Win Conditions to prevent the stalling of games.

Maybe put victory point like metal value losses vs current stuff cost? For example enemy lost 100k in units in game and they have only 8k left so countdown started. And also every resign deepens that points.
Metal value negative thing however is about that enemy lost detris but game is still not lost for them because they have just better eco and they can feed more. Still have feeling that win conditioning should be around how much unit cost have between teals and when there is very large difference. Like my team have 100k in cost but enemy 8k in cost. So my team won.

I've removed storage from future wars. Should try out unit based storage in 0.18. If it plays well, it's a pretty simple system to integrate into base game. Problem is getting people to test how it plays.

quote: we might also do it through metal income. As having less mexes might not necessarily mean losing due to the exsitance of supermexes and overdrive, so pure metal income is a better measurement of economy.

I think victory conditions need to be clear and consistent. Pure metal income is bad because it can fluctuate without the player 'doing anything', which makes it hard to tell how much territory is required to meet a given threshold. Players would be incentivised to stop production (boosting their overdrive) or even to reclaim their own structures at some point. Using any metal income besides base metal income seems to cause too many problems.

I started writing this thinking that base metal income was a good candidate for territory victory. It is easy to calculate 75% of the total base metal on the map, and its maximum doesn't change. I have rethought this position though, simply because tying victory to a number of mexes is far clearer. I don't think supermexes are a problem since they tend to be in the middle of the map or arranged symmetrically.

I see where objections such as "might not necessarily mean losing" comes from, but I am not giving them much weight. There are two ways to think about a new victory condition:

• The victory condition is meant to predict who would have won (ie, destroyed all the enemy units) and truncate the game when this becomes reasonably obvious.
• The victory condition is how the game is won, with no consideration for who could have otherwise destroyed all the other side's units.

I am using the second approach at the moment because I don't think the first is useful at this point. As such, I am trying to judge victory conditions by the types of actions they encourage and whether the resulting games seem fun.

quote: I feel like anything that is a win condition should be visible to players the entire time, which could reveal more information about the game state than might be desirable. For example, revealing the current balance of metal income, or mex control, or total cost of living units, is basically giving each team some important economic information about the other, which would definitely affect how games are played.

I would be wary about adding global information, but luckily I don't think win conditions have to be fully visible all the time. For example, a 75% metal spot victory condition could create a warning when the other team is above 65%.

quote: Designate Commanders and Factories as "Command" units (along with other units that the designers think are significant enough). If only one player or team owns any Command units, a command-victory countdown begins; at the end of the countdown, the player/team with the Command units is declared the winner. The countdown stops if another player/team builds or regains a Command unit, or if the player with Command units loses all of said Command units.

I think losing when you have no commanders or factories is the cleanest version of this. The rule becomes very messy once the list of command units is extended beyond commanders and factories. To preserve some FFA behaviour the rule could be you win if you are the only allyteam with commanders or factories, but perhaps FFA would benefit from easier player elimination.

quote: I have another one of my stupidly dragged out game. that match was Super speed metal and our team had won, except some lobster decided to take the comms of their resigned allies and terraform them into a deeeeeep pit (like starlight-proof levels of deep). I didn't know how the game ended as I lagged out but I do know it got dragged on for like 30 minutes.

This is beyond the scope of this discussion. I think there is general agreement that a side with a commander alive shouldn't lose.

quote: Is this a solution in search of a problem? How often does someone hide a flea just to delay inevitable defeat. But if this is a serious problem, just have a 60s timer or something start when one team has 100x the metal cost in living units as the other team. A small number of hidden fleas will quickly be overwhelmed in terms of material cost and you don't need to show the players until it has happened.

Overwhelming Advantage is a 25x multiplier, and barely ever happens, so this rule would do very little.

The problem as I see it is one of psychology. The lack of a reachable win condition in ZK puts the responsibility on players to decide when to stop playing. The game is training players to be on the lookout for when the game is hopeless, when to give up. The game ends when one side feels like it is pointless to continue. This all seems pretty bad on a psychological level.

If games concluded themselves, without input from the losing player, then players would be free to hang in there as long as the game allows them to. Losing the game while still trying seems better than losing it at the moment you give up.

Most of the win conditions discussed here would not have triggered, but I won this game despite all but one of my team resigning besides me and my team was pretty far behind as well. I am glad that we didn't resign even though a majority of the players on my team already had.

I think the most reasonable condition is: A team has won if its unit value is at least 3 times as high as the sum of the other teams. If it is 2 times as high, a global warning can be shown. In contrast to income or mex number, this condition considers edge cases like only detriment left, speedmetal, metalmult=0 and high OD base in map corner.

I like malric's idea of running a test. In this case, the teams' mex numbers, incomes and unit values at the moment of resign should be stored and compared.

quote: The problem as I see it is one of psychology. The lack of a reachable win condition in ZK puts the responsibility on players to decide when to stop playing. The game is training players to be on the lookout for when the game is hopeless, when to give up. The game ends when one side feels like it is pointless to continue. This all seems pretty bad on a psychological level.

If you set other win conditions, then this will simply change the point where people think it`s hopeless. "In 3 minutes, they will get all those mexes and our team is not able or willing to do anything about it -> resign-vote."

Especially:

quote: The lack of a reachable win condition in ZK puts the responsibility on players to decide when to stop playing.

I think this is actually pretty desirable over the game itself forcing the players to play on or to quit. If i couldn`t resign for example i would sinply exit the game or go afk.
I mean, I can see where you are coming from, but you don`t seem to take into account that the feeling of being in control about one`s own fate seems to be a general desire of the absolute majority of humans.

quote: The aim is to remove the emotionally draining part of trying and still losing by making the game forcibly end before you drain yourself too hard.

Some people like to fight against odds. I am sure there is a "distribution" of how much you want and enjoy fighting and how much you get emotional drain. I honestly don't get the ones that seem to get high emotional drain (hence, they want to abandon) but still not resign and watch the game as a spec. Many times without them the game can even be won, but I feel like they have some "fear of missing out"...

quote: If the encourages everyone to play to a particular point, a point that is a compromise between resigning what would be called 'early' or 'late' in the current system,

I think you miss the cases in which people want to resign and the game is a win for that team! I have seen many games (sadly don't have numbers) in which there are several resign votes in the team that wins. I think you can agree that can't be included in the "resign late" category, hence I am afraid another system will still suffer from completely misguided wishes to resign.

Maybe we should have another system like an in-game button with "want to play another game with at least 6 players and resign current game". When the number of players is reached, everybody on that queue is automatically resigned and moved to the new game. Such a system would

• reduce waiting times for people in the lobby that want to play
• eliminate the need to wait as a spec if you resign "Early"
• might encourage multiple hosts (oh, the holy grail!)

One more idea: allow in the first 3 minutes of a game more people to join (if they can be balanced). Not sure if this is possible without engine modifications (probably not), but would be a nice concept because it would give you a bit of time to "assess" the current situation and decide if you want to "switch".

Edit: and if we are at crazy idea, would be cool to have continuous game, like be able to join at any time (subject to some balancing) and join an existing team. But that would need engine save/load state...

quote: I'm pretty sure adding spectators to teams is possible on the fly.

Do not think it would help much with the scenarios above. If you are halfway through a game that you think is lost and a new team games "just started" (meaning couple of minutes), you can't join the new game. You could be offered the choice before the start (the "I want to switch game queue"), but not after.

quote: I'm pretty sure adding spectators to teams is possible on the fly.

It is. See: Commshare. Use Spring.AssignPlayerToTeam

https://zero-k.info/Battles/Detail/953939

look at that battle. i guess we would have been shut down by artificial win-conditions.