AI cheats at this game.

[rah]

in this case it's the old rand making BS thing... just for lulz load the save 1T before and switch some mine worked into not working mine and do the turn

Changing whether a mine is worked changes the RNG seed??

 

[coanda]

Yes, because there is a small random chance each turn that a mine you work discovers a resource (like gold). So working a mine adds one extra call to the RNG.

 

[rah]

Ah, didn't consider that. Interesting to know. I'm not a normal reloader but if I mis-click, sometime I'll reload and I'd like to make sure that I don't change anything.

THANKS

 

[TheMeInTeam]

as for AI not declaring at pleased/friendly...I actually like it. I saw how bad diplomacy can get when design goes the "AI wants to win no matter what" in Civ V, basically meaning that game has NO diplomacy whatsoever.

The problem with civ V AI is not that they try to win, but that they do ludicrously stupid things with bias against the human only. In other words, civ V AI doesn't try to win either so it's no an appropriate example to make your point.

Yip decayed fast, can still be annoying thou cos they might just decide plotting on you during those few turns

I dun like that mechanic, but would also not call it cheating..AI can miss out too, if i make that tech trade with somebody else cos they are grumbling.

AI can plot on you the turn it meets you, so it can indeed matter if you wind up being it's "worst enemy" for the sole fact that you traded with a worst enemy.

And yes, the mechanic is cheating...whatever justification you want to use for the cheating is one thing but call it like it is.

Soren says that the AI more or less use the same automated workers as is used for the human.

Soren lied, unless by "more or less" he meant "the AIs workers will not move onto the borders of civs they're fighting, but automated workers will gladly do such suicide". That isn't more or less. One side's workers routinely suicide, the others don't. AI workers also build improvements based on its preferences which sets them apart though you can use emphasis to get your workers to do something similar.

Why post here?
Well..is winning this battle with >33% hps left, chances lower than 1%, anything else but cheating?

I've lost a full health knight to a damaged catapult, while the game literally displayed 100% odds, so you're still beaten in RNG misfortune .

Ah, didn't consider that. Interesting to know. I'm not a normal reloader but if I mis-click, sometime I'll reload and I'd like to make sure that I don't change anything.

I generally only reload when the games user interface tries to stop me from playing it by moving units I'm trying to command against my will, such as catapults next to AI "pillage" knights...and the occasional TROLLFACE WE MOVE BUTTON AS YOU TRY TO CLICK IT episode.

Since kmod fixes these things I'd be happy if S&T went to kmod lol.

 

[Mylene]

I've lost a full health knight to a damaged catapult, while the game literally displayed 100% odds, so you're still beaten in RNG misfortune .

But that was not game over for you i bet
Watching your second city being burned down is thou ~~

 

[TheMeInTeam]

But that was not game over for you i bet
Watching your second city being burned down is thou ~~

Oh, well if you want to see something arguably worse, look up my youtube game titled "haxmaster darius", where I lose TWO archers defending a city to 2 barb archers and a warrior. The odds of that happening were almost certainly lower than your fight too, and as a bonus it ALSO cost me a 2nd city though I don't think it was burned.

I won anyway because it was on immortal, but if it were on deity? Probably not.

 

[Mylene]

Yip i just gave up, could have reloaded but meh.
Back to 1 city while Joao already has 8 on Deity, uh i'd rather start over

 

[Pangaea]

Soren lied, unless by "more or less" he meant "the AIs workers will not move onto the borders of civs they're fighting, but automated workers will gladly do such suicide". That isn't more or less. One side's workers routinely suicide, the others don't. AI workers also build improvements based on its preferences which sets them apart though you can use emphasis to get your workers to do something similar.

He didn't say "Yes, they are the same," which was noteworthy in itself I thought. How much he puts into it I don't know - and analysing phrases like that can be a bit OTT anyway. What we do know is that they are NOT the same, as our workers never run back to the city to hide, while AI workers quite routinely do (or certainly away into the fog).

 

[Um the Muse]

I should have said in that last post that I was only talking about certain cheats. Certainly, there is a lot of diplomacy shenanigans that really shouldn't be part of the game. I was really talking about things like the AI knowing where units are from farther away than the human could. Ok, sure, that is a different ability than what the human has, but the human can see the tile underneath when the AI can't (fog gazing).

There's nothing in the game or manual to indicate to the player that the AI uses a different set of rules, or that it plays a different game entirely. It's semantics at that point...for those of us who know these things already and agree to them it's more grey area...but to people who don't know about these things it's definitely cheating.

"Nothing" is a bit strong; I strongly suspected before I joined the forums. The thing is, I expect games to take a few shortcuts.

They're the same thing. Literally, the AI is a part of the game.

In a literal sense, of course they are. I know there isn't a magical fairy that really controls the computer players

But they are separate in a couple of ways: they don't "know" the same things and they don't have the same goals.

Technically, the AI was coded in BTS to follow a strategy for winning culture; this is the only VC it is actively designed to pursue, and only some of them will do it. However, you can still cheat at a game without actively trying to win it. If you need a human example of this look no further than hackers boosting in call of duty; boosting is certainly not something that one does to try to WIN the game...but hacks to make it easier are definitely cheating even if the cheats are not directly applied to victory...and they share the AI PAssal tendency to screw a player (human or not) at random.

Um, I don't know anything about Call of Duty. Looking up boosting in CoD, I see that it means two people killing each other, apparently to rack up kills? How are they hacking?

I suppose the analogy sort of fits: these guys are playing their own game, and just "happen" to be using the same map.

I agree that the whole Peacevassal mechanic is broken, though. That's the main cheat that bothers me.

That's what difficulty bonuses are for. We don't need hidden, difficulty-independent and esoteric things that the AI can do that the human can't due to game rules. I don't want to hear "but the AI neeeeeeeeds the cheats!". Warlords II (mentioned yet again because of just how mechanically superior it is to every civ game ever made, despite having less depth) could certainly still be challenging, or even close to impossible at times.

The difficulty bonuses also aren't published. Just because you agree to their terms doesn't mean that you know what they are

I was disappointed to find out that the AI never becomes smarter at higher levels. I understand that programming that sort of thing must be difficult, but as you pointed out, there are plenty of games that do adjust the AI according to difficulty.

Or, they could make a nice, simple system where they can detect trades between known civs only. I can't imagine that routine would be too difficult...I'm not convinced people can consistently [figure out who traded what to whom]. I'm betting it wouldn't be so easy to always tell who made what trade without seeing IBT.

That would've been nice. And I know that I don't usually realize it. It's also ridiculous that trading with worst enemies can invoke a bigger penalty than being at war, or that the AI's get so upset when you don't give into their frequent demands.

I guess that my whole argument can be summed up as: I agree that there's no good reason to not know what the rules of a game are if you're interested in learning them, but not knowing them doesn't constitute a cheat necessarily; different rules for different players are acceptable if everybody agrees to them; and besides the controls not working and the diplo minigame being completely obscure and broken, the other differences are fine

 

[rah]

I always get irritated that the AI can irrigate wherever it wants.

 

[Choggy]

Ignoring issues with the GUI (which is just crap design), it seems there are essentially two separate arguments going on here.

The first one concerns game balance. AI civs get certain abilities that human players don't have (cheap upgrades, ability for Naval units to 'see' a move ahead, not valuing strategic resources when trading with other AIs, tech rate bonuses in late eras, a bias towards stopping a human player winning over getting a victory condition). I have no issue with this, simply because these are designed to negate the one thing a human has over the AI - intelligence and strategic decision making. Simply put, humans are better at this game than the AI because we are smart. You may as well just play at Noble level if you want a level playing field, by playing at higher levels you are giving the AI bonuses anyway to make the game more of a challenge.

The second one is the issue of automation - where you turn round to the AI and say 'Play this element of the game for me'. Autotmated workers, automated build of cities, etc. The fact that the if you hand over control to the AI it behaves differently to how the standard AI behaves is just wrong, because it penalises weaker players and makes the game less enjoyable because it forces you to micro-manage. Quite why Firaxis made this deciision is beyond me.

Regarding the diplo penalty when meeting new AI civs - I think that comes under the first category - humans are better at Diplo than the AI. An obvious example of this is when an AI civ with no religion converts to the first religion that gets founded in their civ. Experienced human players won't do that, they'll wait to see how the religion spread pans out and then make a tactical decision on what religion and when to convert. Whilst the -4 penalty is frustrating it also is a game balancer, however clumsy it may feel.

I agree that the whole Peacevassal mechanic is broken, though. That's the main cheat that bothers me

The whole vassal system smacks of a rushed implementation. It wasn't thought through very well at all. Why can't I demand techs of a vassal but not DoW him if he refuses? Same with the AP - you can lose a game purely through the RNG spreading the AP religion to your civ. Having played Vanilla extensively before moving to BtS last year, whilst Vanilla is more limited it's also better written and not so....flakey.

As a final note, my main gripe with the game is that the AI is simply not good enough in general. I've never wanted to learn how to play at Immortal or Deity level (although I do respect the skill of those that do) for the same reason I don't enjoy playing chess against somebody where I have to give up both my rooks to even the game out - there's only one strategy in those kind of games and that is to claw back the deficit as soon as I can and then win. Even playing at Monarch/Emperor level annoys me a little bit, I only do so because winning every game at Noble in the Renaissance level gets boring very quickly.

I haven't tried K-Mod yet but will probably do so soon. If it's half as good as people say it is then I'll never go back to standard BtS.

 

[TheMeInTeam]

But they are separate in a couple of ways: they don't "know" the same things and they don't have the same goals.

It would be nice if the players in a game played it.

Um, I don't know anything about Call of Duty. Looking up boosting in CoD, I see that it means two people killing each other, apparently to rack up kills? How are they hacking?

Combine it with wall hacks or aim bots and there you go. Hours of "fun".

That would've been nice. And I know that I don't usually realize it. It's also ridiculous that trading with worst enemies can invoke a bigger penalty than being at war, or that the AI's get so upset when you don't give into their frequent demands.

You can become a worst enemy on no other demerit than trading with a worst enemy.

 

[Um the Muse]

It would be nice if the players in a game played it.

I agree completely. You know what's even worse? When you realize that the AI beat me without even trying. No, seriously, that is a big drawback to this game.

Combine it with wall hacks or aim bots and there you go. Hours of "fun".

Glad I never got into it then. Not that I have the tactical awareness that I'd need to play anyway, but at least this gives me a good excuse

You can become a worst enemy on no other demerit than trading with a worst enemy.

Yep. I already said that I hated all the cheaty aspects of civ4 "diplomacy". The AI is definitely colluding against you, and you have to bend over backwards to get them to break their cartel. And even then you can get screwed because somebody rolled the dice and you get shafted.

I think that Choggy framed it well when he said that some cheats are good for game balance (that's not to say that they couldn't have been better designed, but it's understandable that the developers would rather not do all that work when they've got a reasonably good shortcut quickly) and some are just shameful.

Anyway, I forgot the original questions, so I hope you don't mind that I don't have anything useful to add.

 

[coanda]

AI collusion against you honestly is one of the few situations where the AI does show signs of playing to win. If you win 75% of the games you play, with 8 players in each game, the other players should darn well collude to block you out if they want to win. Conversely, if you win just 5% of games, they should be willing to cut you a great deal of slack. But it's almost unheard of for a human to play at a difficulty where they win less than the average (12.5% with 8 players), so AIs ought to be treating you as especially dangerous if they actually are trying to win.

That said - it's infuriating to play against. It's infuriating when humans collude against you from the start simply because they think you're a better player, and equally infuriating when it's AIs doing it. I can't say I mind the deliberate choice the devs made to make the AI a bit "stupider."

But the diplomacy is generally pretty bad even beyond that. It's full of factors that shouldn't matter but do, and things that should matter but don't. It's pretty transparent because you get to see the diplo. modifiers ("-1: You made an arrogant demand!"), but it's so stupid that it feels like manipulating a set of rules, not building a relation with another player.

 

[TheMeInTeam]

The vassal diplo averaging is probably one of the more horrid things in terms of diplo itself. Why they thought that was a reasonably ideal is beyond me.

 

[Manco Capac]

^
So true! It literally kills U.N. flexibility! Honestly, I really don't know what stopped HoF, XOTM, SGOTM...and foremost BUG to allow such tiniest modification that is so broken. The AI's doesn't even suffer from this feature and once you get one vassal, better to roll one AI after the others as the U.N. gets a lil harder, especially if there is a big heterogeneity of peaceweights and religions.
Because manipulating an AI to be liked is not always cakewalk!
Especially if that idiot likes to trade with our friends' worst enemies!

Damn you Willem!

 

[Pangaea]

Don't think it's been linked to yet, but this table has a good overview of human and AI bonuses, research rate, and so forth. Not sure if this should really go under cheating, as it's how difficulties are managed, as opposed to smarter or dumber AI, but it does feel a bit unfair that the AI can build and research much faster than us, and have smaller costs for unit supply, civics, etc.

 

[Um the Muse]

^ Thanks for that chart. I've seen it before, but then I couldn't find it again. Good stuff. It's also proof positive that AI gets certain boosts regardless of difficulty.

I never realized that they get a boost to their growth on higher difficulties. Also, does anybody know anything about that iAIPerEraModifier? Does that affect all aspects of their economy?

 

[Pangaea]

It's a good table, though I'm not sure what everything means; such as iStartingLocPercent and the one you mentioned, iAIPerEraModifier. One thing that surprised me is that already at Warlord difficulty the AI has a relative advantage over the human in fighting barbs (25% vs 20%).

Ran a test game while watching the football, with automated workers and complete governor control from the start of the game. I have problems believing the AI uses the exact same "AI" (if that makes sense) as automation for the human. I'm far from an expert, but some of the stuff the workers and gov'nors did were shocking.

I'll spoiler some picture below. But what isn't apparent in them, is that without exception the workers first built roads to resources and then built the improvement. And once the Wheel was in they started building roads pretty much wherever they happened to be right then. Took forever to connect the cities too. Although Barcelona was founded early, the 2 and then 3 workers stuck to the capital for a long, long time.

Some of the cities are crap btw, but don't mind that, nor the lack of whips as I mostly couldn't be bothered much with it.

Spoiler :

Here the city governor prefer to work an unimproved tile instead of the 5-hammer copper, while building a barracks. City grows a little faster of course, but production is much worse.

An overview of the whole empire in 600 BC, numbered in the order in which the workers improved tiles. I unticked the two workers related options, and you can see that on a couple of occasions they first built one improvement and then changed it when new techs were online. The forests marked with Chop but that are still standing, are pre-chopped.

The four cities at 600 BC. Lots of unimproved tiles.

Of course possible the AI civs also have this terrible workers and city management, would be a little weird to have two different sections of code after all. However.... It just looks so bad!!

 

[Um the Muse]

IIRC, iStartingLocPercent governs the odds that the AI gets a "better" start than the human player. The map generator struggles with evaluating the strength of a start, though, so the odds that their start is better has a lot of leeway.

AI workers build roads everywhere they go, too. In fact, it's part of what makes worker stealing so profitable (as long as the AI doesn't squish you with their archers afterwards). You can count on another one to follow the captured worker once the city deems it safe and then turn off its survival instinct.

Cows are better tiles than copper, though it was definitely a fail that the worker didn't improve that tile first. I'm pretty sure that the citizen working the forest tile would switch to the copper once you hit size four. Since grass hill mines are more efficient (two GHM gives you six hammers per two food, whereas the copper mine on tundra gives you five hammers per two food) in the long run.

That's interesting about the pre-chop. I didn't know that automated workers knew how to do that.

Edit: I thought that I should add that according to that chart that you linked to, the AI gets to build worker improvements faster than the human on any difficulty above Noble (even Prince, which I thought just gave the AI a bonus to their science).