before they even open the envelope on the unit cap idea, I hope we'd have key shortcuts to find t1 units, destroy 1/2 of what's selected, and other niceties to help trim the ranks. oh, and painfully obvious alerts as you approach the cap... please... thinking about trying to do that in the middle of everything else is making my head start to hurt.
If the actual goals is simply performance, I agree we should wait till later whwn the devs start their full effort on networking and client side optimizations. From the devs goal with engine was to make it scalable, only restricted by the power of the server and client. And so far from prevoius large game tests (changing client files that resulted in 40+ player games) the tech performed admirably. So theoretically the engine can support more than a million units in the games as long as your system was powerful enough. This is much different from many other rts where vast amounts of units couldn't be calculated in the engine, and it would crash every time, because the tech wa never designed to handle that many units. Uber's is designed from the ground up to handle an unlimited number of scaled number of units to the tech you are using, and in a far greater ratio than most engines. This will be improved as they go back and change code and take out poor calculations, meaning it will be have optimized code and sped up greatly. Wait for that to happen toward the end of testing before requesting to add a unit cap only based on performance. As for a unit cap for game design and balancing, no. The game, like its engine, is built to be ever scalable. Reducing the number of units may have its benifits in micro gameplay, but this is a macro game where attention isn't on little details, but looking at the whole, or several locations at once. Along with the fact you would have to split a unit cap across multiple worlds, it hampers the expansionist experience. The only way it would work, is if it was so rediculously large it wouldn't hamper gameplay, and at that point not fix the current issues of preformance .
if there's a unit cap, there's gonna come a time when I forgot about that group of t1 bots... and I will have to choose whether it's worth delaying further unit productions as I wait for them to march to their death 1/2 way around a planet (where they'll only do minimal damage anyway), or get on with production and self-destruct them now.
The only way you can hit a unit cap is by not playing. This is not a problem that matters in any realistic way.
If it comes to it could this possibly be a server setting? Atleast then those who want the no cap, or those who want a cap for what ever reason can set it up. If its not really attainable as a setting then Id rather play without a cap.
Methlodis has already pretty much answered this thread. Going with a unit cap is a bit silly considering the whole point of the game is to have massive armies. The performance problem with the number of units is very heavily networking and inefficient calculation techniques. They are actively working on fixing any issues related to performance, and while every few updates the performance usually drops off, it generally gets better and better. No point asking for something so temporary considering the power of the servers and code when things just need some time to sort themselves out.
A good first step is maybe making them a bit less spammable, and somewhat less able to pack themselves together into a blob.
actually, I do see one benefit of a unit cap - right now if one army is huge, everybody gets a lag, especially those with weak systems / internet connections. That means that the player with the strongest machine has an interest in spamming cheap units to impose lags on everybody else, giving him/her a clear advantage not associated with skill. Even with optimizations, that would still be an issue, no? For the game to be competitive, everyone with a decent gaming system should have an equal chance to win, right?
Client only need to update information about units you see, same for rendering. It's doesn't mater how many units here in total and even if you see whole map there is number of ways how hardware load can be significantly reduced. E.g client can only render icons instead of actual unit models and then it's basically doesn't mater if there 10000 icons or 100000. For bandwidth there always will be some limitation between server and client. So after you have some recommended connection speed you won't get any advantage over other players who have this speed too. Players with better hardware will always have some advantages over other who don't have it, e.g players with two monitors will have advantage, but it's not reason to cripple experience for them. Anyway at this moment simulation performance drop is way bigger problem than client performance drop or bandwidth limitations.
Is this something you know, or something you assume because it would be sensible? All I have to go by is being in games where the other teams started complaining when my team started spamming units? I do not know whether the other teams had significant radar coverage of our base at the time. that's not what I'm talking about - in your monitor example, the benefit of my monitor does not depend on how many monitors the opponent is using.
Alright I just had a game where I ran out of virtual memory. I ran out. Of virtual memory. Cap needed.
I've said it before and I'll say it again, if lag is a problem, fix the lag, not band-aid fixes like units caps and such. Mike
Unit caps, or the lack of them, are not the problem here, folks. It's that whole "Game isn't done yet" that's the problem. In the meantime, try smaller, shorter games. Huge ginormous games that include kajillions of units will happen, but if you know in the current state of the game it breaks things.. my best advice is to not do those things for now. We will get there.
I can say for sure that in current version of game server do not give you any information about units you do not see visually or on radar. You can basically start game with AI, move to other planet and defend here. You'll notice there mostly no bandwidth in this time even if AI have tons of units.
It has been stated many times AND proven that the servers are the ones taking the load of simulating the game. Then sends that information to your client in which you can then choose what to do, sends that information back to the server and interprits it. The unit movement, projectile physics, fligght simulation, and most similar information is done on the server. If you watch any of the dev playtests recently they occasionally check to see how the server is doing under the load of simulating so much, and they continuously say that the server is doing fine. Which means its a problem with the client and more important the networking. The client issues are mainly do to render (try to draw so many units). These are fixed by not being optimized which is a continuous process, and as more of the game is done will be more and more of a priority, as more major features are completed (along side proper balance). And the fact that MOST EVERYTHING is being run off of the cpu including graphics that is usually is run of the gpu. Now the gpu is doing something now, but they are going to start using it for MUCH more including dumping much more on the VRAM. Once these changes are in place the client will run much better. The second problem is networking. Now this is still coding within the server, but its not server load from similating the game, its processing what information should be send to whom, and its not properly optimized to send that volume. This net code will be optimized in the next few patches (or so we've been teased), and will most properly fix alot of the problems. And as I said before, the server has been tested to run VERY MASSIVE armies in 40+ player games in previous patches when the code was half decent before additional features broke the code. And additional features always break code which is why during feature/asset development not alot of performance work is done. UNIT CAPS IN THIS GAME ARE DUMB. (If you only read this part you are an angry troll). ((and lazy)).
A million units is a really big goal. Give it time. There's plenty of other ways to make the team feel bad for now.
NO UNIT LIMIT. Terrible idea. (Although setting it might not be a bad thing.) Imagine trying to take other planets with a 100 units. It takes that just to make an army alone.
Units being OP is for 5 year olds. Let the dev's decide if the unit works or not. If they notice "Hey look this unit is ruining our game." They'll change it.
