1 unit per hex. Poll

[Jolly Rogerer]

Wow...I figured "For" would be winning, but not at the rate of 5:1! That's good to see. I would be super bummed if 1UPT isn't the norm going forward.

That's a reflection of the number of people who have given up on Civ5 and moved on. 1upt still sucks for a game of this scale, being more appropriate for a RPG than a grand strategy game which measures turns in years and tiles in square miles. Even for those who don't care about scale or realism at all, it still is unworkable for the AI, and every game which uses it with anything like these sorts of densities and movement rates will suck no matter how powerful processors become.

 

[Toulouse]

I upt is one of the worst aspects of Civ V. It doesn't work; it isn't fun; and it isn't realistic. Limited stacking is the way forward.

 

[eric_]

That's a reflection of the number of people who have given up on Civ5 and moved on.

Does that also explain why there are so many posts decrying every and anything about ciV?

It's hilarious to look at the poll results and then read the posts.

 

[redmosquito]

I voted against, but only because the AI simply cannot handle it in any meaningful way whatsoever. In a perfect world I'd be for 1 unit/hex...

if only the designers could understand that

 

[Ahriman]

I upt is one of the worst aspects of Civ V. It doesn't work; it isn't fun; and it isn't realistic. Limited stacking is the way forward.

I voted against, but only because the AI simply cannot handle it in any meaningful way whatsoever.

Why do people somehow think it would be easier to design an AI that could intelligently deal with limited stacking than with a 1upt version?
Limited stacking is even more complex, because of the need to consider synergies between multiple units on the same tile.

 

[Jolly Rogerer]

Why do people somehow think it would be easier to design an AI that could intelligently deal with limited stacking than with a 1upt version?
Limited stacking is even more complex, because of the need to consider synergies between multiple units on the same tile.

That's actually a lot simpler for the CPU than the pathfinding problems 1upt introduces. The pathfinding is a pain for a human, but an abomination for the CPU. The reason is that for movement there is no shortcut possible, it's brute force all the way. Imagine trying to figure all the possible moves every turn, including the random elements like combat outcomes. With 1upt not only does every unit need to be figured separately, the order of the movements is critical as well. That's a lot of factors to figure and rank.

Synergies in unit stacking on the other hand are a relatively simple mathematical ranking (though dependent upon movement as well). That's why the SoD was capable enough to give you a challenge in Civ4, at least for a while.

 

[Beef Hammer]

I upt is one of the worst aspects of Civ V. It doesn't work; it isn't fun; and it isn't realistic. Limited stacking is the way forward.

You're clearly in the minority

 

[SJN]

When I heard about the 1 unit per tile I was so excited.

After just a few games, I'm quite disappointed. Here are my reasons.

(1) I do not enjoy the transport of units over water.
(2) I cannot give "group orders" (honestly don't know how you would implement, but it's really painful to try and move a large army... this is related to #1)
(3) It doesn't scale to the right sizes on the map. I played Shaffer's Earth Map (spelling?) and I had a ballista shooting across the Aegean sea. That kind of ruined it for me.
(4) I can't have "bases" where I keep my army. I hate having to have my army just sitting all over the place, rather in a single place I can look to when a war breaks out.

If I had any suggestions for the developers, it would be to consider movement/storage as a different thing than actual combat. Maybe an easy way to deal with this would be to allow for unlimited stacking with the following limitations:

(1) For each unit within a square, combat effectiveness decreases by some factor (maybe even 50% each time).
(2) Maybe going back to a Civ-I style penalty, you could have multiple units destroyed in a single attack.

The key here would be, you can stack as many units as you want when you aren't fighting. You can move them as a group, and you can store them (at a base) as a group, but don't think about fighting that way. This also adds an interesting tactical approach to the game because you can try a "surprise attack" on a country and try to destroy as many units "in storage" as you can before they can get on tactical footing. Also, during a war, you could have missions to try and destroy units in-transit if possible.

And please, before you just rip apart my ideas, I recognize that what I'm proposing has some big holes in it. But I think it would be an interesting starting point to experiment with.

-- SJN

 

[Jolly Rogerer]

You're clearly in the minority

He'll just have to console himself with the fact that he's right.

 

[MadRat]

He'll just have to console himself with the fact that he's right.

ouch horrid pun.

Rat

 

[Nares]

I voted for, because it's awesome and is the upgrade to combat that the Civilization franchise needed.

Obviously it still needs work.

 

[Sonereal]

Holy crap! 80% of the people who voted are for the idea of 1UPT.

Big. Freakin'. Shock.

It isn't as if this poll didn't come up thirty times BEFORE CiV came out. I'm so glad we're continuing the tradition of posting redundant polls and threads.

It just means people mostly want the Tactical AI to improve, not regress to stacks.

 

[Ahriman]

That's actually a lot simpler for the CPU than the pathfinding problems 1upt introduces. The pathfinding is a pain for a human, but an abomination for the CPU.

The AI problems with 1upt aren't efficient path-finding, they're figuring out when to move which unit where.

If you start making the net stack effectiveness of moving a unit to a tile a tile depend on the other units also on that tile, the problem becomes even *more* complicated, not less.

 

[CHEESE!]

i suspect it'll be the same as all of the other polls, roughly split.

20/20.

 

[Jolly Rogerer]

The AI problems with 1upt aren't efficient path-finding, they're figuring out when to move which unit where.

If you start making the net stack effectiveness of moving a unit to a tile a tile depend on the other units also on that tile, the problem becomes even *more* complicated, not less.

No, the larger problem is the pathfinding, just as it is for you when you have to figure out exactly which order to move your units so they end up more or less where you wanted them in that tight spot at the front or that bottleneck as you move through it with an army. The problem is considerably more difficult for the CPU which doesn't have your built in spacial reasoning capabilities. Deciding what unit you want in a particular spot is hard for a human, but the CPU has an easier time of it by using heuristic weighting rather than trying to calculate and weigh all the possibilities the next turn could bring. Chess programs do this too, and Civ is massively more complicated than chess.

You gain more by allowing units to pass through one another than you lose by trying to optimize your stack's end state. Stack end state values are (again) simple mathematical comparisons, but the CPU has to do an immense amount of extra pathfinding when its units can't pass through one another (multiple paths per unit times multiple units times multiple orders of movements). So yes, stacking adds more complexity, but it removes a lot more. There's a reason why there weren't stacking limits in the Civ series previously aside from the fact that it makes perfect sense at this scale.

 

[Isikien]

For
Much better than the 'human siege tower' approach of the previous Civ

 

[redmosquito]

No, the larger problem is the pathfinding, just as it is for you when you have to figure out exactly which order to move your units so they end up more or less where you wanted them in that tight spot at the front or that bottleneck as you move through it with an army. The problem is considerably more difficult for the CPU which doesn't have your built in spacial reasoning capabilities. Deciding what unit you want in a particular spot is hard for a human, but the CPU has an easier time of it by using heuristic weighting rather than trying to calculate and weigh all the possibilities the next turn could bring. Chess programs do this too, and Civ is massively more complicated than chess.

I think the other thing is that the idiosyncratic irrationalities are to some degree aggregated out when you have a stack, whereas in 1UPT this is not the case. (i.e a composite random variable has a lower variance compared to the sum of variances of each random variable)

 

[jprc]

voted: For.

It needs now to be improved by allowing some well defined exceptions (2, 3 units on 1 hex, depending on the type of unit, civil or military).

 

[lschnarch]

voted: For.

It needs now to be improved by allowing some well defined exceptions (2, 3 units on 1 hex, depending on the type of unit, civil or military).

That is, what I will never get.


The whole thread it filled with statements like "I LOVE 1upt. If it only weren't 1upt, but 2/3/xupt".

Guys, I have every kind of understanding for you to disgust an unlimited SoD. I do so, too.

But please, we are asked about 1upt.

If you think that there should be exceptions, you are just not *for* 1upt. Is this really so hard to get?

 

[jprc]

Is this really so hard to get?

What is not hard for me is to know that the stack of doom we had in Civ 4 was killing some fun.
Now, if you propose: "do you want bread or butter?" I think it makes sense to say sometimes that you would prefer butter on bread.
free speech.