Civilization 5 Announced

[Krill]

Well, it didn't slip by unnoticed. The argument then went to why implement 1UPT if a balanced AI couldn't be coded for it, then Sirian said game balance wasn't a particularly important concern of the devs, selling the game to the largest amount of people possible was.

I don't really know how you answer that, short of watching people stop buying C5 because of the word of mouth effect saying it is a waste of disk space.

[Sirian]

Interestingly, there's a Start Bias for some of the civilizations. America is coded to appear near rivers, England/Ottomans along the ocean. Other civs are associated with or against certain terrains: Arabs are coded to start near desert, Aztecs in jungle, Iroquois in forests, India in grassland, and Russia near tundra. Egypt avoids forest/jungle, Siam avoids forest, and Songhai avoids tundra. I'm not sure how I feel about this - realism limiting the gameplay perhaps? Seems like a strike against Arabia if they're usually going to be in the desert. Sirian, if this was your work, could you explain the thinking?

For Civ4, I intermixed the terrain more than any previous Civ. This improved the variety and balance of tiles at almost every city, but it also made cities more like one another and less challenging to manage. This might be described as improving balance at the micro level but losing variety at the macro level. Every city with a smattering of grass and plains, hills, forest, etc.

A big part of Jon's vision was to undo that. He wanted more homogenous terrain at individual cities, with large, sweeping terrain features at the macro level. This was both for "look and feel" and for gameplay purposes. I was not quickly won over, fearing what this much terrain change would do to the game balance, but it was my job to support his vision. We would redo resources from the ground up and use those to bring terrain back in to balance, including reversing the order of start placement and resource distribution.

Changing terrain in to larger patches came first and fairly easily, although what he wanted for mountains was quite tricky. (A forest spanning a sixth of the Old World on a huge Terra map is something to behold, I must say!) Civ4 could never have made the resources-as-balance work out, but Jon came up with the idea of going back to Civ3-style "bonus" resources, which would affect only tile yields, and that fixed the problem. We could now put as many of these bonus resources on to the map as needed, where needed, to bring competitiveness to the handicapped base terrains.

Distance between start points is another tool for balance. Civs starting in a fertile region get less of their own space. Their neighbors will be closer, on all sides, where a civ starting in a less fertile area will get more space. This lets the rough-terrain civs go for "more, shorter" cities, while the crowding factor for the fertile starts tends to support "fewer, taller" cities. The happiness system, being global, has a mix of fixed sources of happiness (luxuries, etc) and per-city sources (mostly buildings, but some policies or wonders, etc), with at least the aim of maintaining balance between the two expansion approaches. The one thing that is supposed to be difficult if not impossible is to rapidly obtain both number of cities and high pop count per city. "Many" and "Taller" are supposed to be magnetic opposites, repelling one another.

The terrain bias concept resides on the fact that -- now -- there actually is diversity of terrain at the macro level. Yes, almost every city in the huge forest will be like one another, but the forest civ will be unlike the desert civ who will be unlike the jungle civ. This tableau is all new: it is unlike any previous civ, and there is far more macro balance between zones than ever before. Using resources and other tools to bring a reasonable balance to this new type of diversity was a pleasure to undertake.

The civs with biases generally have terrain-based Unique Abilities. So the biases help to bring more value to these abilities by ensuring a reasonably decent match: you will get to use your Special Abilities as these civs, instead of them being weaker civs on average because of the diceroll of whether your ability will even matter. The start bias ensures at least some application of these abilities, wherever possible. There is also something to be said for the flavor of playing a civ tied to certain terrain types.

We made sure to enable an easy way to shut this off. But I think you will find that the civs with terrain-based special abilities are, in fact, made weaker (compared to the rest) without the biases.

There is one other balance factor to mention: location of military resources. You will find plenty of horses in the fertile, open lands, but not terribly much else. Nearly all of the iron, oil, aluminum, uranium is off in more rugged locations. And those food-rich regions will tend to be hammer poor. Forests and hills are now of equal value without regard to base terrain, so those desert hills are just as good as a grassland hill. Neither desert, tundra, nor jungle are as inhospitable in Civ5 as they were previously. There is years upon years of interesting map goodness to explore now, even just counting the core map scripts and the sheer variety of terrain combinations that can appear. The biases are part of bringing more value to the interaction between civs and the terrain.

With my help, Jon has been able to bring his vision about the maps and the large, sweeping regions of terrain, to fruition. He was right about the potentials that existed if we moved boldly in that direction.


- Sirian

[HouHou]

Well, it didn't slip by unnoticed. The argument then went to why implement 1UPT if a balanced AI couldn't be coded for it, then Sirian said game balance wasn't a particularly important concern of the devs, selling the game to the largest amount of people possible was.

C'mon, the Civ 4 AI , especially vanilla, being held up as some type of AI paradigm? Civ 4 was so good that it allows the players to, among other things:

* Have 100% sneak-attack immunity. It also allowed players to completely break Tokugawa's intended AI behavior with 100 hammers.

* Delay war for 10s of turns by begging one gold from an AI.

* Sell a useless tech to an AI, then declare war the same turn. Thankfully this was fixed in BTS, and then it got unfixed .

* Constant war being the default go-to strategy in Vanilla. It is unacceptable to have a game with so many ways to rush production with the end result being that the AI would only have 2 archers in a border city.


Just adding some balance into the discussion

[Gold Ergo Sum]

I wasn't around for Day 1 on Civ IV, but all of those things were buried pretty deeply in the gameplay I imagine. None of those exploits were probably widespread in the first month, let alone the first day or two. And some of them hardly count as an exploit at all.

Those aren't exactly "core" mechanics. Those are pretty small, "gamey" issues.

[Sirian]

Well, it didn't slip by unnoticed. The argument then went to why implement 1UPT if a balanced AI couldn't be coded for it, then Sirian said game balance wasn't a particularly important concern of the devs, selling the game to the largest amount of people possible was.

In Civ4, balance was given the top shelf. It was the foremost concern, behind only simplicity.

Civ5 still cares about balance, but for instance, balance for World Wonders means a lot of fairly mild effects. There is virtually no "Wow Factor" to Civ4 wonders of the world, because so much effort was put in to making wonders balance with the rest of the game. A desire to bring back any "Wow Factor" to world wonders means, of necessity, loosening the reins on balance. This is one example of where Civ5 decided to reduce the priority on balance. This is not the same as balance becoming unvalued.

There are areas where Civ5 works much harder than Civ4 to try to be balanced. Roads, gold, and unit promotions come to mind. With gold now able to buy almost anything, the amount of effort put in to preventing it from busting open and snowballing away has been gargantuan. You can see some of the fallout from this in the slow pace of the early game economic buildup, the high maintenance costs on units and buildings, and so on.

Perhaps the biggest fault of Civ5 is not being balanced about its approach to balance. Some areas, there are purposely unbalanced options, trying to breathe life in to areas that were deemed too dull in Civ4; other areas, the reins are on pretty tight, trying to keep control over new features that are more wide open than in the past.

I simply haven't played enough to see where it all ends up going, so as yet, I have no personal opinion about the release build. Civ5 is a new game, intended to be different from Civ4, Civ3, etc, but still to play and feel like Civ, overall.


- Sirian

[HouHou]

I wasn't around for Day 1 on Civ IV, but all of those things were buried pretty deeply in the gameplay I imagine. None of those exploits were probably widespread in the first month, let alone the first day or two. And some of them hardly count as an exploit at all.

Exploits are exploits; would it matter if a car's brake defects only be revealed 6 years after the fact? Some more examples:

* In border wars, 100 culture is paramount, because every time a city's cultural borders expands, it adds a hidden culture bonus to the ring before it. 10 culture adds hidden culture to your first ring, while 100 culture adds hidden culture to your second ring.

* Barbs (not barb cities) CANNOT spawn within 2 tiles of a unit. Picket your coast and barb galleys are gone!

* Civs know when you are building a wonder, and thus won't prioritize it unless they are already building it. That's pretty nice for Oracle eh?

* We Fear You Are Becoming Too Advanced (WFYABTA) took forever to be understood (first off, what does that even mean in plain English?) In addition, certain Civ's WFYABTA threshold changes with their attitude. Where was this documented? You get NO WFYABTA if you are in the bottom half of the scoreboard and trade with a Civ in the bottom half of the scoreboard? Once again, documentation?

And no they are not just "small, gamey" issues. 10 turns in Civ 4 is an eternity; you can whip a city-buster stack in 10 turns. So yes Civ 4, thanks for giving me the opportunity to buy cheap peace. You might not declare on me @ pleased, while I can declare on you at Friendly.

And you say you lost a cultural victory because your friend sneak-attacked you. Well..if you played your cards right, that would be categorically IMPOSSIBLE in Civ4.

Read some of the civfanatics Succession games. People have done such things as:

* Civ 4 Vanilla with no civics, no cottages
* Civ 4 BTS: Win on against teams of 2 Deity AIs (this one is good, they used every trick in the book).
* Warlords w/ BTS (probably toughest "release" combo), Liberalism in 110 AD, followed by ~20 Cavalry and 6 Trebs in 305 AD!

[Txurce]

Sirian's comment about how difficult 1UPT tactics are to code for an AI kind of slipped by unnoticed in this thread, but rings dead on to me. Think about chess, a similar game of front-based 1UPT tactics, but with a smaller gameboard, homogeneous unit behavior and abilities (no ranged units or promotions), and a very tightly defined objective and win condition. Chess took about four decades of research by some of the smartest programmers on the planet to match the best human players.

That makes all the sense in the world. On the other hand, someone else mentioned Panzer General coding experience as a potentially great shortcut. Is there a reason why approaching it from a similar perspective wouldn't work?

[haphazard1]

I was having this exact problem & found this solution suggested: http://forums.civfanatics.com/showpost.p...stcount=11 It seems to have worked for me (only one try admitedly). Note that you still don't get a default save game name, and you can't save till you give it a name (obvious, but it makes it look like it's still broken when you look at the save game screen).

Thanks, pling! This worked for me, although needing to go into the INI file to get saves working...ouch. Betas, betas, anyone got a beta tester?

Exploits are exploits; would it matter if a car's brake defects only be revealed 6 years after the fact? Some more examples:

* Barbs (not barb cities) CANNOT spawn within 2 tiles of a unit. Picket your coast and barb galleys are gone!

We apparently have totally different definitions of "exploit". Fogbusting is a basic game mechanic, not an exploit. Galleys are non-ocean-going ships, so presumably such barbs are really operating from the land. Stationing sentries (patrolling the area, in effect) should eliminate spawning of galleys, even if the sentries do not actually take to the water themselves.

A semi-related question for those playing Civ V: I am finding it to be very common that a barb encampment on the coast has one or two barb triremes stationed offshore. Is anyone else seeing this? It seems to be more than chance, and definitely makes clearing that camp a lot more difficult (which is a good thing, IMHO).

[Yazilliclick]

That makes all the sense in the world. On the other hand, someone else mentioned Panzer General coding experience as a potentially great shortcut. Is there a reason why approaching it from a similar perspective wouldn't work?

I think that person is remember that through rose coloured glasses. AI in those games was never all that great. It's true they had an AI on a hex map but it was not spectacular. Besides that it had to deal with a heck of a lot less than the Civ5 AI does.

[pocketbeetle]

On the other hand, someone else mentioned Panzer General coding experience as a potentially great shortcut. Is there a reason why approaching it from a similar perspective wouldn't work?

It was Bullstrode's post a couple of pages back, and personally I agree with him.

The game is 10-15 yrs old and features a hex grid, 1upt and ranged combat.
These are not 'new' innovations that Firaxis has had to boldly break ground to overcome.