Ideally, I want something realtively simple – something like the Axis & Allies board game or maybe a step up from that, for a PC platform. But what I really want is a really smart AI opponent. The ones I’ve tried have been remarkably dumb and repetitive.
FYI, I saw that there was PC Axis & Allies game, but I saw two different reviews panning the AI opponent.
Any suggestions?
I’ll sign on to see the suggestions. 
Only modern military?
Hadn’t thought about that. Probably, but if you have suggestions, I’ll certainly consider them.
These might not be what you’re looking for but what the heck. The gold standard of easy to learn, hard to master, reasonably complex is Civilization, I’d suggest IV or V. V is smoother and simpler.
And a perhaps odd suggestion, the old Advance Wars game that ran on the Gameboy Advance. It was fairly simple and challenging.
I’ve heard of Civilization but never played it. Can you give be a brief description?
The latter game sounds like it’s for a different platform. Is that true?
The Civilization series has reasonable AI, although exploitable. You can adjust the settings to create a more militaristic game. Civ V was something of a step backwards as the AI has difficulty with one unit per hex. (On preview: Hi control-z, what a good suggestion.)
Galactic Civilizations II has a reputation for good AI. I’ve only played it a bit.
Rise of Nations was challenging, IMO.
There must be a hundred PC Risk clones, versions, and homages out there. Even if the AI is weak, you can turn up the challenge by increasing the number of opponents.
Not really my thing then, but I can recommend a few:
Total War would be my first suggestion if modern isn’t a requirement. Medieval 2 and Shogun 2 might be good places to start with that series. At the harder difficulty levels and/or with AI mods the games can be quite challenging. Empire total war also features the most modern settings for the series… I believe it goes up to the early 19th century.
WarGame: AirLand Battle might be more your speed though. It focuses on battlefield tactics over grand strategy however. Heard really good things about the campaign, but I haven’t played it myself yet so no comment on the AI, but I heard it’s good.
DEFCON plays out almost like a board game. In fact the UI would be at home in a game of risk. It’s all about strategically placing various assets around the world in the prelude to a nuclear war. It’s fun and strategic but I only ever played online with others, so I can’t speak to the strength of the AI either. I know, some help I am!
Blasphemy! 
It’s a turn based game. During each turn you control a number of units and manage a number of cities. You start with one city on a map (the size and composition of which you decide) and your civilization (many to choose from can also create a custom one) starts out in prehistory (this can be changed in the settings. you can absolutely start a game in modern times if you want). During each turn you can scout the map, discover various points of interest (barbnai encampment,s natural wonders, ancient ruins, resources, etc), build new cities, research new technologies, negotiate with other civilizations, or wage war.
I found the AI in Civ V (with expansions) to be rather good. The end goal and how the game plays out is available for you to control. You can change the technology level and even make certain victory conditions unavailable. Course the cool aspect of the game is that you don’t need to win by military might, though you can if you want to. Political, scientific, economic and cultural victories are also a possibility and change up the gameplay quite a bit.
I will attempt a brief description, most all of this applies to Civ4 and Civ5. Civilization is a turn-based game, you start off with one settler and one warrior and explore and expand by building or conquering other cities and civilizations around the map. Each city can work on producing one unit or building at a time. There is a tech tree, you research one thing at a time. It starts off in stone-age times with things like the wheel or ironworking and works through the different ages up to modern times and beyond. Your tech knowledge dictates what sort of units and buildings you can make. You encounter different civilizations during your exploration of the map and there are basic diplomatic options, such as sharing maps, declaring war, trading resources, etc. You can win a diplomatic, technological, or military victory. Real civilizations are used, so each game is sort of an alternative Earth history. Like Russia might build the Eiffel Tower, China might build Machu Pichu, etc. You can play on an Earth map but default is randomly generated maps.
Advance Wars was a game on the Gameboy Advance handheld game console, but there should be ways to play it through an emulator on the PC if you’re determined, or just get a cheap Gameboy Advance. The coolest thing about Advance Wars is it’s pure strategy, no luck is involved. Almost like a more complicated chess. It does have a sort of Japanese-feel to it, but overall it’s just a neat little strategy game.
I’d suggest going to YouTube and searching for “civilization 4 gameplay”, “civilization 5 gameplay”, and “advance wars gameplay” to see what the games in action are actually like.
Even if they don’t match exactly what you’re looking for, Civ4 and Civ5 are highly recommended and I don’t think you’ll be disappointed. Plus they are often on sale from Steam and Gamersgate for well under $10.
You’re looking for a white whale.
Any wargame with enough depth to be really interesting is completely out of the realm of good AI. Just look at chess; it’s a rigidly-defined game with a limited set of potential moves for any single turn, that’s had decades of computer science research thrown at it, and it’s still a recent development that computer play can be considered on par with a good human. It’s only achieved that because the advance of hardware has reached a point where there’s databases large enough and brute force fast enough to do limited solutions.
Anything with more complexity (and your typical strategic wargame will have a far larger action space per unit time than chess does) and without the depth of data and algorithms behind it is a lost cause. You’ll never find anything that plays as well as even a middling human.
Strategy games played against AI get their challenge not by playing well, but by presenting the player with an unfair fight, and there’s no chance of that changing anytime in the near future.
This isn’t entirely true, AI researchers have actually had really good success using various methods (mainly genetic algorithms) in playing Starcraft II. They’re not quite able to beat best-in-the-world players, but they’re shockingly close (mid-high masters level apparently).
However, you’re right that looking for any AI out of the box in a commercially distributed product that’s that good isn’t going to happen.
Wargame ALB is a unique little entry into RTS that anyone interested in the subject should check out, but AI is not one of its strengths. It’s a multiplayer game first and foremost, and even the campaigns are intended to be played between humans, i.e., they’re all structured as equal, or at least not totally slanted, odds.
Anyone with a decent familiarity with modern tactics and how the game interprets them will just beat the pants off the AI. Players crushing AI is pretty typical for RTS games, but it’s even more so in this one, because there’s no macro- or micro-management hijinks that the computer can hope to save itself with. One of the game’s design goals was to try to make the player interact with units at a higher scope than usual, and they’ve succeeded well enough at it that the computer is hamstrung pretty badly.
Yeah, the AI stuff being done with SC2 is pretty neat, but I’m a little leery of calling it legitimate play.
I’m not at all current with the last year of SC2 game theory, but as I understand it, the AI is seeing success not by really competing with humans on the tactical or strategic side of things, but by simply not being so terrible at those that its enormous advantages in sheer mechanics and efficiency can win through.
It’s not really even playing the same game; there’s quite a few bits of SC that are designed around the core assumption that player time is a valuable commodity, while a computer has that in nearly infinite supply.
Have you looked into the Paradox games? Hearts of Iron, Europa Universalis, and Victoria II are country games. You control a country through its time period. That’s it. How much you go to war and with who is your decision.
Difficulty is based on which country you choose: there’s a huge difference between Germany and Bavaria.
Without turning this into a huge philosophical discussion on AI, I disagree that it’s “illegitimate”. As far as I know, it’s not cheating. It doesn’t have any knowledge that any other player has doesn’t have access to (e.g. it’s still limited by fog of war), and it doesn’t have any resource advantages. I think it’s part of the grand disillusionment with AI when something like this is criticized for being “real” – “it doesn’t do this like a human so it’s cheating” is kind of a criticism that kind of keeps all AI into the realm of “cheating” because the fact is we don’t really know how humans do it, so it’s tempting to call any artificial solution wrong.
Either way, the outward effect is an AI that can achieve high levels of play with the same information a person can have, regardless of the specifics. There are real players who make up for shitty macro with really good micro, and all sorts of other things. Sure, the AI has an even further advantage in that it can micro every marine individually – but I view it as just another player quirk. You can argue the game wasn’t “designed for someone who can do that”, but it’s kind of a wishy-washy e-bushido argument. It’s acting within the constraints of the game, sure it’s not doing it in a way the developers necessarily intended, but it’s still doing it.
[sub]This is a really similar to the issue I have with the “Chinese Room” argument, Searle seems to really rely on the fact that biological organisms are doing something “magical”, and anything different is a cheat and has no real understanding. But the catch, as with this, is we don’t know how the biological organisms are doing it, so therefore all artificial attempts are illegitimate cheats[/sub]
The best AI in this line is in the Paradox game Hearts of Iron 3. But it’s an insanely complex and detailed game, so may not suit you.
That’d be Victoria II. Will you be swallowed up by the Southern German Federation or Prussia as the German states try to unify or will you successfully hold back the German horde and lead Bavaria to great destiny? There’s a burgeoning threat to your north and south as two countries consolidate but there’s opportunities to your east as the foundations that the Hapsburg built their empire on begin to show its cracks.
I’d highly recommend Unity of Command. The AI is very clever in that it often plays the game the way a person would. Instead of always forming pretty straight lines across the map, it will block the road to slow your advance, and may just bunch up around important objectives. More impressively, if you leave a valuable unit exposed and it calculates that it can be destroyed, it will send multiple units to obliterate it in a coordinated attack. And it’s especially good about being cognizant of supply: the AI notices when you’ve given it the opportunity to punch a hole in your line or go around your flank to cut off your supply lines (which can be devastatingly effective since you’re usually racing against the clock).
And it’s just a really fun game. It doesn’t hide the fact that it owes a debt to Panzer General, but it makes several improvements on that game.
Going back a ways, I always thought that Sid Meir’s Gettysburg! (followed by Antietam) had a great AI. It’s not like the computer never made mistakes, but it managed good situational awareness. If it noticed you marching troops off around its flank, it would adjust that whole side of its line compensate. It was also better than a human could be at keeping troops abreast and in proper formation, which was a big part of the game, and that ensured that it would be competitive against even skilled players.