Sid Meier describes the allure of Civilization, his groundbreaking 1991 game:
We really didn’t design it in, but as I look back, I realize there is a really interesting growth path in the game. In the beginning, you have one or two military units, just a couple of technologies, and just a couple choices to make. The game opens up and unfolds gradually at your own pace. And before you know it, you’re dealing with lots of interesting decisions.
There is also that one-more-turn quality. There are enough different things going on that there is never really a good time to stop. In one city, you’re building something, and when that is done, you’re exploring this other continent. And then you meet the leader of another civilization, and you’re wondering how that is going to turn out. There are enough different threads in your imagination at any one time. One of the reasons that Civ has become this well-known phenomenon is that people remember the night when they stayed up to 3 a.m. playing it. It’s these experiences that stick with you.
So true.
The game was originally going to be about the rise and fall of civilizations:
There would be occasional setbacks, such as the Dark Ages, that you would have to overcome, and the glory of overcoming them would be satisfying. But what we found was that when bad things happen, people would just reload the game. They were not interested in the fall of civilizations. Just the rise of them.
So we ended up with a game of constant progress. We actually started to understand the psychology of gamers. When something bad happens, often they blame it on the computer, or the designer, or some other outside force. They would think it wasn’t fair, and they would reload the game.
We also found the same phenomenon when nuclear weapons came into play in the game. Players did not have much hesitation in using nuclear weapons against the AI-controlled civilizations. But if somehow the AI used a nuclear weapon against them, it would be: “wait a minute, that’s not fair.” The message of Civ is that [nuclear weapons are] a lose-lose for everybody. But we found that we couldn’t allow the AI to use them, because it was destroying the player’s experience. If the player is destroying the AI’s experience, then it’s only the computer that suffers.
Stupid computer…
Constant seems to be very standard for any game which represents a long time (as opposed to e.g. games of a single battle or combat mission). It might be because people don’t like decline, but I don’t think Civ is strong evidence for that. The specific kinds of mechanisms that Civ prefers for decline — global warming, corruption unless you turn Communist — could be getting under players’ skin in other ways. And more generally the real-world dysfunction in recorded declines — like all the shenanigans in deposing Roman Emperors — seems to resist simulation even with modern computers, much less 1991 computers.
Maybe they did hear that phrase from the playtesters, but I think the main problem in that area was a little different. IIRC the early Civ AI was extremely annoyingly aggressive against the player nation (compared to aggression between NPC nations). Mixing that with nuclear weapons naturally promotes the annoyance to a new level. I replayed Civ III recently and was reminded that (at least at the Monarch difficulty) its AI is less consistently anti-player: NPC nations are aggressive against the player, but also aggressive against each other, which would be somewhat less annoying with nukes. But the AI is still annoyingly prone to starting horribly destructive war with or without nukes, probably unrealistically so. No one really understands the game theory of nukes, but the rate at which Civ AI started nuclear wars is hard to reconcile with our incomplete understanding (and with fifty years of history); some might say “that’s not fair” but “that’s ridiculous” is another natural reaction.
Ah, generalizations…
And this right here is the difference between the Wikipedia Fantasy crowd and the Dwarf Fortress crowd. :]
When wargaming started it was a Prussian exercise meant to teach Prussian soldiers how to think. It did not claim to be fun or entertaining. In the intervening years gaming has mutated in many directions. One such mutation was Civ.
I usually played Civ as a conquer-the-world game, where I tried to get some kind of unfair advantage so that the enemy factions did not pose a challenge. Probably Sid Meier would say I missed the point; probably Sid Meier would have wanted me to respect the other factions, allow them to flourish, and still beat them with an Alpha Centauri victory.
A well-designed game with a lesson to teach should have the courtesy to tell the customers, “Look, we are going to take 40 hours of your life to teach you OUR lesson, and OUR lesson may not be fun.” I may respect such games as learning experiences, but not as entertainment.
Since the days of stand-up arcade consoles offering “Battlezone” and “Missile Command,” MOST games were sold to gamers as fun consequence-free fantasy violence. If a Civ-style game is meant to teach me not to use nukes, it should not sell itself as violence-for-entertainment.
Too many game designers make something that confirms their biases and never stop to question their assumptions. They are one side of the problem. Many gamers think they are strategic and tactical geniuses because they have enough Adderall-fueled twitchiness to beat a game without cheating to get extra ammo. They are another side of the problem. And then there are the gamers who won’t even *start* to play the game unless they can load up Cheat Engine or a similar “trainer” program to give themselves unlimited ammo. When they become game designers, they tend to produce games like the Bethesda Fallout games, where even the guys who code the game make sure the console codes can give them unlimited ammo so they can figure out what they want the lesson to be.
It’s so much worse than this.
Too many game designers make something that they THINK will confirm their biases.
Then never bother to analyze what actually goes on and why, until it turns out the result either does not work this way, or is broken outside their pet scenario. Nor even test extensively by people unfamiliar with their notions of how it “ought” to work.
The best will learn, the worst will complain.
Most designers are simply not good enough at understanding of either logic and math of the model they use, or their own vague notions or to project the latter onto parameters of the former without much blind wandering.
Hence all those releases that should have been beta versions (half-baked code is one thing, but half-baked data is about as common) and interminable seesaw of “nerf”-“buff”.