Part 10: Mechanics: The role of a good AI in game design

I think what you're talking about is a real problem with the game; however it's one that's entirely separate from the combat system. The AI in this game just sucks. Even a completely new human player with no grasp of the mechanics beyond that high Attack units do well attacking and high Defense units do well defending can at least stalemate the AI and will probably make them suffer much higher casualties. And diplomatically smart players can run circles around the AI and make them give away free money or declare war or peace almost at will or trade great techs for junk.

The AI builds terrible unit types, makes terrible city improvements (no Harbors EVER), never founds enough cities, is treacherous at stupid times, etc. It's not a smart AI at all. And it plays like an idiot at every aspect of the game, not just using the combat rules.

But does this actually matter? Well quite possibly not. Let me play devil's advocate for a bit here.

At the end of the day, a good AI is just a means to an end: creating a challenge for the player. Plenty of challenging and successful games don't have an AI at all. And no Fire Emblem game has ever had a good AI. Heck, the Fire Emblem AI understands the extremely simple Fire Emblem rules LESS well than the Civ 2 AI understands this game's complicated rules. Fire Emblem AI enemies will suicidally attack people they have a 0% chance to hit without a second thought. And their whole battle plan is just a disorganized charge- or even for some people to sit still and never make a move at all. But Fire Emblem games can still be good strategy games despite a pitiful AI. The AI only needs to be good enough to order mindless attacks in that game because the developers gave them such advantages in positioning and numbers and often equipment and stats that you must play very cleverly to win.

Similarly, the Civ 2 AI cheats HARD, especially on Deity. Sure they have like 3 cities and those cities are small, but those cities nonetheless turn out an endless flood of units and buildings and wonders and can nearly keep pace with you in Science. And they're immune to most forms of unhappiness, don't have to deal with a senate as a Democracy, have options you don't have in diplomacy, don't suffer decreased reputation, never lose Triremes at sea, flagrantly break the rules by attacking directly off boats, know things they can't possibly know without embassies, flagrantly break the rules by sneak-attacking on the same turn they break an alliance even though that's 100% impossible for a human, etc. Plus they're all semi-secretly on a team against you and will all openly band together- even if they haven't actually met you- to form one giant alliance and smash you when you start doing too well.

So the Civ 2 AI can present a serious challenge which can only be overcome by good tactics and strategy despite playing stupidly, same as the Fire Emblem one. You beat the Fire Emblem AI by standing on forests and switching to a weapon with triangle advantage to drop their hit chance to 10%. You beat the Civ AI by standing on Hills and switching to fortified mode to drop their chance of winning a battle to 10%. In both cases the AI will fall for it. In both cases they're too dumb to try anything similar themselves. And the result is similarly a massacre.

In fact, from a game design perspective it might actually be GOOD to have a dumb AI in this kind of game and rely on unfair circumstances to present a challenge. The whole point of a game is for a human to have fun playing it and the entire idea of a strategy game in particular is for the player to have fun because they're thinking their way to victory. If your game puts the player in a situation where they would be guaranteed to lose to an opponent of equal intelligence, then you make the player feel smart when they win anyway. Even if they're only average or something for a human. If any human can out-think your AI but the AI has enough of an advantage (ideally scalable with difficulty modes) that they present a challenge anyway, then you've created a game that any human can probably enjoy on at least one difficulty mode rather than one which only a few very smart people will ever feel good at.

However, I don't entirely agree with the above argument. There are some good points to it and I think it's more or less spot-on with regard to Fire Emblem. But there are at least two important distinctions between the dumb AI with massive resource advantage situation in Fire Emblem and the dumb AI with massive resource advantage situation in Civilization 2.

First of all, in both cases the AI 'cheats' hard enough that they play a fundamentally different game than the player. In Fire Emblem the player can't receive mid-battle reinforcements out of stairwells. The player has 0 expendable units. The AI consequently only needs to kill a single player unit to force a restart and thus win, so the AI is designed to do that instead of to actually defeat your army. The AI units often have absurdly high stats or do things like break the weapon stat caps with custom-forged uber gear in Awakening. And of course they don't have some unit limit of only 8 people per battle or whatever. Nor do they need to manage funds and resources over a long campaign; they always have unlimited gear for free.

In Civ 2 the AI just gets massive amounts of free production and Science and gold, doesn't need to deal with unhappiness (usually), doesn't need to deal with a senate or an international reputation, can make all kinds of illegal moves, etc.

In Fire Emblem it's not a problem that the AI doesn't play the same game as the human because there's not actually any kind of story-equivalence between them. Most Fire Emblem games are about a tiny, poor. ragtag group of mercenaries, refugees, disenfranchised lordlings, children, and an occasional genuine badass going up against a vast and powerful and often outright inhuman empire. Of course the enemy has 10x as many guys, of course they have unlimited gear, of course they get reinforcements, of course they don't care about casualties the way you do, etc. The asymmetry is appropriate and actually even kind of satisfying. The good Fire Emblem games don't just make you feel proud of your strategic prowess for overcoming a materially superior opponent, they also make you feel sympathetic toward these brave underdogs you're always using.

Civ 2 is different. The whole premise of Civilization is that you and the other leaders are fundamentally the same. I'm not going to win because the Romans are a small band of heroes and the Sioux are a giant, evil empire; I'm going to win because I made different choices than Sitting Bull. If I have fewer units, it should be because I chose to build fewer units. If they're richer than me, it should be because their economy is actually bigger or they chose to devote more of it to building their treasury rather than because the Romans are narratively supposed to be poor and ill-equipped. In Civ 2 I am supposed to be playing the same game as the AI leaders, but I am not. In Fire Emblem, the AI's advantages and cheats fit the narrative and premise. In Civ 2 they undermine it. And this is a problem.


The second problem is one of HOW the game creates its challenge without a good AI. Not all things that make a game challenging are equally fun to deal with. A well-designed game should present a challenge which is actually fun to overcome. This is an area in which some Fire Emblem games fail too. For example, Fog of War chapters are dull and obnoxious luck-fests the first time you play them because there's no way to actually plan a strategy when you don't know what's out in the fog. And Fire Emblem: Binding Blade sucked from beginning to end because the game's 'difficulty' came entirely from vast hordes of spawn-moving reinforcements which poofed out of thin air to force what amounted to guess-and-check restarts. Good Fire Emblem games don't rely on stupid things like that to create their challenge.

Civ 2 similarly has too many guess-and-check and luck-based elements. The AI always knows the exact type and number of units present in all of your cities, but there's no way to know what's in theirs without spending Diplomats or risking death in an attack. And they just know where stuff is in the Fog of War. Giant armies will show up at your arctic cities which have never been scouted by an enemy at any point, etc. Oh and barbarian groups can spawn-move from thin air at any time. That's EXACTLY like Binding Blade reinforcements. In fact, the way the enemy Triremes can cross oceans without sinking and the way non-Marine enemy units can cheat to attack off boats is similar too. An enemy army can suddenly appear from out of visual range in an impossible direction and sack a city that should have been safe instantly. And don't even get me started on goody huts. Who in the world thought it was a good idea to give them a chance of spawning a giant army of barbarians? If you get that in the early turns, it's game over. And speaking of starting stuff, there is a huge and permanent advantage if you have an extra starting Settler or a bunch of extra techs (ESPECIALLY if those techs lead to Monarchy or Republic). Or the game could dump you in an unwinnable starting situation like being trapped on the south pole with nothing but Glaciers around. Too much luck!

And it gets worse from there. See, the single biggest way that Civ 2 creates challenge on the high-difficulties is not by giving bonuses to the AI but by giving massive penalties to the player. I am referring mainly to the catastrophic amount of unhappiness your citizens have on Deity difficulty. Deity is a fundamentally different game from Emperor because allowing your cities to grow AT ALL is instant disorder. And if you build more than 4 cities, even one of your STARTING citizens will be unhappy. Keeping your civilization functioning at even the most basic level requires constant micro-management of every single city you make. This is a gigantic, frustrating, and annoying slog of checking dozens of cities every single turn. Deity games are just not much fun because of this. And a game that isn't fun has failed completely. Like Fire Emblem: Binding Blade!


So in conclusion I think I agree with you that this game has a very weak AI and that's a problem with it. The game would be more fun if it relied somewhat more on the AI being reasonably clever and somewhat less on the AI being allowed to cheat to create challenge. And it would also be good if the difficulties and problems it imposes on the player were less luck-based and less time-consuming to deal with.

However, I disagree that any of this reflects badly on the combat system- or the multiplicative stacking of Defense bonuses in particular. That's a great idea for the reasons I explained before. This is a separate problem.




"Smallpox" and "Bigpox" are actually the typical terms among Civ 2 people and especially Freeciv people as far as I've seen. "ICS" caught on as terminology mainly with the later games (especially Alpha Centauri it seems). "Celebration" is indeed just my word for the clunky "We Love the X Day" that more people use, but I'm sticking to it.

Also, I regard both pure Smallpox and pure Bigpox as badly inefficient strategies but they're not equal. Smallpox is better. And the ideal hybrid strategy for most games is much closer to Smallpox than Bigpox.




Spaceship wins would look really boring to you guys. Actually, they're really boring to play too. I mean, a spaceship win essentially consists of a standard rush to maximum science and then just sitting down and pulling back all military units and ordering all Engineers to sentry so I don't need to bother with them and can just skip 50 turns in a row while the parts build in my cities. The ideal spaceship win strat will result in no military action going on and nothing interesting to do for centuries as I tech up and then build the parts.




It's obviously true that getting, say, 1000 beakers this turn is better than getting 100 beakers for 10 turns starting with this one. But that's not what happens here. What happens here is a question of 100 beakers for 10 turns starting with this one being obviously better than getting 1000 beakers 10 turns from now.

As I pointed out, all of the other Science wonders are available BEFORE Darwin's Voyage. And most of them cost fewer shields to build too. They start paying dividends long before Darwin's Voyage can even exist. And since earlier techs cost fewer beakers, they actually give you even MORE techs than my analysis showed.

Another point which I didn't bother making because I was pointing out the relative badness of Darwin's Voyage instead of its absolute badness is the question of how many shields are actually worth spending to get a tech advance. Here's the thing: Darwin's Voyage unlocks with Railroad. Pre-Railroad, Production shields are in extremely limited supply. In the endgame, massive Production is available, but this is because of Railroads built everywhere, Factories, Offshore Platforms, Power Plants, and eventually Manufacturing Plants. You can't HAVE any of those things when Darwin's Voyage unlocks. So you don't have good Production available. So building it then is a big investment of a rare resource for a small payoff in an abundant resource.

Let me give a concrete example from my recent Emperor difficulty spaceship win. In the year 1630 I had not yet developed Railroads but was about to. I had put them off for a very long time because I didn't want to destroy my Hanging Gardens until I'd used it to help Celebrate all my cities. I had approximately 50 cities, 34 of them at size 12 and 12 between sizes 7 and 11. The others were tiny, recently founded villages. So celebrations were basically over and it was now time to learn Railroads. Average Production of one of these size 12 cities? 4 after unit upkeep. A few much higher, many as low as 1. There are many ways to generate massive Food pre-railroad. There are many ways to generate massive gold or Science pre-railroad. There is no way to generate massive Production pre-railroad. All you can do is pray for good terrain: like bunches of Hills with Coal or maybe Deserts with Oil or Mountains with Iron by a city which also has bunches of nice Grasslands so that it can grow big enough to actually work many of those good tiles. Oh and you can build the King Richard's Crusade wonder, which is really just not a great deal most of the time.

In short, Production is at a premium and useful for many things in that phase of the game (Like a bunch of actually good wonders that unlock around then or getting started on all the factories that unlock 1 tech later and will actually let you have abundant Production in the future). Science is abundant and it's really easy to be pulling down a tech every 2 turns- or faster. I sure was and I didn't even have Libraries.




I can't comment on Alpha Centauri, but Caravans are actually decently balanced in Civ 2. They completely break everything only under idealized circumstances at the very end of the game. Circumstances which require you to already have multiple absolutely massive cities that produce insane amounts of Trade already (and oodles of Food to get that massive) and to already have very late techs like Corporation and Refrigeration and Automobile and Advanced Flight and then have built a bunch of expensive improvements in each of these cities. And THEN build 3 Freights apiece. In other words, the game is already over and you already DID all the research and you already HAVE all the Trade you need. So the uber Caravans serve very little purpose. You could have easily won the game in the time it took to set them up.

During actual in-game circumstances your Trade routes are likely to produce something like 3-8 Trade each for much of history. Which is nice but not game-breaking. The number is much lower if you're not playing a Celebration strategy because only Celebration people are likely to have numerous high-Trade, high-size cities early on. To give real numbers again, in 1853 in that Rmperor game I've talked about a couple of times I had a bunch of Trade routes between my size 12 Trade-heavy cities. Most produced 3-4 Trade. I had a few giving 6-8.



This is the key. If you have fun playing that way, keep doing it. Nothing wrong with that. All I'm saying is that it is, as you said, not the most efficient. Of course, what I do isn't actually the MOST efficient either. The maximally efficient strategy is really boring.

For those of you out there who don't already have your own playstyle or have never played civ 2 and want to know why I would say stepping-stone Monarchy is not a good idea compared to either early victory pure Monarchy or pure Republic, here's why I disagree with many of these specific arguments:




1) Alright, hold up. You're trying to have your cake and eat it too. A unit can either scout OR do martial law. It cannot do both at once. Let me put some numbers on this and look at the very start of the game (when still a Despotism). Assume a typical, goodish city site: it's founded on a Grassland WITHOUT a shield on it and has a Grassland WITH a shield nearby. As I mentioned previously, founding on Grasslands without shields is a simple way to boost your Production forever for free. Anyway, at the start of the game this city has 2 surplus Food and will therefore spend 10 turns filling the Food gauge and then grow on on the 11th. It has 2 Production, so it builds the Warrior after 5 turns. Thus there are 6 turns between the Warrior being built and the city growing. This means you can scout out 3 squares and then walk 3 squares back. That's not much scouting since the city site already reveals a lot of the land around it, but it definitely helps a bit. So minor benefit there.

If you're a Monarchy, you actually get 3 Food per turn and therefore grow on turn 8. Which means you an only do 1 square of scouting before needing martial law. In other words, you can't scout at all.

So you MUST choose between scouting and martial law and cannot have both from the same guy. Now as a Despotism you COULD build a second Warrior who will finish just in time. The first one can then go off scouting while the second stays home. Ok. But this is a serious Production problem. Your Settler is now delayed 10 turns, which is a really major problem if, as you said, the goal is to do hybrid smallpox/largepox. You need to be founding cities earlier than that. Furthermore, once your city trains the Settler, it now is a size 1 city and has 3 units. Since you started as a Despotism, this means it eats a -2 Production upkeep penalty. Which means it produces NOTHING until it grows again or the Settler finds a site and builds there. Crippling.

Being a Monarchy at least solves that upkeep problem, but you're still suffering a massive delay to build multiple Warriors before finally starting your Settlers. Fortunately, you probably don't need many scouts anyway. So you don't need to do that with every city. Nonetheless, I think I've made my point that the same unit cannot both scout any useful distance and enforce martial law and that building 2 or more Warriors so you can do both is problematic.

Furthermore, Warriors may be relatively cheap but money is extremely tight. Your starting income as a Despotism/Monarchy with size 1-2 cities is about 1 gold per turn per city on average. Maybe less. Buying a Warrior that's already halfway finished is 8 gold. That's not cheap at all. That is 8 whole turns of income at the start, and quite possibly for some time thereafter. And there are so many better things to spend your money on or save it for.

And finally, Warriors are not a good defense. They aren't even an ok defense. A barbarian Archer has 4.5 Attack vs the fortified Warrior's 1.5 Defense. Slaughter. And your city shrinks by 1, possibly being obliterated. If the Warrior attacks, it's 1 Attack vs 1 Defense (with defender's advantage). You'll still be crushed. And don't even think of trying to swarm him with a bunch of Warriors. First of all, you're likely to turn him into a veteran before you finish and therefore get completely flattened instead of chipping him down. Second, the sheer number of Warriors required would take dozens of turns to produce, crippling you with high upkeep and delaying your production of anything useful. And that's against just 1 Archer. And assuming he doesn't get to sit on any good terrain near your city! They often come in pairs. Warriors also get smashed by Horsemen, but you DO have non-zero odds of being able to beat a Horseman if you can first-strike it. Which has roughly 50% odds of being the case. So on average you have a small chance of beating one barbarian Horseman if you've built a Warrior. Yay?

But it's worse than that. Warriors are often worse than no defense at all. See, barbarians will NEVER try to ransom a city that has a defender.They will ALWAYS attack it instead. So as long as you have Warriors in the city, the barbarians will attack the city- and thus decrease its population or obliterate it. If you have 0 defenders in the city, the barbarians will let you pay them to go away instead. Which is a much better deal.

Furthermore, no other military unit is actually a good bet either. Phalanxes are crushed nearly as hard and cost twice as much. Horsemen actually have a chance (and also cost double), but you dismissed their tech as terrible to get early.

In short, Warriors are not actually very cheap, can't both be scouts and enforce martial law, completely cripple your early-game Production if you build too many, and are less than worthless as defenders.

2) Partial fair point, high Settler cost IS slightly painful as a Republic. On the other hand, it's not THAT bad, as I'm demonstrating with my current game. For one thing, a Settler stopping to make just a bit of Irrigation solves the problem permanently for that city. For another, the problem only persists until the Settler founds a new city, which should only be a few turns usually. Third, it doesn't actually hurt your overall Settler production rate much at all. The limiting factor on how fast you can make Settlers is typically production, not Food. You need 40 shields vs 20 units of Food and automatic Irrigation + easy Irrigation of many tiles makes Food easy to come by once you're not a Despotism.

Where I really have to disagree is that you don't switch to Republic until you can start mass Caravan production. That's not just delaying a few dozen turns, we're talking possibly a hundred or more as a Monarchy now. That is a TON of time spent with cruddy Science and Tax income. Definitely not worth it. The other thing is that this mass Caravan point is actually when a pure Monarchy smallpox player suddenly blossoms and becomes great. This is the point to double down on Monarchy and steamroll all opposition, not to completely switch strategies to focusing on Science after setting yourself back by most of a hundred turns.

3) Yes you can. But this is wasteful. It's better to not build the unit in the first place if you don't have to than to build it when shields are precious and then throw it away for only 50% of your shields back at a time when shields are more common.

4) Republic does this better. Not even close. That's Republic's main specialty.

5) I agree 100%. The MC is the best wonder in the game regardless of strategy on the high difficulties. No ifs, ands, or buts about it. J.S. Bach's Cathedral is nearly as good and you definitely do want both ASAP. Plus Hanging Gardens if you can get it. Those are the big 3 that I really want early on, with Great Library a distant 4th. However, it turns out Theology doesn't actually give that +1 effect to all cathedrals in Civ 2. It does that in Freeciv and I thought it did it here, but it doesn't. In Freeciv Cathedrals start off working on 3 people, upgrade to 4 with theology, and then downgrade to 3 with Communism. In Civ 2 they just start working on 4 and then downgrade to 3 with Communism.

6) True. More conveniently, a Republic has massively more Science than a Monarchy anyway though so you'll still get there faster.

7) Hold up. First of all, no early tech is unwanted. Some are higher priority than others, but ALL are needed eventually. And most are needed quite soon. Let's look at what's required to get to only Monarchy and Theology and Trade. Alphabet is required. Bronze working is required. Ceremonial Burial is required. Horseback Riding is required. Warrior Code is required. You can only do without Masonry and Pottery and Map Making even in this extreme short-term approach.

Pottery is a great tech since it unlocks the excellent Hanging Gardens wonder. Well worth picking that up since we're talking mainly about how to get the anti-unhappiness wonders like MC and JSBC. Hanging Gardens is the cheapest and earliest of that type of wonder.

Getting access to Triremes early is also extremely valuable, but you can definitely put that off until you have at least MC and possibly also JSBC done.

Every single one of those tier 1 techs and nearly all of the tier 2 techs and most of the tier 3 techs they unlock are necessary for the critical Railroad tech that opens the industrial era of the game.

Those low tier techs are precisely the ones that the Great Library will get you while you forge ahead to advanced goal techs like Democracy or Gunpowder. But because it will take some time before you actually BUILD the Great Library, you have ample time to get to Republic or Theology or whatever you want before completing it. Thus no slowdown. Only a massive speedup going into the rest of the game.

Second of all, the amount you get slowed by learning a new tech is really quite small anyway. This is not some kind of massive, crippling problem ever. I'll concede that there are certain, very narrow circumstances in which you don't want to learn all the techs you can because you want to absolutely blitz to something key like Republic or Monarchy, but other than those situations you are better off getting as many free early techs as you can. Particularly if you're playing as a Monarchy for a hundred turns or something and therefore have a bad Science rate.

8) This I have to strenuously object to. The potential benefit of a goody hut is absolutely enormous. Getting a permanently free Settler early on is amazing. Getting an extra city early on is also usually amazing. Getting a solid military unit that you don't even necessarily have the tech for? Very handy. THOSE are your scouts. THOSE beat the barbarians for you. Look what I did with the 2 I got in this game. Getting a pile of gold when your income is like 2 per turn? Great. Getting a free tech? Also great because once again, you need every single early game tech eventually and the amount of slowdown you get (assuming the random tech you receive isn't actually one you want for your current project in the first place) is very small. The only thing you have to fear is a massive barbarian swarm. And honestly it's not unreasonable to just restart or reload that turn if you get one of those. Those are a stupid possibility and should not be in the game early on.

9) Wait what? You NEED Horseback Riding to get to the all-important Monotheism. And Horsemen are the first actually GOOD military unit. And you need Warrior Code to get to Theology.

10) Yeah, this is the other big one. This can be catastrophic and can destroy even the largest technology leads you have in a single turn. Do NOT let someone else get the Great Library if you can avoid it. Unless you're going for a low-tech early game win, in which case why ever go Republic?

11) Yeah, I won't do this. Yes, it's powerful and efficient. But it's an absurd and cheesy exploit when you take advantage of the poorly thought out and secret "Ref Civ/Key Civ" mechanics like that. You're supposed to have higher prices if you're ahead. That's supposed to be part of the game and it's a good part of the gameplay, if poorly implemented.

Again, there's nothing wrong with playing the game you want to play, but I definitely hold that the Great Library is excellent, if not top priority, and that spending a pile of time as a Monarchy before switching to Republic is inefficient compared to either rushing Republic or staying Monarchy and going for early conquest.

Delenda est Carthago!