Terminology that drives you crazy



  • Continuing the discussion from Blakeyrat Is Building A New Gaming PC 2: Electric Boogaloo:

    @tarunik said:

    My take (overall, not on the latency thing specifically) is that:

    For the love of $DIETY, people, the problem isn't latency, it's jitter


    Ok, got my rant out. But man, it always bothers me (as someone who does network admin at an ISP) to hear people complain about their latency when they're lagging in an online game, when what they really mean is the jitter (inconsistent variations in latency). If I remember, I'll link in the Riot Games post on what they're doing for their NA servers, because I think it did a decent (if overly simplified) job of explaining The Lags for the masses.

    But what other terminology abuses drive y'all crazy?



  • That's not a terminology problem. That's a "two different words for two different concepts" problem, and you assuming tarunik was using the wrong word when in all likelihood he typed exactly what he meant to type.

    How about, from the Enlightenment thread, using "signal" and "slot" instead of "event" and "handler"?



  • Synergy

    EDIT: And The Cloud


  • Discourse touched me in a no-no place

    Surely latency is also a problem?


  • Discourse touched me in a no-no place

    @IngenieurLogiciel said:

    Synergy

    Synergise.
    Solutionise.



  • Extreme latency in comparison to other players (or the length of the reaction window for the PvE mob's attack), sure. The sub-300ms latency that exists on nearly all non-saturated high speed connections in first world countries? I sincerely doubt it.

    As an anecdote in point, one complaint that's really easy to find on a League of Legends subreddit or fan site is the latency difference between NA west coast and NA east coast players. Apparently west coasters typically see latency in the 50-80ms range. I, a Midwesterner closer to the east than the west (Ohio), have 90-120ms latency. And apparently New York/New England players typically see 100-130 (before adjusting for ISP poor quality or in-home congestion). There's just no real difference there -- you're talking 0.05 seconds difference, less than you can even score a game winning basket in from an out-of-bounds play.



  • At least on Dota 2, the difference between 70ms and 150ms is night and day.



  • @dstopia said:

    At least on Dota 2, the difference between 70ms and 150ms is night and day.

    Right; with the faster connection you get to see:

    faggot newb fag fuck your mom

    In the chat window a whole 80ms faster!



  • Yup, I can also take the piss out of everyone who gets mad at words on a screen much faster! It's amazingly efficient.


  • Discourse touched me in a no-no place

    • Revert back to me. (Mentioned topics (and boards) passim ad-nauseum.)

    • "Can I hash-tag that line out" (ditto)

    • Mixing up IMEI and IMSI

    • Mixing up <abbr title="bits"">bps vs Bps

    • Class A, B, C network (this one was from a sysadmin of all people - and he's younger than me - in fact I'd be surprised if he'd even remember, let alone worked with, named classful networks.)

    • On-boarding (initiation/start of probation period)



  • @PJH said:

    Mixing up bps vs Bps

    "I get 50 megabytes [per second of bandwidth] on my internet connection."
    "You mean megabits, right? 50 megabytes is like 400 megabits and you probably don't get that."
    "Whatever."
    "NOT WHATEVER!! IS IT 2500 INCHES FROM SAN FRANCISCO TO NEW YORK!?! NO!! IT'S 2500 MILES!! BIG DIFFERENCE!!!"



  • Though, in the US, 50 MB/mo is was probably a feasible data cap before the evil data-grubbing ISPs started charging overages before the saintly FCC stepped in and made that illegal even more likely to happen

    🔥


  • Java Dev

    A colleague of mine has a gigabit symmetrical. Another has 500mbit. I make do with 50mbit.


  • BINNED

    @PJH said:

    Mixing up IMEI and IMSI

    I get people confusing GPS and GPRS a lot :wtf:

    @PJH said:

    Class A, B, C network

    *shakes fist*



  • Having not played or researched DOTA at all, beyond having a vague understanding that League is a "dumbed down" version of it, I presume that the primary difference in 70ms vs 150ms latency is ease of last-hitting the NPC creeps?

    I wonder two things (and in true "weasely economist" fashion, I'm just gonna chuck these out here as "questions wherein it would be nice to see further research", because fuk if I know how to test for them)...

    1. Is all else equal between the 70ms and the 150ms connection, in terms of connection consistency (jitter and packet loss)? To me, it would seem reasonable to presume that the 150ms connection is travelling through more networks (more interconnects, more hands) along the way than the 70ms connection. After all, the speed of light would suggest that "minimum" latency to get all the way around the world would be somewhere in the neighborhood of 140ms (light can make 7 laps around the world per second, in a vaccum), so unless the 150ms connection is literally halfway around the world (12,500 miles/20,000 km) further than the 70ms connection, there must be some cause other than distance that is increasing the latency.

    2. More importantly, is the difference independent of skill? I would expect that "casual" players would see an improvement in their ability to last hit, because they're more dependent on reactive last hitting, rather than anticipating which minion to attack and when. By contrast, I would expect pro and top-tier players to play at basically the same level regardless of latency for any connection under 300ms. (Very cursory google-fu seems to suggest that various online tests of human reaction time get median reaction times around 250ms, and this site seems to estimate machine latency as adding 30ms to the results of their test, which would imply a human reaction time closer to 200ms than 250ms).

    Again, if it wasn't obvious earlier in my post, I'm making speculative statements and ponderings, not attempting to claim some sort of fact.



  • "Leverage" (as a verb...)



  • There's not much to discuss, you just need to sit down and test Dota at 70ms and 150ms and, unless you're not very good at coordination, you'll instantly notice the delay between your actions and what the hero is doing. It affects everything, not just last hitting.

    For reference, I'm a 3.2k player, which is about average for Dota 2 for people who are decent at videogames (it is technically above 75% of the playerbase, but most of the playerbase is terrible at videogames in general, as is probably the case with LoL too). For top players, for obvious reasons, the effect is a lot more noticeable and there have been issues between teams about server locations for online tournaments.

    I know nothing of how LoL netcode works. Maybe they sort of compensate for that server-side, but I have no clue how that works to give everyone the same experience. But I'd say that saying there's no difference between 70ms and 150ms is quite preposterous.


  • BINNED

    @dstopia said:

    It affects everything, not just last hitting.

    You mean there is other stuff to do besides hitting things? 😮


  • BINNED

    @Luhmann said:

    You mean there is other stuff to do besides hitting things? 😮

    Typing abuse into chat window?


  • BINNED

    @Onyx said:

    Typing abuse into chat window?

    That's just hitting random keys again, isn't it?



  • @izzion said:

    I presume that the primary difference in 70ms vs 150ms latency is ease of last-hitting the NPC creeps?

    Isn't that robbing Peter to pay Paul? I mean, you're both on the same team, why would you be competing for creep kills?

    @dstopia said:

    I know nothing of how LoL netcode works. Maybe they sort of compensate for that server-side, but I have no clue how that works to give everyone the same experience. But I'd say that saying there's no difference between 70ms and 150ms is quite preposterous.

    Without knowing how the netcode works, I don't think you can make any judgement at all about it. As far as you know, LoL only does 250ms time slices or something crazy.



  • Hrm.

    So, rather than trading anecdotes all day (but flame wars are fun!), I decided to be unlazy and try to Google search for data to support my point. It seems that the data don't support my point for FPS type games (where studies suggest that there is a sharp break point in effectiveness at 100ms latency, as of Nov 2006 at least). But that study and a couple others suggest that some other game types such as Madden are less sensitive to symmetric latency -- they basically are defining latency as one way travel time rather than two way, apparently because Madden is P2P within the game -- especially with latency totals under 500ms.

    This study seems to suggest that players who play more often ("expert" players) are much more likely to report poor satisfaction with the game at higher jitter levels, even though "these 'expert' gamers who are presumably the most sensitive to latency/jitter are also the most adept at compensating for impaired conditions as observed from their final game scores (which are not presented in this paper)."1

    So, I guess I've found some data in support of my point 2 (more experienced players will be less effected by changes in latency and jitter), some data in support of your point (150ms jitter is noticeably worse than 70ms jitter, at least/especially for FPS games), and little data thus far at all in regard to whether latency or jitter matters more.

    1See page 9 - the Conclusions section, second paragraph.



  • In League of Legends, at least, you're competing against the 0ms latency NPCs on your own team (your creeps, who are busy attacking their creeps).

    Although your creeps attack in an anticipatable pattern, with fairly small damage packets, hence my theory that professional-grade players should see less relative decrease in their ability to last hit than casual-grade players under 150ms latency conditions.



  • @blakeyrat said:

    Without knowing how the netcode works, I don't think you can make any judgement at all about it. As far as you know, LoL only does 250ms time slices or something crazy.

    I never said anything about how LoL works. I was disputing the claim that latency under 300ms doesn't matter unless you're playing an FPS or something, or saying that

    there's no real difference there -- you're talking 0.05 seconds difference, less than you can even score a game winning basket in from an out-of-bounds play.

    is all complete bullshit. If you sit down and play a game of Dota 2, there's a world of difference between 150ms and 70ms.

    Maybe LoL compensates for lag, I have no clue. In Dota 2 you can't compensate because there can be no inaccuracies in the flow of the game, so your character input (right clicking to move, using your skills) gets the short stick so that everything on the server (character positions and spell usage) gets synchronized correctly. As an addendum, it's also far easier to do it in a game like Dota given that pathfinding takes care of most of the issues for player input, as opposed to an FPS where the player has full 3D input on their movement, which I guess is the reason why latency generally handled more lazily in FPSs.

    @izzion said:

    So, rather than trading anecdotes all day (but flame wars are fun!), I decided to be unlazy and try to Google search for data to support my point. It seems that the data don't support my point for FPS type games (where studies suggest that there is a sharp break point in effectiveness at 100ms latency, as of Nov 2006 at least). But that study and a couple others suggest that some other game types such as Madden are less sensitive to symmetric latency -- they basically are defining latency as one way travel time rather than two way, apparently because Madden is P2P within the game -- especially with latency totals under 500ms.

    Dude, look for all the studies you want, or just download Dota and try different servers. There's not much to argue in this respect, this is something that 100% of the playerbase that knows something about videogames understands of Dota. And that's not even getting into FPSs, which I had the displeasure to play at >150ms latencies on a regular basis due to where I live.



  • Irregardless.



  • @izzion said:

    For the love of $DIETY, people, the problem isn't latency, it's jitter

    If I have 700ms ping to a game server for a game that requires near-instant reaction time, my latency can be constant and I'll still have a shit time.



  • Emoji.

    They're called emoticons. Any contradictory comments are wrong and also a rape culture.



  • Emoticons are images like :rolleyes: :rofl: :spittake:

    Emoji are Unicode characters.



  • Oh nice, I can build my own custom animated emoji too!



  • Nope. Both are emoticons.
    Rape is wrong.

    Protip: this is the awful terminology that drives you righteous thread. Not the bad ideas debate thread.


  • BINNED

    @CreatedToDislikeThis said:

    Rape is wrong.

    Even if I want to rape people who talk about cloud synergy?


  • BINNED

    @Onyx said:

    cloud synergy

    I propose we move past the cloud and into the FOG.



  • http://www.gamefaqs.com/boards/975212-super-street-fighter-iv/56322081, post 6:

    This game is already easy mode with the enormous reversal window... and you want MORE?

    Two frame links are possible possible to hit very consistently with enough time to practice. The majority of one frame links can be plinked, turning them into two frame links. Jeez with a 5 frame window on all the combos, you could just smack your fist against the board and would probably be on par with the pros execution-wise.

    Your buttumptionspeculative statement that 80ms wouldn't matter to a pro gamer is way off.



  • My understanding of Street Fighter combos is that you're comboing with yourself. Meaning that, for completing combos once you got the first move off, consistent 200ms latency (with 0 jitter) wouldn't make completing combos any more difficult than consistent 10ms latency. So I think your example is poorly chosen, and that you picked an example case where jitter matters more than latency.

    However, I will concede that getting the initial move off would feel/be less consistent in a higher latency situation, since you'd be more likely to get interrupted by an opponent's move you hadn't seen when you tried to start the combo.



  • Dude, fighting games aren't only about comboing, they're about knowing yours and your oponent's neutral moveset and exploit the frame differences between them. If you have a 200ms delay on your input, you're at a major disadvantage.

    You seem to have no clue of how videogames networking works. This is something you learn by actually playing the games. I suggest you do the same instead of quoting random studies.


  • ♿ (Parody)

    $DIETY

    I won't go into a long explanation here. It's been done elsewhere, though discosearch might fail you (but not for the last time, I'm sure).


  • ♿ (Parody)

    @dstopia said:

    This is something you learn by actually playing the games. I suggest you do the same instead of quoting random studies.

    Quoting the studies sounds like a more productive use of time, though.



  • I'm not about to argue with that, but if you're going to talk about latency and input delay in videogames, you should probably test that before even beginning to think about the topic.


  • BINNED

    Jeffing in 3... 2...


  • ♿ (Parody)

    @dstopia said:

    you should probably test that before even beginning to think about the topic.

    Because who needs causality‽



  • Yeah, all I was getting at there was about the order of magnitude of what constitutes noticeable changes in timing. I mean, I am by no means a pro gamer, but from what I hear about people who are, even a single frame change in a character's animation or response time can make all the difference between viable and worthless trash.

    Anyway, what you're saying about jitter does make sense; I'm not trying to refute your central thesis here. It's just that some of the conclusions you seem to be drawing are inconsistent with what I've gathered from seeing people talk about that kind of stuff. Like, you seemed to be saying that people don't know the difference between jitter and latency, but maybe what you should have said instead is that you hate it when people refer to ‘jitter’ as ‘lag spikes’, or whatever they're calling it these days. Because you can bet your ass the pros know what's going on in their game, timing-wise.



  • Do you play many competitive videogames over the internet? Because it definitely sounds like you don't.

    Look, I don't care about whatever random study thinks about networking in fucking Madden. If you sit down and play Street Fighter over the internet on 200ms ping, you're going to have a shit time. This is undeniable and anyone who plays these games even remotely seriously knows it.


  • ♿ (Parody)

    @dstopia said:

    This is undeniable and anyone who plays these games even remotely seriously knows it.

    I just thought it was funny that you said someone should test something before thinking about it.

    @dstopia said:

    Do you play many competitive videogames over the internet?

    I just said (somewhere), Solitaire sucks over a network.



  • @boomzilla said:

    I just thought it was funny that you said someone should test something before thinking about it.

    Well, before formulating any theories, the least you could do is record your own empiric experiences with the subject matter. That will prevent you from making asinine statements such as "latency doesn't matter in videogames, it's just lag spikes that do".


  • FoxDev

    @dstopia said:

    You seem to have no clue of how videogames networking works. This is something you learn by actually playing the games.

    So the person who wrote the first ever networked videogame; what game did (s)he play to learn how videogame networking works?



  • You know what I meant.


  • FoxDev

    @dstopia said:

    You know what I meant.

    Yes, but that's not why I asked; I asked in order to point out a major flaw in your argument. Playing the games will teach you the practical aspects (maybe), but you won't develop a proper technical understanding unless you do some background reading too. And part of that is the studies you so readily dismiss.



  • It will teach you the practical aspect and the practical aspects are what is affected by latency.

    It's not very hard to understand.


  • FoxDev

    Oh, of course. So we should just ignore the technical reasons that people can then use to change stuff that makes the practical impact less.



  • @izzion said:

    My understanding of Street Fighter combos is that you're comboing with yourself. Meaning that, for completing combos once you got the first move off, consistent 200ms latency (with 0 jitter) wouldn't make completing combos any more difficult than consistent 10ms latency. So I think your example is poorly chosen, and that you picked an example case where jitter matters more than latency.

    However, I will concede that getting the initial move off would feel/be less consistent in a higher latency situation, since you'd be more likely to get interrupted by an opponent's move you hadn't seen when you tried to start the combo.

    @dstopia said:

    Dude, fighting games aren't only about comboing, they're about knowing yours and your oponent's neutral moveset and exploit the frame differences between them. If you have a 200ms delay on your input, you're at a major disadvantage.

    You seem to have no clue of how videogames networking works. This is something you learn by actually playing the games. I suggest you do the same instead of quoting random studies.

    The bottom line is that constant latency is easier for a player to compensate for (by adjusting timings, leading, etal) than jitter is. It doesn't mean that latency doesn't matter at all -- just that it matters less than jitter does.

    I'd also strongly argue that making a game where a constant 100ms latency differential gives you a nigh-insurmountable advantage is extremely bad design -- there is a geographical lower bound on latency, and even the best transoceanic fibers near 100ms by themselves.

    Filed under: in b4 a horde of angry Aussies and Kiwis show up on your doorstep because they can't play your game worth a damn and there's nothing they can do about it


Log in to reply