive, you get a jitter indicator whenever you turn on the “net graph” options in the console:
The difference is of course that in CS you have all players connect to a single server. In SFV, I assume that the Capcom servers only do matchmaking and setting up the peer to peer connection between players, but I don’t see why you couldn’t have the same kind of feature in this game if you really wanted to. You could then play “low jitter” players across the Atlantic if you wanted to, or maybe US-Japan games, but without the ability to filter on that variable, most such games will be bad.
Another factor that I think is interesting to discuss is how the netcode most likely will expose your public IP address to everyone you play (assuming that it uses some kind of standard peer-to-peer model). This exposes you to DDoS attacks, something that has been a problem in other PC games whenever they are played on a competitive level online, or when people Stream. In Counter Striker though you need to get the IP address through other means than just playing the person since all packets flow through the game server, and never directly between players. If SFV exposes your IP address to the other player, you probably never want to use your real nickname if you are any kind of notably figure. Only playing on PS4 should also protect you to some degree.
As someone who deals wih DDoS on a daily basis during his day job I can tell you with full confidence that majority of “DDoS” attacks aren’t attacks at all but self-absorbed kiddos who think they are the center of the universe being incompetent.
People throw that term around like it’s a joke, as if every kid with his mom’s credit card has access to a bot-net…
You can’t imagine how many times I’m on the phone with people who swear that they are being DDoS’d only to find out they miss-configured their router…
Besides, just in case you didn’t realize this yet, as soon as you connect to the internet all your privacy goes flying right out the window…
Any twat with wireshark is easily able to find your ip-address within seconds.
If you, for some reason, think that someone would seriously DDoS you then grab a VPN because that’s pretty much all you can do.
Edit: oh, and as a side-note. Counter-strike isn’t exactly the poster boy for good netcode either…
The interpolation is a good way of doing things but it also allows for really weird shit to happen like getting shot behind walls or bullets that seem to disappear into thin air.
Nope, one end user is not getting access to another end user’s IP address in non peer-to-peer applications. The owners of this forum will know everyone’s IP address, but I don’t know yours and you don’t know mine. “DDoS for hire” is real and very cheap, especially if your target is someone with very limited bandwidth, like a home user. And no, getting a VPN is not “all you can do”. Not exposing your IP address to people that could attack you in the first place. Very relevant for top players online, and maybe not something that think about if they’ve only played on console before.
Has it been confirmed if SFV will have a search result list for you to choose your matches from (like SF4)? Or will it be only a randomly found game each time you search (like MK)?
see this is why I shouldn’t be commenting on these types of statements.
Someone takes a OSCP class they know everything there is to know about network security…
I’m actually interested in this stuff so I’ll definitely try to “proof of concept” this when the game comes out to get determine how close to reality my hypothesis is. Also, I’m not just some guy that “takes an OSCP class”, but it seems like you’ve already decided what I am so we should probably not discuss this further.
Ah… how nice of you. It’s not like your post contains grammar errors or anything really, such as having no comma after “…first language and that’s fine[,]”. That is a really great way to start an argument against someone.
You might as well tell the vast majority of SRK that English is not their first language while you are at it. This actually goes along with almost every forum, considering most people can tolerate even the minor errors of grammar in forums, acknowledging that recipient’s time and patience he or she has for a simple forum post.
Again, while it is better in its own regard (depending on location, I would imagine that the areas close to California or New York will happen to get the most success and reward for rollback), saying that it is “infinitely” better is misleading. It is a biased way to express it, but apparently when someone gives their own experience from both netcodes and shows any sort of favoritism towards delay netcode, it makes some sort of sense for you to tell those to keep their posts to themselves and do not post any sort of opinions without proof.
Yet, the choice of using “infinitely” from your own past opinion is wrong. If you want to follow along with your first assumption based on your first line you replied back with, the use of infinite would mean that it is a variable that has no end. I just pointed out one of the few negative factors that rollback has over delayed netcode, and while the long-term result could end up with rollback being the optimal choice for online gaming, it still does not fit the criteria of being “infinitely” better.
DP’d like crazy… so, you would do it often in your example? Well there are two things that can occur, either the delay is so horrendous that a single button confirmation would show up in the screen at approximately 1 second, or the opponents you play focus too much on reaction rather than doing something in between prediction and reaction and have the thought process of punishing DPs ready.
Either way, this is somewhat of a poor example, because the same outcome would occur on a bad connection regardless of rollback or delayed netcode. In delayed netcode, there is a strong possibility that your inputs just won’t come out properly in time, resulting in a punish that you could not rely solely on pure reaction. In rollback netcode, the same outcome would occur in a sense, but instead, the animation of DP being blocked would feel skipped, resulting in a DP that would have skipped quite a couple frames already (a teleporting effect).
So, even with some (or one) of those videos showing that one specific player receiving the bad side of rollback problems, while the other player does not experience any faults, the first thing one should do is to “blame” the guy that is lagging? Well, obviously it cannot be you. Something lags? Best thing to do is blame the other party. Let us not worry about how far the location from one another is, as well as any other technical difficulties that could arise.
From my experience of seeing one of my own matches archived from a live stream, the gameplay on Valle’s side looked nearly perfect, yet I recall that my match with Valle had a few skips throughout our match. One thing to note was that he was “streaming”… and while it shouldn’t matter at all if they have a sufficient amount of download/upload speed provided from their ISP, it’s worth noting that most of the comments about this problem seems to involve around a player that streams. For the question of curiosity, my download speed is 50mbps, while my upload speed is 5mbps. No streaming of any sort occurred during my time with the beta… and I had a couple matches that meet my satisfaction.
We may have to take these posts as a grain of salt, but if their network statistics show that they are well capable to have good matches (and the person they are playing is close by), there has to be a good explanation as to why this problem occurs. It could be fixed (hearing that Skullgirls had this issue and fixed it through some time), but the last thing that comes to mind is to blame players in this case.
My matches have been pretty good otherwise, but I did get some matches that definitely did not felt like 5-bars… and sure, there was a tweet stating that we should put the search criteria at “3-5 bars” for the best matchmaking performance, but it’s too late for someone like me who only had time to play it on the first day for a couple hours.
I’m going to assume that your last line is not intended towards me, since I had a couple satisfying matches outside of three of four games. I merely mentioned how a lot of the visual aspects from “watching” a live, online match (SFV netcode) does not accurately show what a normal match would look in a visual sense (NOT in a player vs player psychological sense).
