The discussion has moved over to Discord

Home Forums Overload Multiplayer What Overload multiplayer needs to succeed

RSS Feed
Viewing 15 posts - 46 through 60 (of 72 total)
  • Author
    Posts
  • #16665
    CHILLYBUSCHILLYBUS
    Kickstarter Backer
    Topics: 61
    Replies: 173

    I suspect that for every player like me there are hundreds or even thousands like Chillybus who don’t enjoy limited mouse movement so for the success of the game it might be necessary to allow mouselook-like turning in MP. (Didn’t work out in d3 but those were different times)

    I’d hope with 20 years of hindsight we can find a happy medium between the two of us!

    #16667
    Hunter
    Kickstarter Backer
    Topics: 6
    Replies: 69

    Birds was king in early years and Zero really was in charge in later D3.

    Awesome pilots yes, but there was a ton of players back in the D3 days (99-early 2000’s) that rivaled those.

    #16685
    Pilot
    Kickstarter Backer
    Topics: 3
    Replies: 23

    I’ve been having some concerns about the turn rates things(stick vs mouse), but let’s still comment here, that I don’t think there is nothing wrong with how the ship handles currently with joystick. I actually love the way it feels. And now when I just yesterday bought a new joystick which actually has a working hat(until I break it 🙂 ), after little practicing on ace was kinda suppriced that was able to add +100 on insane in one map. Being able to dodge seems to do a lot.

    #16690
    SiriusSirius
    Kickstarter Backer
    Topics: 9
    Replies: 406

    How about a server which can only accept players who have NOT opened any insane maps?

    The catch with this is the Jediluke scenario: someone who has never played the single-player mode (or even just not to a high difficulty level) and goes straight into multiplayer. You’d need another way to distinguish them. Some kind of pilot rating system could potentially be used, although it’s still probably possible for someone to create a new unrated profile and break into games they’re not supposed to be in.

    Getting behind other players or chording well should be rewarded to some degree, but too much control limiting greatly removes counterplay options and strongly favors just flanking or being flanked, which is frustrating.

    Reflecting back on how this worked in the games that had it (D1/D2 mostly)… I’m pretty sure I recall flanking being an issue for people in free-for-alls, but mostly at lower skill levels. One of the skills people learned as they progressed was how to avoid showing your back to places you were likely to be attacked from. Another was how to evade attacks from unseen angles long enough to turn around or reach cover (this didn’t always work, but most of the time it did limit the problem to just “took some damage”).

    In 1v1 it was definitely a factor, but more in a “tool-in-the-toolbox” sense than everything revolving around it. There are multiple ways to score a kill, and sneaking up behind someone was definitely an effective one if you could do it. But you couldn’t always do it, because your opponent knew you might be trying for it. You’d just look for opportunities, and if “I saw him but he didn’t see me” was there, then great – but maybe it was “he’s in a confined space near a closed door, I might be able to flood it with quads and land some damage before he has time to open the door”. Or “I think he’ll come out of this door, let’s drop a smart in there and see what happens”.

    Removing the flanking component will thin the game out, but as Birdseye discussed that might be a sacrifice we need to make. Hard to be sure.

    #16691
    luponixluponix
    Participant
    Topics: 4
    Replies: 24

    How about a server which can only accept players who have NOT opened any insane maps?

    opens the can of “smurf” problems where people create accounts and dont unlock insane to be able to join these servers.
    only solution i see is some kind of MMR as in chess .
    A number that tries to reflect your skill and is based on how often and against who you win but that sounds quite a heavy bit of work -> not realistic

    •[OOTS]•

    Prepare for Multiplayer

    #16694
    hurleybird
    Kickstarter Backer
    Topics: 6
    Replies: 41

    Removing the flanking component will thin the game out, but as Birdseye discussed that might be a sacrifice we need to make. Hard to be sure.

    “Thining” may be a bit of a reach. Obviously, the more that the outcome of a match is determined by aim, the less it will be determined by movement, and vice versa. But the assumption that positioning is more important than aim, or vice versa, seems to be an unfounded one.

    When it comes to actual skill ceiling, which is what comes to mind when I think of how “deep” or “thin” something is, you want as many different dimensions as possible that are as significant as possible. You don’t want to have one area that is overwhelmingly important at the expense of others. This is admittedly an oversimplification, but generally, the highest skill ceiling will be realised when aim and movement are equally represented. Mathematically, this is like multiplying the two together. 0.5 * 0.5 > 0.9 * 0.1.

    Of course, just having a higher skill ceiling doesn’t make one game better than another, and often a high skill ceiling comes with its own trade offs too. I certainly agree with the consensus here that FPS + significantly more involved movement is likely to be more appealing to non-6DOF FPS players than FPS – a significant amount of aim + a signification amount of movement. The former is compelling because it’s a super set of the FPS formula; a way to evolve it so so it’s more skill based and interesting, albeit at the expense of approach-ability. The later is less comparable and more of a “totally other thing.” Even less approachable, and arguably (assuming the implementation goes really far in the direction of movement at the expense of aim) not that much deeper, if it can be called deeper at all and not just… different.

    #16697
    SiriusSirius
    Kickstarter Backer
    Topics: 9
    Replies: 406

    It doesn’t de-emphasize most parts of aim, though, just the “twitch” aspect. Aiming is still a strongly necessary skill in D1, as anyone who’s seen the likes of Cyrus light people up with Vulcan can attest. But there are more facets to it; with any gun except that, being able to spin on a dime and accurately shoot won’t matter much because they’re slow weapons. You have to lead them, and you probably won’t benefit from a higher maximum speed when you’re doing that. (With the possible exception of extremely close range, but in that situation you’re also unlikely to miss.)

    I’ve played enough D1/D2 to know that those two games wouldn’t gain much from an unlocked maximum turn speed, and certainly not more than they’d lose. What’s less clear to me is that the additional tactical complexity is compelling enough to justify the reduced accessibility. In Overload’s case, it really might be something better added later if needed than imposed up-front whether needed or not. That’s why my primary concern is that whatever happens, it’s fair to all users (particularly on console/gamepad); the game can still be good with or without fast turn.
    And even then… somehow people play Fortnite on phones where there is a demonstrable disadvantage. Sometimes you just can’t predict how players will react. 😀

    #16858
    ayleejennayleejenn
    Participant
    Topics: 1
    Replies: 4

    I’d hope with 20 years of hindsight we can find a happy medium between the two of us!

    This line caught my eye. Why: I like the reference to experience. It lead me to think about something connected to a game’s success. Yes, the concerns brought up (mechanical, balance, programming, networking, etc…) they are all valid. But there is one component that I feel to mention: Developer Motivation

    I think that a community can have a profound impact on that motivation. If it is a mature, experienced community like this one seems to be who does not just expect the developers to feed a good game to them but also takes care to cater to the developer’s experience I think it goes a long way.

    Simply put: I think it is important not just to win player-hearts but also to win the hearts of your developers 😀

    The developers are in it with their heart and soul. But if they are even more so motivated then I think the game will only benefit from it. ^.^

    Prepare for Descent …

    #17014
    LotharBot
    Kickstarter Backer
    Topics: 1
    Replies: 133

    When it comes to actual skill ceiling, which is what comes to mind when I think of how “deep” or “thin” something is, you want as many different dimensions as possible that are as significant as possible. You don’t want to have one area that is overwhelmingly important at the expense of others. This is admittedly an oversimplification, but generally, the highest skill ceiling will be realised when aim and movement are equally represented. Mathematically, this is like multiplying the two together. 0.5 * 0.5 > 0.9 * 0.1.

    One of my rules of applied mathematics: often solving the equation is the easy part. The hard part is making sure you’ve modeled the right things so your equation solves the right problem.

    In terms of deciding what it means to have the “highest” skill ceiling, if you treat aim and movement as two equally-deep components, you want them to be balanced. If you treat aim as having two sub-components (say, precision and lead-prediction) and movement as having three sub-components (like orientation, reactiveness, and speed) and you want to balance the sub-components, having aim=movement would result in its components being overemphasized. You’d want movement > aim because its sub-components would have more combined depth. If aim had 6 components and movement had 3, you’d want aim > movement to have the highest skill ceiling. And so on.

    Of course, you’re correct to point out that “highest skill ceiling” isn’t necessarily the right goal. Accessibility matters. It should feel good to fly, and every aspect of flight should be a place to find joy. To me, every aspect of flight (broken down into components like aim+movement, and sub-components, and sub-sub-components) should matter. Every aspect of flight should be a way for pilots to gain an edge. Every aspect of flight should be a way for pilots to make up for shortcomings in other areas. IMO that’s what makes a game deep rather than thin, when *lots of different skills matter*, and a high skill ceiling is a side effect. If twitch-aiming ends up being a much more important skill than many other skills combined, it will detract from the overall depth.

    Now, I don’t think “make mouse suck” is a good solution. There are lots of other potential solutions. Modifying the relative power or speed of different weapons can make a big difference. Likewise modifying the hit models for the ship, the size of projectiles, the tracking behavior of homing weapons, etc. can result in greater or lesser emphasis on the twitch-precision aspect of aiming. Even though D1 and D3-as-usually-hosted have essentially the same turn speed limits, D1’s most precise weapon is vulcan, which is not quite hitscan, fires in a cone, and is relatively low DPS — but D3 has the mass driver, which is hitscan, perfectly precise, and does heavy frontloaded damage with a knockback effect. So D3 is a much more “twitchy” game.

    Ultimately, we have to see how Overload plays, and (hopefully) make adjustments to make sure we have both accessibility and maximum depth.

    #17016
    LotharBot
    Kickstarter Backer
    Topics: 1
    Replies: 133

    On a different note: regarding Sirius’ initial point about superweapons — every type of weapon has a particular effect on particular game modes. The point isn’t to say “no superweapons ever” so much as “understand the role of superweapons and distribute them accordingly”.

    A few powerful missiles with predictable spawn locations can drive the action in a free-for-all or a team-vs-team game, giving players a reason to continue to return to the same area, or perhaps to take potshots into that area, or to avoid it entirely, or to set up ambushes, or … you get the idea. There are lots of tactics that can evolve around the control or use of something on the scale of a D1 mega missile.

    When very powerful missiles are spread all over the place and easy to come by, the action tends to become more focused on getting a large number of shots off (running away to find the big missiles), missile-chucking toward high-traffic areas, and learning specific tactics (often involving heavy use of cover / running away) in order to avoid dying to others’ chucked missiles. Engagements tend to be less face-to-face and more standoff-and-suddenly-your-area-gets-nuked. There are people who would call this a “no-skill” type of game, and I would disagree with them. There are a lot of specific skills. But it’s a more specialized and less globally-interesting type of game than one with more frequent direct engagement.

    In games that have a superweapon primary that you can pick up and keep using for a long time, the dynamic turns into king-of-the-hill. Whoever has the big gun tries to string together a bunch of kills before they lose it, and then the next person tries to string together a longer kill streak, and so on. This tends to magnify small advantages, and result in a much wider score gap than with less-powerful weapons.

    None of these are specifically *bad* dynamics. But I’d recommend a focus toward the first dynamic in basic games that new players are likely to find just by clicking “play pvp”, and with drastically reduced superweapon frequency with fewer players.

    #17071
    hypersonic
    Kickstarter Backer
    Topics: 18
    Replies: 220

    What is critical is that all players see the same situation for each timestamp. Though when they receive that timestamp can vary slightly (a few milliseconds.)

    Regarding #1

    Look at it the other way around: if it looks like I hit them, then I hit them. This goes both ways regardless of each’s connection to the server.

    Regarding #2

    Authoritative server is critical, to keep all entities in proper position for each timestamp. Quakeworld lets you see your movement before it happens on the server (client prediction) that seems to work well for the appearance of an immediate response. Though when enabling client side prediction just remember that your real self is where your screen displayed where you were a few milliseconds back. Real self is your position that corresponds to specific timestamps that is the same for all players. (I’m not familiar with Descent Underground’s client side prediction, perhaps it’s not quite the same as Quakeworld’s.) It might not handle packet loss as well.

    Bandwidth since the 1990s has increased by leaps and bounds, too bad latency hasn’t really improved, <10ms ping times were possible in the 1990s, but that’s hard to find in 2018. Latency improves reaction times. With added bandwidth servers should now update at say 60 times/sec, and coordinate position should be more precise now (such as instead of just a value of 3, now it’s 3.18.)

    EDIT: much of this subject is contained in this thread from 2016

    Brief mention of multiplayer

    #17074
    SiriusSirius
    Kickstarter Backer
    Topics: 9
    Replies: 406

    I already covered why dodging is the more important part of the equation than hitting; Overload is dominated by travel-time weapons. It’s a lot less frustrating to miss in that kind of game than it is in the likes of Quake, and conversely a lot more frustrating for dodging attempts not to work.

    My understanding regarding the ship movement thing is that Overload is using a fairly sophisticated client-side prediction system that should hopefully smooth out these problems. Keeping fingers crossed it’ll work out.

    #17077
    hypersonic
    Kickstarter Backer
    Topics: 18
    Replies: 220

    I think it’s important to consider what exactly it is that you are dodging.

    In D1/D2 on the other player’s screen they could be pelting you with laser after laser. While at the same time on your screen you could be dodging laser after laser. Essentially they are firing at what amounts to a decoy of yourself and you are dodging lasers aimed at your decoy. It’s more challenging to dodge lasers aimed to intercept you rather than your decoy.

    With client prediction your real self is a few milliseconds back as movement inputs made by clients take time to reach the server. This is a must so that all players see entities (ships/projectiles) in the same position for every timestamp. With peer to peer there are essentially multiple different realities, with client server there is one.

    History rewrites like
    https://wiki.teamfortress.com/wiki/Lag_compensation
    I suppose could be an option perhaps for firing and dodging, but it does seem complicated to code for. Basically giving high latency players time travel abilities.

    #17112
    SiriusSirius
    Kickstarter Backer
    Topics: 9
    Replies: 406

    It’s more challenging to dodge lasers aimed to intercept you rather than your decoy.

    It sure is! That’s why LAN can be such a demanding environment – especially since the other guy doesn’t have an out-of-date view either and knows more accurately where to shoot.

    But when we’re not on LAN, I’d much rather deal with a lesser challenge that I can see rather than something of any difficulty I can’t see, and instead have to guess the future in order to account for. There’ll be future-guessing when I’m shooting still, but in Descent-like games there’s less pressure on that side of the coin. Most shots miss anyway, so messing that up hurts less.

    Yeah, high-ping players are weird to deal with, that’s a definite downside. But at least it’s a two-way street – they have to lead too. Peer-to-peer or fake peer-to-peer I will at least still consider playing in their games – D3-style client-server, well… that was why I gave up on D3. Dealing with lag there sucked *bad*. D1/2 were still fun on the other hand, the tactics just had to change.

    #17119
    hypersonic
    Kickstarter Backer
    Topics: 18
    Replies: 220

    Ya LAN or near LAN (internet with <10ms ping) is awesome, whatever the network architecture.

    I was thinking lag compensation for dodging could get really weird. It’s already weird for firing (dodge behind a wall, then get hit from someone you can no longer see.) I suppose the client could tell the server what server time stamp (which is already history once your client received it) you made your movement inputs on, and by the time the server receives your input it could already be say 200ms after that timestamp, depending on the lag. Say the server keeps a buffer of all frames back 500ms or so. It replays the last 200ms this time with different movement inputs from you (movement inputs teleported back in time.) Then it determines: OK, with this set of movement inputs you dodged the projectile. So on other people’s screen first you were hit, then suddenly you teleport somewhere else un-hit. Like peer to peer this creates multiple realities (though one gets overwritten by another in this case), and in affect has people shooting at decoys as well.

    Another idea would be having the server add artificial lag to low latency players (using incoming and outgoing buffers) so that all players have roughly the same lag. But then people would just get mad at the high ping player for lagging up the game.

    Could turn off client prediction and play like netquake, where it does display your real self (it’s not back in time a few milliseconds), but movement input response is delayed, a horrible feel when you have bad lag!

Next Unread Topic:
Viewing 15 posts - 46 through 60 (of 72 total)

Forums are currently locked.


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