Which is fine, given that the AI in Civ 6 is terrible even by the standards of other titles in the genre.
I’d also toss Unity of Command under the Very Good header. Agree on OTC. Would probably nudge EU IV to good, given it does diplomacy better than any game, well, ever.
I would probably push some of the AGEOD games into at least fair. Civil War II is at least competent with the tactical and maneuver level. I’m not good enough myself to know their quality at the economic side, but at least in the more focused campaigns seems to hold up well.
But your list is a pretty decent pull. None of them seem egregious inclusions, and there are no omissions of major games.
That said I believe it was Soren himself on his podcast that brought something up. Can’t recall exactly the interview, but the gist was that people don’t want an AI that plays mathematically perfect. People don’t do that. They want an AI that plays believable.
It might have been Sid Meier, because I think the context was ‘if the AI really wanted to win, all factions would declare war on the player turn 1’. But that wouldn’t be fun, likewise the random war decs are not fun in Civ 6.
I played through one game of Civ VI and gave up. I can only blame myself for believing this game would be better than what is was. I set my expectations high, the game well and truly underdelivered, and as a result, the crushing sense of disappointment pushed me away. The ultimate kicker was when I started a game as Pericles, came across a neighbouring Pericles, and then realised I was going to be fighting myself.
I had lofty expectations because like Alstein, I went with the general vibe that the even Civs have been the best. The odd Civs do something new, the even ones refine the new stuff. Furthermore, there are what I consider basics that Firaxis need to look at. The first is diplomacy. Admittedly, Paradox Dev games have ruined diplomacy well and truly now because they do it so well. Civ on the other hand - I declare war on a nation early, I’m a warmonger for life. 4000 years later and the AI still holds a grudge? Awesome. Marry that with the current woeful system in Civ VI where the AI doesn’t give a toss if it is a defensive war and how suing for peace is laughably limited.
Armies… I expected a Call to Power throw back with armies, not this shitty brain dead system. Yay, I can link a general to a unit? Whoopee. Or a builder so it doesn’t get sniped if I recall. Wake me up when they do something better with it please (ie: Call to Power). Sure, the AI might do a better job moving units around, but then, who cares when all it has are warriors to defend itself with while I’m marching around with tanks. Something else wrong when I played was that the AI didn’t invest into a modern army.
I was quiet when Civ V came out. I was disappointed with it on release, and it took 2 expansions for me to fully appreciate the game. Yes, I actually liked Civ V + expansions. It isn’t my most played game, but it is up there in terms of most played. I could look past the tactical AI and appreciate the non-warfare elements better. Even diplomacy didn’t bother me. Yet Civ VI, all that stuff now bothers me because none of it has improved to any meaningful extent.
I look back at the games Firaxis have released lately and I wonder if they’ve truly lost their way with game design, or if maybe I’ve acquired a new taste in games (ie: Paradox). Once I thought Firaxis was synonymous with quality, but from Civ V onwards, it has steadily gone downhill. And yeah, I didn’t mind XCom, but holy shit both games had some major flaws. One of which was being able to game the system for the first game by only activating “pods.” So to get around that, the second game they include patrols (yay) and timers (hohum). I don’t mind the timers, not a game breaker, but again, they needed to look at games like the original XCom, or better yet, Jagged Alliance 2 (with 1.13 patch) and see how the AI responds when ambushing for instance. It is possible to stealth whole missions in JA2 for instance, and by firing, it is enough to pull patrols into the general area. Furthermore, XCom provides no recovery mechanic when a mission goes bad, and the aliens, well, certain ones merely stop showing up after a while. So much for being able to train rookies up…
At this point it looks like it’s not even game design. There’s not even enough will or wherewithal in the development team even to make the most ridiculous of bugs/logic failures out of the game.
I decided to buy Unity of command so we will see… You are probably right about EU IV, I just haven’t played that much or that recently.
I think both Soren and Sid have made that point. I attended an AI talk Soren did at GDC many years ago, and he made lots of points about the players need to believe that AI was playing in fair and believable manner. To the point that if you had 4 or 5 bad rolls in row players would think the AI was cheating, so that they changed the RND algorithms so that after 4 bad luck rolls you were virtually guarranteed to get a good one on your 5th attempt.
Maybe they can fix VI, but like Stellaris it has been put in my wait and see category. If a new patch or new expansion really makes a difference I’ll play it, but I don’t think I’ll be playing a lot until that time.
I just hope Brad’s reaming of the genre recently in general on reddit leads to something. It sounds like he’s going to try next year.
I would include a number of Slitherine/Matrix titles as having good AI, notably:
Hannibal:Rome & Carthage in the Second Punic War
Victory & Glory - Napoleon
Pike & Shot series (including Sengoku Jidai)
Vietnam '65
War in the East (and, I assume, West, though I haven’t played it.
Decisive Campaigns:Barbarossa
It’s too early to tell but I’m beta testing a 2017 release of theirs that appears to have good AI, but that’s subject to change :)
It’s a little easier to do wargame AI compared to 4X AI, so that’s not entirely a fair comparison.
I’ve not played most of those @tgb123 but, given I include several other of their games, I would definitely not discount that possibility. Perhaps one day I’ll give Pike and Shot a go. That magical day when I have time…
Yet still hard. Many wargame AI’s are of the bottom-up statistical sort, where everything is given a value (think Norm Koger or Gary Grigsby) and added up to other values, which then go into algorithms that make decisions. There’s usually supplementary logic of some sort specific to the situation but essentially it’s a pretty mechanical thing, which works (when it does) because the player gets tons of fiddly bits to play with, and sometimes obsessing over the number of riflemen in your battalion can take the cognitive place of worrying about actual tactics or strategy.
What rarely if ever happens in wargames is AI that plays holistically, using actual doctrinal philosophies as guidelines and making hgh-level decisions based on something more than totaling up arbitrary valuations of firepower or whatnot. There are a few that try, with varying degrees of success, like Command Ops, but it’s rare and usually quite iffy in execution. You still need values–computers being computers–but there’s a world of difference in rating soft factors and fitting the values into a logical construction based on analysis of doctrine, and totaling up the number of jeeps times the firepower value and comparing that to the values derived from the number of cooks with butter knives or whatever usually happens.
In either case, though, it often seems like a wargame has decent AI, because either the actions are hardcoded (attack here on this turn) or the systems are so complex we feel like stuff is happening logically, as we’re engrossed in mastering the systems.
Civ games are less suited for a bottom-up approach, leaving the more difficult top-down conceptual approach as the necessary one. And the results, to say the least, are mixed.
Wow, I feel almost guilty enjoying Civ VI, AI flaws and all. I think it’s fun. And I still haven’t beaten it on a level above King.
I’ve had fun with it, too. It’s just that usually, a Civ title is compelling me to play, and this one isn’t for some reason.
Goodness me no, don’t feel guilty at all. You are in with the majority of people who enjoy Civ. The Steam reviews say it all really, with 80% positive hit rate. For me, I guess it really was a case of setting the bar too high with my expectations, thus the disappointment.
Just to remind you all, the mod AI+ is still getting better and better:
Quick question. I don’t have the game installed.
Is this list of keyboard bindings complete?
http://civ6.gamepedia.com/How_To_Play_Guide_for_Civ_6#Keyboard_Controls_.2F_Hotkeys
Do the arrow keys do anything, such as scroll the map?
Does ALT+F4 exit the game?
Thanks.
In my current Jadwiga games on Prince I started on a large land mass sharing it with only India & Japan, but 9 or 10 city-states. So far I haven’t seen the AI do anything obviously stupid. I established a friendship with Gandhi from the start, and going in the Renaissance we’re still BFFs. - surprisingly he hasn’t DOWed me. I think Hojo is going for a religious victory as I am, but I haven’t seen any missionary spam - yet.
So far, so good.
Arrows do scroll the map, and ALT+F4 prompts you with the question on whether you’d like to quit the game (to desktop, btw). ESC also brings up the menu for saves, quitting, options, etc…
In Civ 6 we the player are thinking 20-30 moves ahead or more, maybe much more in some areas as you start to plan your victory type as early as the middle ages or sooner. This in a system with at least three categories of strategy (economy, combat, culture, etc). Even in AoW3 or wargames that are held as examples of good AI, I don’t see the look-ahead depth or inter-relationships as complex. How many moves ahead does the AI search in a popular hex-based wargame, honestly? Or EUIV with a lot more pre-defined assumptions in place (could be wrong, would be interested to hear)? I’m curious to discuss some comparisons. Coding look ahead depth can be easy in some cases (chess), bit seems much harder in Civ.
Also we get so upset that the AI declares a war it can’t possibly win despite just declaring friendship and receiving resources from us, then sends a wimpy dozen archers against our comparative stealth bombers. Yes it’s unacceptable, but maybe the AI sees our exponential growth and it is behind in tech doesn’t have stealth bombers, so what exactly do we expect it to do? Standing by and letting a superior foe get more superior has worked so well in history (or with the bully on the playground :).
On the other hand, Civ6 has always been for a casual PC gamer audience, which is increasingly incompatible with the people reading this thread (no insult intended, a lot of us are looking for something with more depth). Why aren’t there a thousand posts on AI in the Civ Revolution for iPad thread? Because the games are for the same audience, at least more for the same audience than say Hearts of Iron 4.
Given sales and bugets as they are it seems to me like higher-ups at Firaxis aren’t likely to approve a $100k AI improvement budget, let alone a $million one. So I’m not sure a sea change in 4x Civ AI will happen soon. That said, there seems to be some fans out there doing it via mods, so maybe we can all be satisfied in the end?
Long ramble short, I agree the AI is awful. I still enjoy the game, need to play again as I’ve been distracted lately.
I am actually going to give Civ V (with all expansions) and the Vox Populi mod a shot.
As someone who never played much Civ V because the AI sucks, I’ve heard that Vox Populi makes the game (and AI) a lot better.
VP is quite good, yes. IMO it’s up there with Civ 4 + BUG, but I’m a filthy V lover to start with ;)