Dhoomstriker
Girlie Builder
- Joined
- Aug 12, 2006
- Messages
- 13,457
Why would 3-5 UPT be necessarily any better? Would an AI necessarily be able to properly stack ranged Units to reach the maximum Units allowed on a few tiles before attacking? Or, isn't it more likely that an AI would, on average, in most combats, use 1 to 2 UPT against a human player using the max of 3 or 5 UPT and the human would thus easily win most combats? I would even go so far as to claim that a lot of people enjoy feeling that they can find ways to work within the rules to abuse an AI who cannot take full advantage of a ruleset (i.e. the human able to better understand how to fully stack ranged Units on multiple tiles, be it 1 UPT or 5 UPT, before advancing their attack, and thus better able to cope with the imposed rule) and that if an AI were produced that could take full advantage of a ruleset, a good majority of players would complain that the game is too hard.Sticking with 1 UPT was the stupidest decision IMHO with regards to building a decent AI. It's too complicated to code a AI with 1 UPT as humans can process by seeing placement and code can't easily replicate that. Not saying we need Civ IV SODs but should have gone to 3-5 UPT based on a support calculation.
Where is the evidence that a Civ game should have ANY limit on UPT? Where are the game editions where an AI can properly play with such restrictions? There is evidence that AIs can be competent and challenging in Civ games without limits on UPT, and yet a good portion of the community somehow got a "bad taste" in their mouths about the idea of no limit to UPT. But, it seems more like a feeling that some players have developed, probably for a combination of reasons, without there actually being a good solution for the AIs in the alternate approaches (1 UPT or any UPT limit). To be honest, I'm disappointed at how the community fractured around such a concept in a way that led to Civ 6 continuing to include a UPT-limiting rule, and I think that people getting vocal about how stacks of doom are supposedly worse is what led to such a situation.
Where is the actual evidence that stacks of doom ruined Civ games in a way that if the exact same game edition had a UPT limit that the AI would play better and that the game would be better? Evidence shows that AIs can leverage an unlimited UPT.
Civ 4 Beyond the Sword's AIs are capable of dealing with a modest attempt at using a stack of doom against it, and while there are many ways to exploit the AIs, if one is playing on an appropriate difficulty level for one's personal skill level, one does generally need to employ good tactics in battle. When it comes to using very large stacks of doom against the AIs, again, if playing on an appropriate difficulty level, it generally requires a technological leap in military technology or a sufficient number of turns with slowly advancing stacks of siege units in order for a stack of doom to truly dominate, which, on an appropriate difficulty level for one's skill level, presents economic challenges and building stacks of units is a far cry from an instant "I win" button.
Maybe it is time for the community to agree that we're going to get better games without UPT restrictions and do what we can to encourage the developers to drop the UPT limit on Civ 6, in the interests of providing us with more competent AIs. Evidence shows that the Civ 6 AIs are already playing as though there isn't a UPT limit, what with spamming a lot of the same types of units which have nowhere to go and their inability to build up enough courage to attack with singly-stacked units, while removing the UPT limit could address both of those issues.
The idea of combining units into armies (as seen in Civ 3) just creates another area in which an AI will be incompetent and easily taken advantage of by a player.
But, the theory behind the idea is a good one. How do we find a way to implement it in a way that AIs can be expected to realistically leverage, without human players complaining that the AIs just keep building these massive combined armies that are unstoppable?
If you want some sort of a limit, you need to make it an intrinsic one, based on the situation, without limiting the quantity of units on a given tile. For example, you could say that if there are 5 units or more on a given tile outside of a city, then each unit will defend at a 20% reduced rate, due to some sort of overcrowding effect. On attack, if you have 5 or more units on the same tile or 5 or more units on all adjacent tiles for 2-movement-units (to avoid the trick of just moving 2-movement-point units one tile away from the stack before attacking), then your attacker gets a 20% reduced effectiveness. Or, make the penalty limit "3 units of the same type," which will encourage unit diversity, while not limiting you from stacking a large amount of units if they are the only type of unit that you are able to build.
In other words, a penalty gets imposed for "excessive" stacking (as per however the game chooses to define "excessive"), but stacking limits do not get in the way of unit logistics in areas that aren't on the front lines of a battle, and you can still field your stacks of doom, but they become less effective when sent around together as a large rabble. Even have successive penalties, such as an increased penalty for every level of 5 units (10 units would have a 40% penalty, given the example numbers that I used).
In Civ 4, one of the best Traits that AIs are able to leverage is the Financial Trait, which provides an intrinsic extra Commerce to any tile with 2 or more base Commerce already on it. AIs tend to do much better with intrinsic-types of benefits and penalties, as it tends to be easier to program an AI to leverage them properly.
AIs can and do field mixed-unit types of armies, probably often better than human players, so I would see both strategic and tactical value in a unit-type-limit penalty (i.e. 3 Pikemen, 3 Macemen, and 3 Swordsmen can all stand on a given tile without receiving penalties due to not exceeding a 3-per-type value, while a 4th one of any given type would fight at reduced odds, but the other 3 of that type would fight at their normal odds if the 4th one got killed off). That said, it would be good to also have a total-number-of-units penalty, to avoid a human player "gaming" the system by keeping a ton of obsolete units around just to be able to get around the unit type penalty while having a larger stack.
So, if you have 2 Pikemen and 4 Macemen on a tile, with a penalty for unit types at 3 and a penalty for total units at 5, then if you attacked with one Maceman, it would have both penalties applied to it. If your Maceman died and you attacked again with another Maceman, it would only have the total unit penalty applied to it, as there would be 2 Pikemen and 3 Macemen (a total of 5 units) remaining on the tile.
It's not a perfect model, but what I find in Civ 4 with mixed-army attacking against mixed-army defenders is that when you have MOSTLY one type of unit in your stack and you mostly attack with that unit type, you end up getting around the benefits provided by mixed-army defenders in a way that an AI isn't all that great at coping with*. So, the unit-type penalty addresses this situation, meaning that if you rely too heavily on one type of unit, those units of yours will fight with less effectiveness. Meanwhile, if you balance your army, the nature of different defending units having different defensive bonuses will mean that more of the defenders that are good at dealing with a particular type of attacker will face off against the type of attacker that they are good with.
Implemented in this way, these penalties exist primarily for combat-purposes, without getting in the way of pathing logistics.
* An example from Civ 4 Beyond the Sword:
Spoiler :
You are facing off against 2 defending Archers, 2 defending Spearmen, and 2 defending Axemen.
Your attacking Axemen do very well against defending Spearmen, but less so against defending Axemen and defending Archers. If you attack with Axemen, they'll face off against the Axemen and Archer defenders. But, if you attack with enough Axemen, eventually you'll be facing off against the defending Spearmen, which you'll easily kill.
Your attacking Chariots do very well against defending Axemen, but less so against defending Spearmen and defending Archers. Similarly, if you attack with Chariots, they'll face off against the Spearmen and Archer defenders. But, if you attack with enough Chariots, eventually you'll be facing off against the defending Axemen, while you'll easily kill.
Your attacking Swordsmen do very well against defending Spearmen and do all right against defending Archers, but do poorly against defending Axemen. Similarly, your first Swordsmen attackers will face off against the tough Axemen, then later against somewhat-tough Archers, and then against the easier-to-kill Spearmen, if you attack with a large enough number of Swordsmen.
But, if you attack with 3 units of each type of Axemen, Chariots, and Swordsmen, instead of 9 units of a single type (assuming that you just have 1 stack defender of another type), you'll have much worse odds than if you were to use 9 units of a single type, due to the defenders eventually "running out" of the "most appropriate type" of non-wounded defenders.
Your attacking Axemen do very well against defending Spearmen, but less so against defending Axemen and defending Archers. If you attack with Axemen, they'll face off against the Axemen and Archer defenders. But, if you attack with enough Axemen, eventually you'll be facing off against the defending Spearmen, which you'll easily kill.
Your attacking Chariots do very well against defending Axemen, but less so against defending Spearmen and defending Archers. Similarly, if you attack with Chariots, they'll face off against the Spearmen and Archer defenders. But, if you attack with enough Chariots, eventually you'll be facing off against the defending Axemen, while you'll easily kill.
Your attacking Swordsmen do very well against defending Spearmen and do all right against defending Archers, but do poorly against defending Axemen. Similarly, your first Swordsmen attackers will face off against the tough Axemen, then later against somewhat-tough Archers, and then against the easier-to-kill Spearmen, if you attack with a large enough number of Swordsmen.
But, if you attack with 3 units of each type of Axemen, Chariots, and Swordsmen, instead of 9 units of a single type (assuming that you just have 1 stack defender of another type), you'll have much worse odds than if you were to use 9 units of a single type, due to the defenders eventually "running out" of the "most appropriate type" of non-wounded defenders.
You could also have the same types of penalties apply in other related ways. For example, when healing from combat damage, the same two types of penalties could apply--there are only so many mace-fixing specialists in your empire, or perhaps there is only enough space in the grassland for some of them to effectively do their jobs at the same time as each other, so they operate at a reduced efficiency (i.e. your units heal more slowly) when all of the Macemen are healing on the same tile as each other.
Last edited: