3041

But how do you determine what buildings, if any, survived the attack?
When you raze/annex you can go into the city menu and see the building list on the right, but I can’t access the city menu of a puppet.

You settle your settler where it is. If you don’t like the spot start a new game. I think Civ 4 had an instant restart option in the menu for this Otherwise you will always be behind the other players in terms of growth.

3042

The bolded part is wrong. Other games, famously including the very Panzer General from which the new system was taken, manage a one-unit-per-hex system with complex terrain & positioning bonuses quite well. So it’s really a (fixable) defect of the AI and not of the combat system. The Civ5 combat system is definitely not beyond the ability of modern CPUs – only beyond the willingness of Firaxis to pay for a decent AI, it seems.

3043

Yes, but you were genuinely helpful, I was just setting up a pun.

Other TBS games have managed more complex systems, with AI equal to the task, but they are generally much more simple games, where the combat is basically the entire game. Civ is a much more ambitious beast. They have to decide where they are going to expend their effort. And if they don’t think its worth spending time on a complex combat AI, they need to simplify the combat.

I’m not saying its impossible. Fraxis have more resources to draw on than the creators of Panzer General. I’m hopeful a patch will improve things.

My point is that AI and game design are linked. If you just spend all your effort trying to design a cool game system, and then decide “oh crap we need to teach the AI to play this game”, then you end up with Empire: Total War :)

3044

Of course I agree it’s a defect in the AI. The system is not that complicated, and there tend to be much fewer units in a given tactical situation than in most straight-up wargames.

But PG is not the world’s best example of good AI. Panzer General certainly has better AI than Civ 5, but it’s not actually good as such. By coincidence I recently got it from gog and played a couple of scenarios. It has a strong tendency to just sit around doing nothing rather than counterattack, or if it does counterattack, it just randomly does something with a frontline unit because the unit sees an enemy. It seems incapable of coordinating the movement of multiple units in the open field to take an objective. In PG, this isn’t such a big problem because either the player is the attacker or else there is some simple scripting for the initial assault the player has to weather as a defender.

But of course Civ 5 can’t have any scripting, and if the AI can’t figure out how to attack a human player, it’s just hapless.

However, I must admit while some hex-based wargame AI is worse than others, I’ve never really seen good tactical AI. The best I’ve seen was probably in Atomic games’ stuff like Operation Crusader, and that wasn’t actually good as such. And again, with fixed historical scenarios, it was no doubt helped out a lot by scripting.

3045

Yup. There are lots of social policies that increase global happiness as well.

3046

You don’t want to waste turns scouting with your initial settler. You can move your warrior first to uncover a little more territory, then either create your first city right where your settler is or move it one hex and then settle there (make sure that one-hex move doesn’t take up all of your movement points due to terrain, otherwise you won’t be able to use the “settle” action until next turn).

The game should put you in a decent starting spot to begin with, though occasionally you do get a little screwed.

3047

However, I must admit while some hex-based wargame AI is worse than others, I’ve never really seen good tactical AI. The best I’ve seen was probably in Atomic games’ stuff like Operation Crusader, and that wasn’t actually good as such. And again, with fixed historical scenarios, it was no doubt helped out a lot by scripting.

Agreed. I’ve played many hex-based wargames (both board and PC games), and I’ve never seen a really good tactical AI. Some games get by with scripting; “War in the Pacific/Admirals Edition” will give you a good game, at least for a while. But scripting won’t work with a flexible game like Civ.

I’ve seen worse than the Civ AI. At least the Civ AI tries to attack; in many hex-based wargames, the AI plays OK on defense but hardly moves on offense. I’ve had the Civ AI (on Prince) invade me amphibiously and successfully; naval AI in wargames often can’t manage that at all unless they’re scripted, and I don’t recall the Civ 4 AI doing that as well. (I think the new embarkation routines may have helped Civ 5 a bit.) If my military is inferior to a Civ5 AI and it invades, it will often succeed in taking one or more of my cities, even if it makes inefficient use of ranged/melee capabilities and loses units needlessly. Unquestionably the Civ 5 AI wilts if it’s outnumbered, and in general it plays poorly on defense. Lots of room for improvement there. In general, though, the AI doesn’t have to be fabulous to keep me entertained. I think Civ5 is fun.

Also, some aspects of the Civ AI are traditionally designed to give the player a good experience, not necessarily to win at all costs, or at least so said Soren Johnson in his video presentation “Playing to Lose”. E.g., Civ AI’s roleplay their characters, even if it’s not in their interests to do so, whereas we humans (and Vulcans) play as we like. The AIs roleplay to make the game more fun for us, even if it hurts them. Likewise, the AI’s diplomacy choices may be more constrained than ours, so that the AI’s diplomatic choices don’t seem unduly annoying – or at least, that’s what Johnson said. Johnson also said they took some AI options off the table in Civ 4, such as building annoyingly close to the player; I wonder if this restriction has been somewhat relaxed in Civ 5. Last but not least, we can reload, and AIs can’t. We tend to take this huge advantage for granted, but it is a huge advantage for us humans.

Somewhere I recall reading Firaxis say they were taking some AI tactics off the table at lower difficulty levels in Civ5, but letting the AI do everything at the highest levels. Is the tactical combat AI any better on Deity than on Prince, say?

3048

I was a tester on that game and helped with the AI scripting. I’d play the game and pretty much document what I did each turn, and the programmers would work that into a script. At least, that was how some of it happened. I doubt my contributions were terribly profound, but the scripting they did helped the AI play a decent game I think.

3049

Spock: I think it was building and settling and hose kind of decisions that the AI is better at on higher difficulties, not tactical AI. At least, someone posted some XML snippet for the King level earlier in the thread, and I don’t think there was any variable modified that looked like it was for combat AI decisions.

3050

So after a week or so, is the “Digital Deluxe civ V” containing the extra civ a value?

3051

Because I paid the extra $10 for Babylon I won’t play any other civ to be sure I’m getting my money’s worth. So take that into consideration :)

3052

I thought so. The extra Civ (Babylon) gets its own unique race trait that’s quite different from others. Their unique unit and city improvement is quite different. They get their own background soundtrack while playing the game (like all the other races get), so I felt it was a substantial amount of content. Is it worth an extra $10? Well, that’s an individual call.

3053

I’d vote yes, but I also got the sale from D2D as well - $10 off and also including the first upcoming double-civ/map DLC.

3054

After playing through a full game, starting another game on a higher difficulty with very specific initial conditions, and messing around a bit I would like to give some of my thoughts on the game. Please note that I am no Civ expert, have probably played Civ III the most (since it worked well on my laptop), and have likely played a couple hundred hours at most of all versions of Civ combined (as opposed to the many hundreds or thousands that others here have played and the many hundreds that I have put into Master of Magic). I am also not that great at TBS games since I am pretty poor at making a long-term strategy and sticking to it .

I’ll start with a brief-ish recounting of my first game. It was on Chieftain difficulty, standard map size and standard length. I chose Japan as my Civ for no real reason other than maybe liking the portrait for Nobunaga. Since, from what I had read, it sounded like a cultural victory was difficult that was my goal at the start of the game (I realize that, at face-value, Japan isn’t a good choice for this). It was a stomp-fest with me conquering my continent early on, working on my cultural and science output until the industrial era, and then conquering everyone else except the two largest remaining civs (we each controlled about a third of the land). I won the game with a scientific victory in 2003 with a score of 4450. Though it was a bit cheesy I saved right before completing the spaceship and by reloading the save was able to get a diplomatic victory (2016), domination victory (2021) and a cultural victory (2032).

What I found striking in the end game was that I was pulling in massive amounts of science (3000+), gold (700+), culture (700+) and happiness (100) every turn. How much of a direct penalty would playing on Prince have had on this? That is, if I was in the same situation how much lower would my happiness, culture, etc… have been? I realize that on a higher difficulty the AI would have likely built more wonders, had a stronger army and been closer to me in the tech race because of their bonuses and increased competence, so making a direct comparison is difficult. I’m just wondering if I would have been crippled by unhappiness on a higher setting or if I wouldn’t have had as much gold to buy off city-states or purchase whatever I wanted at any time.

I don’t want this account to give a negative impression of the game. It may have been easy, but I had fun the whole time (a little over 11 hours) and really enjoyed a number of the new features in Civ V. First off I am a big fan of the interface improvements and how easy it is to see what’s happening and needs to be done throughout your civilization. I was never a huge fan of Civ IV’s interface (and it probably kept me from playing more of that game). I also enjoy the change to hexes and the lack of military unit stacking. This makes the tactical wargame aspect of Civ V much more appealing (and I think brings it closer to MoM in terms of making combat fun) than in previous versions. For the most part I liked the resource and management streamlining, but I was never a high-level enough player to really worry about fiddling with all the details in previous games. I find City-States to be a very good replacement for religion. I also found the Diplomacy options to serve what I think is their main purpose well (making trades for specific resources), and since I don’t ever delve into the diplomacy side of things I didn’t notice any glaring weaknesses.

As for negatives, what I noticed the most were some of the weird rough edges. For having such a polished and intuitive interface there were a number of strange graphical/presentation glitches: unit icons appearing over tiles with no units, tile resource graphics moving as I moved the world, roads not connecting up properly in adjacent tiles, occasional checkerboard textures on tiles (could be a lack a video RAM), and a few others. At one point in the late game I liberated Monaco which gave me the 150 influence points and their status said they were my ally, but there was also a message saying “is at war with you” (the message did NOT state who was at war with me) with the city-state icon for Monaco beside the message. Monaco also attacked my units and acted aggressively towards me despite apparently being my ally. I “solved” this issue by just conquering them again. Although I never conclusively pinned any examples down, it did seem like a number of treaties/agreements never actually expired (especially “Open Borders”) or expired at the incorrect time. This is not even mentioning the well-known problem of mods not installing automatically if Steam isn’t installed in the usual place on the C: drive (mine is at d:\steam) or the lack of a satisfying end-game summary/graph. All these bugs/glitches/omissions make me think the game was released a little early and could have used a couple more months of clean up. Hopefully we will see at least the same level of post-release support as previous Civ games.

As for the AI the many previous comments have covered it far better than I could, but I certainly noticed that tactically they couldn’t hold a candle to human controlled units (I have played a bit on Prince difficulty and this also holds true at that level). Strategically and diplomatically I haven’t noticed any particularly poor decisions, but that is likely a function of my expectations for the lower difficulty and my own lack of diplomatic and strategic skill. I suspect that after another 50-60 hours I could recognize some large flaws with the AI, but I don’t see that being any more of an impediment to my continued enjoyment than the brain-dead AI in MoM.

To end this rambling screed I will say that I am very pleased with Civ V and hope to see it polished over time. I am also hopeful that the mod scene will take off enough provide real improvements to the AI weaknesses and produce some interesting new game-modes (An Advance Wars style scenario based campaign? A Civ-ified version of Defcon? Super Fall from Heaven?).

3055

PG is extremely heavily scripted; this is very apparent whenever the player takes an out of the box approach to “solving” the scenario. It is also the reason why Fantasy General - which added new dimensions of complexity to the design - was also the weakest version of the game AI-wise (they compensated for this by stacking the odds heavily against the player).

Though I have to disagree on your assessment of OC. Atomic Games were the kings of hex-based AI, IMO; by the last iterations they had AI that could both attack and defend, rotate units from the front-lines, form pockets (don’t know how many times I drove for the beaches in Gold Juno Sword just to have my panzer divisions cut off and surrounded), and outflank effectively. Sure, it cheated a bit with the fog of war even on “equal” difficulty levels, but it was competent enough to require you to play well in order to beat it. I don’t see that you can ask for much more from an AI until we get positronic brains.

3056

I’m not sure what definition of scripting you guys are using; there’s none at all, using “hardcoded instructions telling the AI to always do a specific thing in a specific circumstance.” There’s no scenario-specific AI packages, there’s just one set of AI for all of them. There’s an offensive AI (taking victory hexes, nothing else) and a defensive AI (defending the current victory hex front line, nothing else). We know this because that crazy guy rewrote the game! :)

3057

I just don’t think the weak tactical AI is that big a deal. I find the whole game to be easier than Civ IV, but I could never get anywhere on a difficulty above Monarch in Civ IV and I’ve played a ton of Civ, so while I’m sure there are plenty of people who are much better than me, I doubt I’m on the lower half of the curve. I have no idea how things scale when you get to Diety, but having to make the most of your smaller resources to beat the AI is exactly what the challenge is about. I just don’t really buy the argument that the game isn’t fun if the AI has to “cheat” to compete with you.

3058

That’s never made any sense to me either.

3059

I didn’t play any of the later games in the V for Victory series, so the AI might have improved in those games. In the large scale OC scenario with a lot of desert to move tanks freely around, the OC AI is very bad, and can’t coordinate any attacks. It’s like Rommel was invalided out a couple of years too soon and replaced with a magic 8 ball.

As I recall the OC AI was much better in the smaller tighter scenarios that are more amenable to scripting. But as I said, even with its open-desert shortcomings, it seemed superior to most other hex-based AI, including of course that of Civ 5.

3060

By “scripting” in these games I mean basically a chess opening book, but unlike chess, it usually only covers a turn or two, with sometimes a bit extra for a planned event like reinforcements or some “story” action. Something like this:

“On turn 1, AI always moves this tank to hex 0421, and that infantry to hex 0422, and AI fires artillery at those defenses at hex 0627.”

So it’s hardcoded for the first turn, and maybe on the second turn there is a script that is overridden if some unit is unexpectedly engaged somehow. In a typical 8 turn scenario on a small map, with a strong AI force, this puts a lot of stress on a human defender to stop the initial well-thought-out onslaught. But then the rest of the scenario may just be mopping up once the script runs out of steam. Scripting works much less well on a large open map or with a scenario that has a lot of turns to play out. Then you have to rely on the tactical AI, and if it can’t coordinate its units on a big map, it’s kind of pathetic.