The discussion has moved over to Discord

Home Forums Overload Multiplayer What Overload multiplayer needs to succeed

RSS Feed
Viewing 15 posts - 1 through 15 (of 72 total)
  • Author
    Posts
  • #15778
    SiriusSirius
    Kickstarter Backer
    Topics: 9
    Replies: 406

    I know preaching isn’t always popular, but these concerns are going to come up, so I figured I might as well get to them now rather than later. πŸ™‚ I don’t want to talk about mere preferences, as a lot of that comes down to “Overload is not D1” and could just be written off as bias from having played Descent multiplayer for 20 years; this is more about stuff I can’t see Overload actually being a fun multiplayer game without. Particularly because I’ve seen what happened without them and it wasn’t always pretty.
    Also, I’m focusing on player-vs-player here, not co-op; co-op in general is really pretty forgiving and hard to screw up really bad as long as you don’t desync. Except for item #2, that’s probably still important for co-op.

    1. If it looks like I dodged it, I should have dodged it.
    Specifically, latency shouldn’t be an issue when it comes to dodging – what happens on the player’s screen should determine their success, not what happens in the server’s world. This is in contrast to most first-person shooters, where the emphasis is more often the other way around – they take pains to ensure the shots you took that looked like they counted actually did count – people hate missing.

    Why this is important: Unlike most first-person shooters, Overload (like its spiritual predecessors) is dominated by travel-time weapons. You’re not supposed to have a meaningful chance to dodge instant-hit weapons outside of forcing your opponent to shoot in the wrong direction, but travel-time weapons actually can be dodged. And players will expect to be able to dodge them. So if you thought you dodged a Thunderbolt shot, then took damage anyway, it’s going to be annoying.
    This gets particularly critical for homing weapons. If you dodge these too early, they’ll turn to track you, and you didn’t really dodge. If you’re too late, they’ll have hit you by then, and again, you didn’t really dodge. The time window to execute a dodge can be pretty narrow, and if you have to guess at how much latency to add to it, it gets very frustrating to get it right. It’s much easier if the model of the game a player has to work with is just what he/she sees on his/her screen.

    Possible approaches: Obviously the simplest way to achieve this is peer-to-peer netcode like D1/D2 had, but that comes with some drawbacks; modern engines may not support it well, and it is easier to cheat. You can still achieve the same effect in a client-server model as long as either the client reports whether it got hit (subject to sanity checks, probably) or the server tries to work back in time to figure out what the client’s state would have been like at the time of the apparent hit (complicated, probably prone to failure).

    Games that got this wrong: Descent 3 (except for P2P mode), and almost every Descent-like game after it

    2. My movement should be dictated client-side.
    This basically means “don’t do permissive client/server”. It can be OK for some games at low latency, but it seems deeply dissatisfying for 6DOFs. When I turn or move my ship, I really want to see that on the screen immediately, as though it were a single-player game.

    Why this is important: Because permissive C/S leads to rubber-banding, and it’s basically unavoidable in a game as complex as a 6DOF even if you do some level of client-side prediction. If you change direction, there will be some degree of mismatch because it takes time for the packet to reach the server to say “OK this player is changing course now”, and that will create an error, and then the client has to correct that error, which will cause the ship to jump. It feels awkward, can exacerbate motion sickness, and makes dodging things harder (see issue #1).

    Possible approaches: Since Overload isn’t using Unreal Engine, it might not be pressured into using such a model. But I don’t know much about what features Unity gives for netcode. Either way it comes down to either making the client dictate their own position, or just not syncing the server-side ship position back to the client.

    Games that got this wrong: Descent: Underground. Sorry guys. πŸ™

    3. Keep the superweapons under control.
    This doesn’t mean there shouldn’t be superweapons – even D1 had the Mega Missile. Rather, they shouldn’t take over most games – if they’re secondaries you can keep the supply low, and if they’re primaries they should typically not appear in multiplayer levels.
    For a weapon not to be a superweapon, it needs to have a reasonable number of counters – strategies you can employ to counteract it, or other weapons that can counteract it when used properly – rock-paper-scissors style.

    Why this is important: Because you usually want there to be something to learn in a game. If all I need to know is “get this weapon and use it all the time”, the game isn’t very deep and I might as well play something else. If the Gauss Cannon wrecks people 90% of the time, the game basically becomes about getting that gun, preventing your opponent from getting that gun, trying to drain their ammo when they do get that gun, and making the most of your opportunities when you’re the one with the advantage. There are real strategies to that, but they’re a lot more boring than a more well-rounded arsenal of weapons.

    Possible approaches: Overpowered weapons and powerups can exist in the game, and probably should – they’re fun in single-player and co-op! But they need to not be available in standard PvP levels. Powerful missiles (Nova, Devastator, etc) should be relatively scarce to prevent missile-boating from becoming the dominant strategy. Top-end missiles (e.g. whatever the Earthshaker equivalent is) should probably only be available in levels designed for them (there were purpose-built Earthshaker levels in D2, and they were actually fun from time to time).

    Games that got this wrong: Descent 2 (Gauss), Descent 3 (Mass Driver, way too many strong missiles)

    4. Turning speeds should be fair.
    I’m not making a value judgement on what exactly the right turning speed limit is, just that it should be the same between mouse users, joystick users, and gamepad users. And that limit should be something that won’t cramp your hands trying to control it on a gamepad or joystick.

    Why this is important: If there is cross-play between PCs and consoles, the console players are going to get wrecked by the mouse-and-keyboard players who can aim better, faster – same reason FPS games typically don’t allow this. Even on PCs, if you can turn near-instantly on a mouse, that means joystick users might as well go home because they either won’t be able to match that, or they won’t be able to aim if they do make their turning super-sensitive.

    Possible approaches: There is always the option to segregate gamers by platform, like everyone else does – which might be enforced when the PS4 and Xbox come into the picture anyway. Or turning speeds could be a per-game setting, although this could prove annoying for mousers if they were expecting something other than what they’ve got. Another option is, if you try to turn faster with your mouse than the ship can actually turn, your view does go there immediately, but it takes a while for your ship/reticle/guns to catch up. I’m not sure whether that’s worth the trouble it would take to implement, but it is something I’ve seen quite a few newer games with vehicles do.

    Games that got this wrong: Thankfully none yet. πŸ™‚

    Criticism is invited – there may well be something I’m forgetting, and although I’m fairly sure about what I posted, it’s still quite likely others will disagree with some of it.

    #15782
    LemurFromTheIdLemurFromTheId
    Kickstarter Backer
    Topics: 3
    Replies: 24

    As someone who’s generally not big on multiplayer (though I have a pretty good understanding of networking), I have to say I agree with pretty much all you said.

    I do have conflicted feelings about #4 though. I’m a KB+M player, and while I’m not looking for any kind of advantage over other players, I do love the feeling of freedom that being able to turn fast brings. But yeah, a more level playing field would probably be important for the success of the game.

    I do think it’s important to point out that limiting turn rate is something that fundamentally affects gameplay (especially) in PvP: if you can’t turn quickly, it’s a bad thing to let other players behind you. This increases skill cap, as it becomes even more important to understand map flow and to predict enemy movement patterns. I think that’s a positive though; I’m generally in favour of a high skill cap, despite the fact that it most certainly doesn’t benefit me (I’m gonna get crushed in MP).

    Anyway, a wonderful write-up, and something the devs should consider carefully. I do have faith in them though, they do have some experience in these things. πŸ˜‰

    Thunderbolt β€” an elegant weapon, for a more civilized age.

    #15788
    yencil
    Kickstarter Backer
    Topics: 7
    Replies: 32

    I think these requests are quite reasonable (I was really worried it was going to be a list of play type preferences, which we all know vary person to person!).

    I would only add that I have enjoyed a few modern games with a high degree of freedom of motion since Descent (such as Titanfall) but I’m not tech savvy enough to know how they handled these things. In Titanfall, I never felt like I wasn’t in control (whether of pilot or titan) and also didn’t experience the lag / needing to lead so far like in Descent. That leading-to-account-for-lag mechanic of Descent feels like something out of the stone age now ; )

    So personally speaking, having a seamless experience like that, where I don’t need to lead my shots by a few ship lengths to account for lag, made me enjoy it very much! I don’t know if that conflicts with anything you said here, though ; ) (also just noting, there is a chance to dodge ordnance in that game, even though most primary shots do travel very fast).

    I’m assuming they’ll get the tech part right and just hoping for a level editor, or even just a few folks being given access to their editor to make levels for the rest of us ^ _ ^

    #15789
    CDN_MerlinCDN_Merlin
    Kickstarter Backer
    Topics: 15
    Replies: 243

    The leading part of Descent/Overload is normal in terms of game mechanics. Unless you are firing a instant hit weapon like the Mass Driver or Gauss. Any energy weapon you will need to lead the shot. Same goes for any space type sim like Freespace.

    Yes the lag/loss makes harder but you end up adapting.

    #15790
    yencil
    Kickstarter Backer
    Topics: 7
    Replies: 32

    I totally get that, but wanted to share what the biggest drawback to multiplayer was, for me. I’m not sure if it’s addressable in a game of this type, but I’d love it if it were! Like I said, all that leading feels like going back in time ; )

    The only recent game I’ve played a lot with really slow shots is Splatoon 2 on the Switch. But in that game, despite players mostly using WiFi due to the Switch, leading isn’t really a part of the game. Anything that could move multiplayer play in that direction would be a big plus, for me! And also maybe connect with younger gamers better.

    Just my take, and not sure if it’s possible, but wanted to share = )

    #15791
    YoshimitsuYoshimitsu
    Kickstarter Backer
    Topics: 57
    Replies: 421

    I don’t have as much Multiplayer experience but these all seem really important. 2 and 4 especially caught my attention so here are my thoughts on those.

    2. Descent Underground’s horrendous and persistent rubber-banding made the game next to unplayable for me. It was one of the top things that drove me away. Client-side prediction was supposed to help but actually made it worse when it was introduced. That seems like a losing prospect anyway when unpredictable movement is one of the most important parts of gameplay.

    4. As a joystick player I must admit that this one has me worried. I’ve watched videos of mouse pilots flying and I keep thinking “how am I supposed to play against that?” I didn’t even take consoles into account and I’ll bet the discrepancy is even worse with a gamepad. Of course the other part of me is intensely curious to see if I could devise flying styles and tactics to compensate. πŸ˜‰ I am concerned about driving mouse players away though. I can definitely see them getting really frustrated if they can’t fly the way they’re used to in Singleplayer. I know I would. I don’t know if there is a real solution to this one other than to remove high turn rates from the entire game (which would probably anger a lot of people), but I imagine that the best partial solution would be a server option to cap the turn rate for that match. Of course the other answer is to just play co-op πŸ˜‰

    #15792
    SiriusSirius
    Kickstarter Backer
    Topics: 9
    Replies: 406

    One of the reasons you don’t need to lead so much might just be that the internet is faster now. Even in Descent multiplayer these days there’s not usually a major difference between what you see hitting other players and what actually does hit them unless you’re on different continents.
    When you’re shooting slow weapons you have less agency over whether they’re going to hit – you can do your best but there’s usually still some chance your target will find their way out. Because of that and the fact that you can’t see exactly how much armor they have, it’s not as frustrating to miss as it would be with, say, a sniper rifle.
    Lest we forget though, there is a pinpoint weapon in Overload… the Driller. Should that be handled differently? It seems like a reasonable question. While I haven’t been bothered by lag trying to use the Vulcan in recent years (it’s usually pretty minor), I know I’m not everybody…

    I have been wondering what impact the schedule change has on the level editor, actually. That’s an interesting question. πŸ™‚

    #15794
    CDN_MerlinCDN_Merlin
    Kickstarter Backer
    Topics: 15
    Replies: 243

    Yes, the Driller will be like the MD of D3. But I think every game has this type of weapon. The new UT (still in progress) has one. I’m sure the new Quake has one. pretty much all FPS have one.

    All energy weapons take time for the beam to travel a distance and this is why we need to lead any shots with them.

    #15796
    yencil
    Kickstarter Backer
    Topics: 7
    Replies: 32

    Great points, Sirius! And maybe we can petition the devs for editor access for the few of you with good experience making multiplayer levels = ) (and time to make a few for launch). I’m so excited for this game in multiplayer!

    #15797
    maniak1349maniak1349
    Participant
    Topics: 10
    Replies: 18

    All of those are valid concerns, but there is coupe of things I wanted to point out.

    1. I’m sure that the modern games design where all the mechanics is dealt with server-side is not to make players pleased but to get the opportunities for cheaters to minimum. If you allow anything to be decided client-side, especially in questionable situations like hit/miss, it will open the way for wider variety of cheats and some of those will even be completely untrackable. Evasion is a perfect example actually – just report more misses in questionable conditions within the margin of error and you will get a completely undetectable advantage.

    2. Pretty much the same as first – if you allow client to decide whether there was or was not a movement it will lead to the same set of problems.

    I know that Overload is niche game and I do want a smooth MP experience as anyone does, especially since coop is a main selling point for me. Hell, I will probably the first to suffer form server-based design, being half the world away from the majority of players. But client-based design will probably lead to most of the public PvP games being ruined.

    4. I think that forced movement restrictions is not the best idea – it will be very annoying for KB/M players, especially if movement will remain unrestricted in SP or they played the game as teaser/during EA. You said it yourself in p.2 – people want to control the game, not to be controlled by it. Sure there probably should be a server option to restrict movement, but I think that configurable balance with carefully thought out presets will be a better general solution for this. Limiting or restriction of the weapons that benefit from fast precise turn rate should at least allow some level of fair play. I don’t know how effective and doable it will be on practice though.

    #15799
    SiriusSirius
    Kickstarter Backer
    Topics: 9
    Replies: 406

    Cheating is always something to try to suppress for sure, and implementation complexity is another consideration, but there is a cost/benefit; there’s no point making a game that can’t be cheated if it’s not fun and nobody wants to play it.

    #15809
    YinutYinut
    Participant
    Topics: 10
    Replies: 55

    1. If it looks like I dodged it, I should have dodged it.

    Sorry man, but as long as it’s online you’re inevitably going to experience some rubber-banding if you want a fair game. Good example would be you making a mad dash for a corner and just as you get behind the corner a driller shot lands on you even though your opponent no longer has line of sight on your screen. It sucks, but it is what it is.

    2. My movement should be dictated client-side.

    This works more or less fine for coop. For PVP, well Maniak already went in-depth why it isn’t such a good idea.

    3. Keep the superweapons under control.

    Superweapons could function almost like a map objective in regular deathmatch mode where they are for example in the middle of the map, they always have a fixed spawn point there and they also have a fixed respawn timer. For example: 1 mega missile spawns in the middle of the map and when it’s picked up another one spawns in 1 minute. This would give players a reason to contest a certain area on the map, because claiming the prize would often guarantee them a kill.

    4. Turning speeds should be fair.

    This is something that has been on my mind as well. Now I would like to point out I have really grown to enjoy the Overload flight/turn model more than Descent’s. As Maniak already pointed out, changing the ship controls between modes would be a so-so move. In it’s current state the mouse users are most definitely at an advantage. I will say this however, hitting people in a 6dof game is vastly more difficult than in a traditional fps game especially if people are actually using all the axes properly to dodge incoming fire, mouse or not. This also leads us to Challenge mode, since there are actual leaderboards and certain players have an advantage, what should we make of that?

    #15814
    yencil
    Kickstarter Backer
    Topics: 7
    Replies: 32

    3. Keep the superweapons under control.

    Superweapons could function almost like a map objective in regular deathmatch mode where they are for example in the middle of the map, they always have a fixed spawn point there and they also have a fixed respawn timer. For example: 1 mega missile spawns in the middle of the map and when it’s picked up another one spawns in 1 minute. This would give players a reason to contest a certain area on the map, because claiming the prize would often guarantee them a kill.

    Man, could I like this 1000 times!?

    I think this is such a great idea! The problem with superweapons in Descent was always that they would randomly spawn all over the map, leading to a mega-fest, and actually incentivizing running away. But making a superweapon spawn in a strategic area on a timer would incentivize fighting there – the exact opposite of what happened previously! You could tweak it so that it wouldn’t spawn unless the area had been empty of player ships for 30 seconds or a minute, to avoid camping… or even have it spawn in one of a few key areas, rather than just one. But I LOVE this idea as a way to include superweapons, but not have them overpower multiplayer play. What a great idea!

    Maybe I’ll copy this post into one of the development sections to make sure the devs see your idea… but I really, really love it ^ _ ^

    On the separate turn speed question – I do think it might be unfair to have a console player with a gamepad facing a super-turn-speed-mouser, but this issue has probably been faced by many games before. I believe in many FPS games there’s a bit of “auto-aim” applied to console players, and that they don’t compete directly against PC players. Not sure what the best solution here is, but I won’t be using a mouse, regardless ; ) (I might try a gamepad, though – I’m interested in trying this game both on PC and Xbox One). It does seem like it would be nice to have cross platform multiplayer, if possible, so as not to fracture the player-base… hopefully you can learn from how other games handled this to make it “fair” enough = )

    #15816
    SiriusSirius
    Kickstarter Backer
    Topics: 9
    Replies: 406

    Yeah putting the big powerups in a “hot” zone worked well for the likes of UT. I would say it’s most effective with one-shot weapons and boosts, like missiles or overdrive powerups; it wouldn’t work so well if it’s a primary weapon you can just run around melting faces with forever. But again, this is why secondary weapons are better places to put the overpowered stuff – it’s harder to hog them.

    One other thing – fixed weapon spawn points are almost necessary in CTF mode. Otherwise a failed base push from one side leaves all their weapons down there, and they’ll be very hard-pressed to get them back. Once that happens you might as well forfeit.

    Sorry man, but as long as it’s online you’re inevitably going to experience some rubber-banding if you want a fair game.

    I don’t really agree that this is inevitable. You can get perfectly fair games of D1 today that don’t behave this way. Of course, you could counter that the only reason D1 holds up is that it’s not popular enough to attract hackers; but you could counter that with that Overload is, being realistic, not the next Call of Duty either, and furthermore that doesn’t prove you can’t nail it down without rubber-banding.
    (P.S. Also, there is still cheating in those popular server-rules games. If they’re determined enough, they’ll find a way. I don’t think it’s a good enough reason to deliver legitimate customers a substandard experience. Especially since travel-time weapons seriously do make a major difference to how annoying this is.)

    One thing I can weaken from my opening post is that there are certain degrees of “badness” you can get away with. D3 was pretty wonky as far as the hit detection goes; people still played it in multiplayer. Some people, anyway. It wasn’t as popular as its predecessors at their peaks, but it’s probably unfair to say that the networking model was the only reason for that.

    For the movement thing, I mean, as long as whatever is being done is good enough to take out the jitters, it’ll work. I’ve played plenty of first-person shooters where this is just not a problem. In Sol Contingency, it isn’t a problem. In D:U, for whatever reason, it is a problem. Let’s keep Overload on the right side of the fence.

    #16254
    mickey1mickey1
    Participant
    Topics: 0
    Replies: 16

    I am some people. This topic is very important for multi-player to be fun. This is almost a constant topic of discussion for some people. Thanks for bringing it up here.

Next Unread Topic:
Viewing 15 posts - 1 through 15 (of 72 total)

Forums are currently locked.


Previous Topic: Hello Multiplayer! Next Topic: Axis binding bug with Saitek Cyborg 3D Gold