I'm prepared for challenges to this, so feel free to speak out: after 10 players the game can be reasonably expected to run poorly on both client and server sides. Won't stop us though. We do it anyway because it's super fun. Don't get me wrong- the lob pot is a hot mess in the best way. It's fucking bedlam... But I think it should not gather the focus of our valuable (FREE) development resources. Can you optimize anything at all as player count grows arbitrarily large? correct me if i'm wrong but - impossible Can you optimize something for exactly 10 players? Empirical evidence shows yes. It seems to me, any adjustments made to performance without regard to a player count can jeopardize the scope of performance, so to put it simply; in my head zk should be optimized for a maximum of 5v5 and any expectations after that are wishful. As always we can just jump in a lob pot 30 of us and have the game run terribly but not care.. The other part of this is I don't get why small team games aren't popular. 2v2, 3v3 to me is where this game is the most entertaining competetively, the options 2 players have leaves such suspense! the lob pot? all facs, hub present and skies are closed after 4 minutes. It's hitting the ceiling right away. spider player is dedicated looking for scythes with flea spam- this is not a waste cause there are guarunteed to be scythes incoming. In 2v2 you can actually use their scouting utility, and take down a mex and wonder if maybe scythes are coming. A metric I'm using for this are the air factories in general. These are support factories, and in the lob pot it's all out there immediately. You should have to earn air vision, or be liable if you go air start. My main objective of course is the development resources. GF just put in a fat heap of love into this game and we've got nothing but more work for him and co. If we can put (even temporary) cap on the expectations of performance, we can get back to unit balancing and vision theory.
+4 / -3
|
|
|
This is partly why the small teams poll has been added at the end of every big teams game. It should be an easy way for players to find a small teams game without missing out on the one big game if their small teams doesn't happen. Unfortunately from these poll results it seems like not even 10% of the big teams players are even interested in smaller games. The matchmaker isn't being used that much either. If we were to actually look at the statistics we should probably be putting all our resources into big team games and abolish the competitive game modes..
+7 / -0
|
Scale back mex/com income based on players. If every 16v16 had just as much units as a 1v1 the game wouldnt lag! The game would be much more noob friendly if everyone only had 2/3 units to micro =0 [Spoiler] /s
+2 / -1
|
High player count in a game is something ZK has that other RTS games don't. If you think ZK big team games are meme-worthy terrible, try the Eugen (War Game, Steel Division) 10v10 where every round is stacked by people clicking on your profile and checking your win rate. These (premium) games are mostly empty now, maybe in part because their execution was flawed by metagaming. ZK is already better. People love to complain about the things that happen in big team games, but they keep coming back.
+9 / -0
|
Have the special team game server flag/ description: to encourage better cooperation and skill training. Special Type 1: Request to make a widget to add shared battle zone, in which not grouped combat units are "shared control in one team"; could make similar widget for constructors as well. Special Type 2: Restrict player number according to map size. Could vote to change the restriction.
+1 / -0
|
So u want a autohost with "!maxplayers 10" - should be possible. Lobsterpots: Client: Get a PC not made out of potatofrags. Get a Internet connection not made of Bamboo. Host: Get a Server that can handle it. (maybe a virtual server 4 15€/Mounth) Get a game/engine that uses more than 1-2 cores of the clients PCs! (sadly (not) possible and asked 2 much - i know) The reason why small team games aint that popular? cuz u have to do everything yourself! Grid, AA, Anti, SpecOps, Basic defence and Microing. In the big team games everyone kinda got its task and u can focuss on the gameplay or microing ur units more whats more fun! And YES there still do exist peeps who dont bulid any defence at their factory and wonder why it got destroyed in minute 3 if they aint aleready asking why/how their com got lanced at front. The difference in small team is that its 1/5 or 1/10 of the players whos facing temporary game over if that happens (also the eco lose on small maps where every com matters). In big Team games its 1/10+, 1/16 or even 1/30! (Yes ive played kinda 30 vs 30 before dunno if it were exactly 60 but 50+ fore sure! Worked fine on my PC and i LOVED it! Yes there were sometime little lags which wouldnt happen if zero-k would use all/more of my 8 4,5 ghz cores!) Decreasing the metal? Hell no! Then ull have ur 5 units and thats it. Itll also shift the community behaviour to the worse! (Saving/rescuing allied units? - no i want the reclaim). Also the game will become a scynthe only game. Shared units aleready exists! It can work suprisingly well or end up in a toxic saltmine. To the map size im against it. There small maps playable with large Teams. And there Huge maps playable with 5 Players. But there also maps which are no fun without a certain amout of players.
+0 / -0
|
Having the option to do everything yourself is one of the benefits of small teams. If you're playing reasonably well, you can retool. In a lobster pot, you can't do everything even if you see the need to act and no one else does. If you try to do several things at once, you can't put enough resources into any of them to accomplish anything. In that respect a lobster pot is a harder game than a small team match. You're usually stuck with what you start with and more dependent on everyone else to fill the gaps. I find that impotence far more frustrating than juggling the challenges I need to in small team game (or 1v1 for that matter). That's probably where some of the current ragekicking fetish is coming from. There are so many times I've tried to take ground- but been stymied by the fact that our air is elsewhere or simply routed- which means I need to wait until I have sufficient AA and lose the moment (and all the while the other side is reinforcing or fortifying.) It's great that Zero-K offers so many game styles, just wish more people shared my preferences.
+1 / -0
|
People need to understand that they can't expect the game to run well in big games when they still use a Core 2 Duo, or some old AMD processor. If your cpu can't keep up you will start lagging. ZK is mostly CPU single core performance bound, so even getting a high GHz core i3 would be a good idea. I have NEVER had any performance problems with the game in any game, even in silly speed metal games with plane spam or 16v16 lobsterpot its still >20FPS. (Admittedly I do have an i7 8700k @ 5ghz and 1080ti) As to your comment about big team games, I would say that ZK would be dead if it wasn't for them. I play the game to have fun, sometimes I will play seriously, sometimes troll. In a big game you can get away with having a bit more fun, in small games you are balancing so many things it is very tiring and you lose the community feel from the game.
+4 / -0
|
Suggest how to make 5v5 much more common without adverse side effects. If hosts are simply limited to 10 players then I'd expect there to be a single host with many spectators.
+0 / -0
|
quote:
Suggest how to make 5v5 much more common without adverse side effects. If hosts are simply limited to 10 players then I'd expect there to be a single host with many spectators. |
Can you explain how or why this would be an adverse side effect? It seems to me like engagement is engagement. Watching other (good) players play is a great way to bootstrap your own learning. Spectating in-game is a more connected experience than a Twitch stream. If this isn't good enough, what exactly is the goal? The 16 player limit feels pretty good. Has anybody been observing the battle list while the teams lobby was limited to 16 players? I saw a separate team lobby with about 4 players the other day, at a glance. There were also some spectators in the 16 lobby, at least some of whom already know how to play.
+0 / -0
|
I went to battle list as soon as I posted this, and the 8v8 [A]'s were in effect, so I thought I'd let the thread marinate for a bit. I'm inclined to just thank everone for their input and agree that lob pots are too fun. I don't experience but a little network lag thanks to my los r/f isp. other hardwares are sharp... I'm more speaking of a design ceiling than performance, though; also in my head that impossibilty of accomodating the higher player count reliably is always there and just needs to see a model. e.g. Q: "is eighteen players on porky_islands going to lag?" A: "Yes, the [hypothetical] design of zero-k spring is 16 players, expect 5% input delay" - or something also A: "there's going to be repeat facs fighting to split 10m/e/s each" I just guess that if everybody's expectations revolved around a defined number, the programming has more of a target, map sizes/categories get clear definitions, people get to spectate games, one+ role/factory per player and strategic air (not less). more...? Then of COURSE you have, everything after that, the lob po+ , the 9v9 + which will still be there and not change other than defined as larger than can be expected to perform well, krow rush lulz, Firepluk with a nuke already done and a disclaimer that things can get a little design cramped/spammy in this mode.
+0 / -0
|
The reason why people play the giant team games is imo two reasons: 1. It's always available if there is anyone playing at all. This game has long deadzone type periods so people tend to tumble into the 20v20 games because at least there WILL be a few games you can play in a row. If you play team MM and get a game, maybe there is no game waiting for you afterwards. aaaand... 2. It's very slow to lose stars in lobster pot. If Godde goes and trolls in 10 1v1 in a row or something he could lose his purple star, in the 20v20s that is almost impossible. PROPOSALs: 1. Make it impossible to lose stars, only gain them. This is a purely cosmetic change but I think it would have a big impact. 2. Do something about the massive deadzones. For example, you could allow people to get up to a certain (low) rank by beating AI or various chickens. Something has to be done to make the queue actually active and more "anything goes" (like the bigger games) then people will use it.
+2 / -0
|
quote:
2. Do something about the massive deadzones. For example, you could allow people to get up to a certain (low) rank by beating AI or various chickens. Something has to be done to make the queue actually active and more "anything goes" (like the bigger games) then people will use it. |
Are you saying you would prefer having AIs populate the lower MM ranks over having the guarantee of playing against humans?
+0 / -0
|
I am saying exactly that, I realise it is a bold proposal. However, this games AI is pretty decent and it should be able to have its own elo (and a randomized name to prevent cheezing it)?
+1 / -0
|
The technology exists, I've been testing upcoming changes to the rating system on the test server using accounts controlled by CircuitAI. You can see them here.
+0 / -0
|
Why not just make a set amount of missions to be completed and you're automatically granted rank/elo that is most populated? Instead of having to go through AIs comp-stompting. If I'm not wrong there's a fair bit of challenge missions that are hard to accomplish and wanna-be-pros can freely go through there and not get wrecked when they get to fight players.
+0 / -0
|
quote:
Are you saying you would prefer having AIs populate the lower MM ranks over having the guarantee of playing against humans? |
This exact thing was done to a game in recent history. In 'The Culling', a small-scale melee focused BR game, bots were added to their MM queues masquerading as real players in an attempt to make up for a too-small matchmaking base. Before jumping into this, it would be good to look at the mistakes they made, and see if they apply in a ZK context if implementing something similar. From what I remember, their veteran player base was hit the hardest. Bots weren't optional in MM, and a lot of vets outright refused to have anything to do with bots, going so far as to leave the game instead of playing against them, even with other human players in-game or in queue. Players that enjoyed the game for its social aspects (it had in-game voice chat with enemies) couldn't do that with a bot. It had the opposite effect of what was intended.
+6 / -0
|
Good point   Porkchop and the reason why I'm really wary of using any number of bots in MM games. I do think bots in ZK have the advantage that they're not cheating - You can do anything a bot can do, technically you can just automatize the game to the point where you are a bot. In first person games, bots always have trouble seeing and interacting with the terrain as a human would. Still, bots would take away from the human interactions in MM, especially in team games. They're generally mostly a bad compromise, that I wouldn't take while the proper solution of spending more time in marketing or MM quality improvements is still open.
+0 / -0
|
Having a tickbox to enable/disable bot matching, defaulting to disabled, and matching a player to a difficulty of Circuit based on their elo could still be useful. Especially if that difficulty of circuit had a shared global Elo rating so it could gain/lose same as players. That way players could find their starting elo by working up through the bots. Actually pretending bots are players is a whole different matter.
+5 / -0
|
