Civ4 AI Survivor 2: The AI-stralian Outback

[LogicalTautology]

But is that actually a departure from normal AI behaviour, though, T-Hawk? My understanding of Sullla's post is that while it may appear as though the AIs are putting points in Espionage to counteract the player, they actually just run a flat 20% Espionage regardless of the Great Spies (as evidenced by continued spending late into the game).

[The Black Sword]

That doesn't seem right to me T-Hawk. Both parties lose 20% of their total commerce. The financial civ has more total commerce, so they throw away more absolute resources into the EP black hole. It seems to favour food-hammer strategies over commerce heavy ones; instead of a river cottage being a 2/0/2 tile, it's only a 2/0/1.6, makes a farm look even better.

[LogicalTautology]

No, T-Hawk is correct that spending 20% into Espionage hurts Civs that pay a higher percentage of their commerce as expenses (those running a lower Science slider). To give a more extreme example, assume we have two Civs: one is running a OCC and so has low total commerce but virtually no expenses and the other has a huge sprawling empire with an enormous amount of total commerce but spends 80% on expenses. Without Espionage, Civ B out-researches Civ A assuming it has 5x the total commerce. With Espionage, Civ B cannot research at all as it spends 80% on Expenses (gold) and 20% on Espionage, leaving 0% for Science.

[T-hawk]

I'm not sure if the AIs always run 20% espionage. The problem is observer bias, that the observation has a requirement that changes the AI behavior. We may always see them running 20%... because we can see that only when we put EP against them! Do they run 20% when we don't have the EP to see it? We don't know.

The right solution is to patch the display screens to see the AI activity without requiring EP from the player.

[LogicalTautology]

From what I've seen in watching gameplay of people that play Civ4 at a much higher level than I do, the AI does like to spend heavily on Espionage even in a 'normal' game where the player invests minimally into Espionage. My hunch would be that the behaviour is coded as to always run roughly (or exactly) 20% Espionage. The one thing that does change is very early teching rate, since unless I am mistaken they would not normally be able to spend on Espionage until they meet their first opponent (although I assume that happens relatively quickly in this game, what with the multitude of free starting units).

Anyways, I certainly seem to have opened a much bigger can of worms than I expected by pointing out a comment I saw in the stream chat! I hope everyone can still enjoy what is sure to be an awesome tournament

[The Black Sword]

No, T-Hawk is correct that spending 20% into Espionage hurts Civs that pay a higher percentage of their commerce as expenses (those running a lower Science slider).

Nope, it hurts the civ with more total commerce, which is consistent with your example.

Consider two civs of equal size, who can choose between investing in hammers or commerce freely. For simplicity lets assume 1h=2c is the correct exchange rate and each civ has 50h or 100c to invest. Both civs pay 20c maintenance.

Civ A goes with all commerce: 20% goes gold to cover maintenance, 60% research and 20% into espionage.

Civ B decides to to go with 50g and 25h. He has to run the slider at 40% gold to cover expenses, 40% research and 20% espionage.

Civ A has thrown away 20g that turn and Civ B has only thrown away 10g.

So, AI leaders who choose commerce investments over food-hammer ones would be at a disadvantaged compared to the base game. Additionally, the value of the traits would be changed, the most obvious two examples are that Financial is only at 80% strength and Phi, which can give a lot of beakers outside the slider, is stronger.

EDIT: Bit of a crosspost. This is all assuming that the AI has it's espionage slider at 0% without lurker interference, I have no idea if that's true one way or another.

[Dreylin]

The right solution is to patch the display screens to see the AI activity without requiring EP from the player.

FFH (& EitB) have no EPs built on the back of BTS and show graphs on meeting, so Qgqqqqq or one of the other guys modding that may be able to help out with an easy solution.

[Jabah]

From what I could see, there seems to (still) be some/all goody huts. Was that on purpose or shouldn't you remove them quickly before starting.
[can't remember if the AI will/could pop unfair results at this level]

Thanks for doing that again (and live on twitch to make it even better).

Jabah

[LogicalTautology]

No, T-Hawk is correct that spending 20% into Espionage hurts Civs that pay a higher percentage of their commerce as expenses (those running a lower Science slider).

Nope, it hurts the civ with more total commerce, which is consistent with your example.

Consider two civs of equal size, who can choose between investing in hammers or commerce freely. For simplicity lets assume 1h=2c is the correct exchange rate and each civ has 50h or 100c to invest. Both civs pay 20c maintenance.

Civ A goes with all commerce: 20% goes gold to cover maintenance, 60% research and 20% into espionage.

Civ B decides to to go with 50g and 25h. He has to run the slider at 40% gold to cover expenses, 40% research and 20% espionage.

Civ A has thrown away 20g that turn and Civ B has only thrown away 10g.

So, AI leaders who choose commerce investments over food-hammer ones would be at a disadvantaged compared to the base game. Additionally, the value of the traits would be changed, the most obvious two examples are that Financial is only at 80% strength and Phi, which can give a lot of beakers outside the slider, is stronger.

EDIT: Bit of a crosspost. This is all assuming that the AI has it's espionage slider at 0% without lurker interference, I have no idea if that's true one way or another.

No, sorry. It hurts the team that runs the highest Gold %.

Civ A has X Commerce, and runs 40% Gold. Normally they would get 60% Science, but instead they get 40%. They lost 33% of their research output this turn, regardless of how high X is.
Civ B has Y Commerce, and runs 20% Gold. Normally they would get 80% Science, but instead they get 60%. They only lost 25% of their research output this turn, regardless of how high Y is.

The thing you're overlooking is that it doesn't matter how many beakers get lost to Espionage, it's what percentage of their beakers that get lost.

[T-hawk]

The thing you're overlooking is that it doesn't matter how many beakers get lost to Espionage, it's what percentage of their beakers that get lost.

Both do matter (you're both right), but I think the percentage is the more relevant number. It's all relative. A straight number of beakers has wildly different meaning throughout the game in say Agriculture versus Fusion, but a 20% percentage always means it takes 8 turns instead of 10. An elapsed turn is the fundamental input to all things economy in Civ.

But even beyond that, even disregarding which effect is more important, the point is that AIs spending EP changes the balance of the game's economic model, in different ways for different leaders. Which leader gets the worse impact is debatable, but that leaders are impacted differently is not. Anything that causes the AIs to deviate from what should be their standard gameplay is not seeing their true personalities and might as well be a mod.