Why did Cathy let me win on Deity?

[Noian]

So I just beat my first Deity game right after beating my first Immortal game (kinda crazy since like a month ago I found King too hard) but it felt so wrong. It was also my first game as Venice but the huge 12 city Russian Empire always managed to get like 100gpt more than me. By like 1825 or so Cathy had the biggest army by far, was 10% ahead of me techwise (I was second place), had like 2700 score while I barely had 1000 and also over 40k gold stockpiled. And yet she let all the city states be my allies and when the vote for world leader came she just didn't care. Sure, we were best friends all game long, but why did she let me win so easy? Is the AI always like this? Deity felt not challenging.

 

[wydon]

I've seen a lot of posters turn Diplo. victory off for that very reason.

 

[EK834]

It's what is the most irritating with Diplo Victory.

The best you can hope for is that a few AI civ will target a single City-State (or 2, and always the same ones which makes it easy to anticipate them) and try to get it away from you, but they'll start too early and usually give up turns before the vote.

I had a game where Casimir was probably 20 turns away from a science victory, with an Empire of 60 cities, a 10,000 gold treasury and 5 CS that were costing me a fortune to keep as he gained influence with them through gunboat diplomacy (I was up to 545 Influence with one of those in my struggle to keep it from him).. Austria was also way ahead of me in score with perhaps 30 cities, and she was sitting on 13,000 gold. She did interfere with my alliances with CS, actually managed to prevent my first attempt at a diplo victory by taking three on the last turn before the vote, and I had high hopes it wasn't a fluke (but apparently it was... at least 2 were quests fulfillment).

On the second session for WL she did took 2 allies from me after the 5 turn notice (not sure if she bribed or not that time), but after I took both back she did nothing at all. With all the CS allied (a good third had been puppetted by me, wed by Austria or conquered by others) and Globalization (with all my spies as diplomats), 2 from World Ideology plus 2 votes from a previous failure at WL I still had just one vote over the bare minimum to win, and fearing Maria Theresa would buy a few my allies in the last turns I expended 3 MoV to get a 18,000 gold reserve, ready to spend it all as bribe on the 10 CS on the last turn, and a few turns before the vote I declared war on Casimir to liberate an easy to take Byzantine city to bring back Theodora and get her 6 delegates... but I shouldn't have bothered. The AI did nothing to interfere, it just let me win. 4 CS taken from me and I was doomed, but it did nothing. Very disappointing.

I don't know if it's because they don't "understand" they can block a diplo win by last minute bribes to CS, or if there's something broken the programming that interferes and makes them keep their cash for something else even though preventing a win should have been the priority (Austria seemed to have the 3rd level tenet to buy spaceship parts), but this is really annoying.

It's too bad because I re-installed the patch midway into that game and the AI's performance in science increased spectacularly. I was 6 techs ahead of Casimir and Austria, but post patch they built science buildings and soon the best I could do was parity with Austria while Casimir ran away (he was 9 techs ahead by my diplo win). So very wide civs can again be very dangerous for science (I hadn't seen that on King for a long while). His tourism was impressive. The only way I could hope to win culturally would be the Internet and that the AI blocked by beelining to get the Great Firewall first. My last hope was diplo but the AI had the means to stop that too. That game was thus rightly lost, but the AI let me win.

 

[Polisurgist]

I don't know if it's because they don't "understand" they can block a diplo win by last minute bribes to CS, or if there's something broken the programming that interferes and makes them keep their cash for something else even though preventing a win should have been the priority (Austria seemed to have the 3rd level tenet to buy spaceship parts), but this is really annoying.

It seems to me that the AI does what it can to maximize its own ability to win the victory condition it's chosen for itself as quickly as possible, but it doesn't really take into account how close anyone else is to winning. So you'll have that diplo victory blocked if you're playing against Greece or Siam, but if you're playing against AIs who aren't prioritizing diplomatic relations, they aren't going to stand in your way...they're thinking 50 turns down the road, that they'll want that money in case they get invaded and need to buy up troops, even if you're 10 turns from winning.

 

[KrikkitTwo]

It seems to me that the AI does what it can to maximize its own ability to win the victory condition it's chosen for itself as quickly as possible, but it doesn't really take into account how close anyone else is to winning. So you'll have that diplo victory blocked if you're playing against Greece or Siam, but if you're playing against AIs who aren't prioritizing diplomatic relations, they aren't going to stand in your way...they're thinking 50 turns down the road, that they'll want that money in case they get invaded and need to buy up troops, even if you're 10 turns from winning.

I think that is because of the Victory conditions

1. Domination.. all players will naturally try to stop a domination win... because they try to avoid losing their own capital

2. Science.. players really can't stop a Science win except by winning some other way first (what they are already trying to do)

3. Culture..players can't do too much to avoid a cultural win (except cutting off Open borders/declaring war).. although it seems like they may have fixed that (ie AI realizes effect Open Borders have on cultural domination)

4.... Diplomacy, here blocking is Easy (unlike Science+Culture.. all you need is gold) but Unnecessary (unlike Domination.. losing CS allies is not as bad as losing your capital)

It probably needs to just increase the desire of AIs to steal a CS from the delegate leader if the delegate leader is anywhere near close.

 

[EK834]

I think that is because of the Victory conditions
It probably needs to just increase the desire of AIs to steal a CS from the delegate leader if the delegate leader is anywhere near close.

Yup, increase the likelihood they'll spend big to steal CS if someone can win a DP. As you said, they stay focused on their own strategy to win, oblivious to the fact if they don't act to block the game was over in a few turns.

That said, pre-patch I had a few BNW diplomatic struggles against Alex and even him didn't bribe as he should have. The challenge was to generate enough gold to keep up with his growing influence, but despite having a huge treasury sometimes, he still didn't block me.

In one game G. Khan at least started attacking and conquering my CS allies when the session for WL was announced. In another Alexander launched a whole lot of units to try to conquer my capital 10 turns before the vote. But most games they just do nothing.

3. Culture..players can't do too much to avoid a cultural win (except cutting off Open borders/declaring war).. although it seems like they may have fixed that (ie AI realizes effect Open Borders have on cultural domination)

Yeah, it was just fine in my first post-patch game. Much fewer AI agree to OB in the late game, unless you're already influential. Casimir started building the Great Firewall in his best city the turn after I built my NVC and before I could get my Internet up, blocking my Cultural victory and forcing me to fall back on a diplo win.

 

[Noian]

Ah well, I hope they get to fixing this AI carelessness as soon as possible, doesn't sound hard to do. It's beyond me why they didn't already. I love DV but don't want it to be easy .

 

[redwings1340]

Yeah, basically what other people have said. My first several BNW games were diplo wins until I just realized it was way too easy, even on deity (assuming you don't die before the renaissance). The computer doesn't care if you win, so it won't stop you. Sometimes they'll randomly ally a few city states before the vote, and it will look like they tried to stop you, but that's just their normal city state focus happening at a lucky time for them.

Turning off diplo victory doesn't turn off the world congress or anything though, and provides another round of normal voting every time diplo victory comes up, which can make the world congress exceptionally powerful late game if you play it right. I like playing that way, since I love the world congress, and if you can't do anything other than propose natural heritage sites, you're either way too afraid of bad diplo penalties (if you don't have a military and have bad relations with your neighbors this can happen), or you're not really fully realizing how strong this thing is.

 

[EK834]

Yeah, basically what other people have said. My first several BNW games were diplo wins until I just realized it was way too easy, even on deity (assuming you don't die before the renaissance). The computer doesn't care if you win, so it won't stop you. Sometimes they'll randomly ally a few city states before the vote, and it will look like they tried to stop you, but that's just their normal city state focus happening at a lucky time for them.

Turning off diplo victory doesn't turn off the world congress or anything though, and provides another round of normal voting every time diplo victory comes up, which can make the world congress exceptionally powerful late game if you play it right. I like playing that way, since I love the world congress, and if you can't do anything other than propose natural heritage sites, you're either way too afraid of bad diplo penalties (if you don't have a military and have bad relations with your neighbors this can happen), or you're not really fully realizing how strong this thing is.

The downside of that is that it handicaps a few civs like Siam and Greece (and Autocrats who use gunboat diplomacy). It's not as unbalanced as turning off Culture or Science though, as the AI don't seem to really work for diplomatic victory (I've never seen one focus to get Globalization early enough or bring back dead Civs to get their delegates, for instance. If they can't win culture or Domination, they all seem to focus on the lower tree for a SV, occasionally getting the Great Firewall to block/delay the human from a C. Vic).

I don't mind so much that they don't work for a diplo win (because their efforts would definitely make them even less competitive for Science and Domination victories), but it would be great if they were at least aware that:

- their massive treasury could be spent on CS in the last few turns to block the human from a diplo win
- if the human brings back from the dead some Civs, they should declare war to reconquer them or the human just got 4-6 extra delegates
- If Marrying/puppetting/conquering enough CS will deprive the human of enough delegates for a win, go for it.

Of course I would prefer it if they changed the system so that the vote takes place between the three Civs with the most delegates. The candidates get chosen, and the vote takes place 30 turns later. At that special session, the AI that's not a candidate would give its delegates following various formulas taking ideology, friendships, cultural influence, military strength into account, also allowing for a percentage of the delegates to be bought in deals. CS would exceptionally provide only one delegate for those votes (they could thus give a big edge, but usually not an automatic victory).

The WC panel would tell you how many votes you have and why.

The three Ideology Wonders could provide extra delegates, a step in becoming "faction" leader.

Something like that, or a streamlined/simplified version, would make it more challenging and immersing. Winning by having the CS alone would be impossible unless many Civs have been eliminated by someone else by conquest. It would mean you pretty much need to forego chances of winning by Science, by going for Globalization for extra delegates, and the Internet for Influence. It would also reward the player who manage to keep or sway enough other Civs to his ideology before the WL session arrive.

 

[KrikkitTwo]

Ah well, I hope they get to fixing this AI carelessness as soon as possible, doesn't sound hard to do. It's beyond me why they didn't already. I love DV but don't want it to be easy .

Well dealing with gold spending on allies is a difficult thing, because if you get into a bidding war you can have some serious issues.

Especially because that gold could be used for other thing (buying spaceship Factories/Research agreements/etc.

However, If AIs that were NOT going for diplo victory and still had cash simply had a tendency to target the CS allies of the delegate leader (ie whichever allies were easiest for them) then it would be good

Or perhaps if a CS would only accept one gold gift every 5-10 turns (from Anyone... ie if Siam gives a gold gift, it won't accept one from Germany or India or Siam until 5-10 turns have passed)

 

[kaspergm]

The infuriating thing is that while some of the AI issues the game has are probably very hard to solve - for instance, coding the AI to use 1UPT and move units properly is probably nigh impossible, at least starting from how things look now - but coding fix for diplomatic win is REALLY easy. It's not like it comes to a surprise to anybody that the vote is happening on a certain turn, and it's very evident to everybody how many votes you control, so all the information the AI needs is available, not only through cheating, but actually available for any player. As such, it's incredible that they haven't put in a sanity check that makes AI buy out the required number of city states - or at least just a couple, and hope for the other AI's to do their part - when they have the gold.

 

[Gort]

To be honest, any AI with more than about 2000 gold should be blowing it all in a huge bidding war on city states. It makes no real sense to focus on one victory condition to the exclusion of all others - after all, having city state allies is perfectly useful for winning domination, tourism, or science victories, not just diplomatic ones.

The issue would then become that the AI has far too much gold on Deity for it to use it sensibly and a human to have much of a chance, so you'd have to bring that down, which is fine by me.

 

[Polisurgist]

I think there may be a more fundamental question involved here (not saying that the AI couldn't use an upgrade in this department, though): if you're playing a multiplayer game that only allows a single winner, how much does it benefit player A to prevent player B from winning if that means possibly losing to players C, D or E?

I was almost in this kind of situation in my last game: Rome was Influential on all but Siam and Greece; Siam and Greece were allied but in a bidding war for city-states, and both were topping the world leader vote every ten turns (but not winning). I was pushing for a science victory, but when it came to eliminating competition, my options were to remove one of the diplo players, leaving one I couldn't compete with economically, or to eliminate both, triggering a cultural victory for Rome. As a human, with at least some emotional investment in the game, my gut told me to go after the one closest to winning, but that wasn't going to help me win, so should I actually have bothered?

I'm not sure that an AI programmed to win would bother to stop you if it wasn't also going to win. An AI programmed to stop others from winning at all costs, would, but then you have the Vanilla situation, where you're pursuing a VC and the rest of the world is trying to murder you.

 

[KrikkitTwo]

I think there may be a more fundamental question involved here (not saying that the AI couldn't use an upgrade in this department, though): if you're playing a multiplayer game that only allows a single winner, how much does it benefit player A to prevent player B from winning if that means possibly losing to players C, D or E?

I was almost in this kind of situation in my last game: Rome was Influential on all but Siam and Greece; Siam and Greece were allied but in a bidding war for city-states, and both were topping the world leader vote every ten turns (but not winning). I was pushing for a science victory, but when it came to eliminating competition, my options were to remove one of the diplo players, leaving one I couldn't compete with economically, or to eliminate both, triggering a cultural victory for Rome. As a human, with at least some emotional investment in the game, my gut told me to go after the one closest to winning, but that wasn't going to help me win, so should I actually have bothered?

I'm not sure that an AI programmed to win would bother to stop you if it wasn't also going to win. An AI programmed to stop others from winning at all costs, would, but then you have the Vanilla situation, where you're pursuing a VC and the rest of the world is trying to murder you.

That's why I think the AI needs to focus on stopping an impending diplo win only.
1. only target the allies of a delegate leader near the time of the vote, and
2. only if it has a chance of winning itself (ie it has a decent stockpile of gold)

Finally, it should only target a small # of CS (if we assume 1:2 starting ratio, there is probably 1 surviving/could win civ for every 3-4 CS... a diplo victor will probably need 1/2 of them... so if the "major players" all took the 1-3 CS that were easiest for them to steal, that would probably stop a diplo victory)

I agree the AI shouldn't stop others from winning at all costs... but it should stop others from winning IF it has a chance of winning.
[it shouldn't be that hard to have a "winning check"... if nothing else is it in the top 1/2 of original players in score... then it might be big enough to spend some money delaying your win while working on its own.

It would be nice for Time victory to be something that occasionally popped up because the major powers each had good enough military/cultural and "diplomatic sphere of influence" Defenses so that no one could win any of those 3 (then just make the techs a little bit more expensive/ stop the 'saving to bulb later' of GS to delay science victory a little more.)

 

[Noian]

This got me wondering, isn't there any mod out there that improves AI behaviour in this sector? If there is, can anyone tell me what it's called ?