(2-VT) AI golden age points from difficulty replaced with GA cost reduction

[Legen]

Proposal:

The AI no longer earns Golden Age Points from AI difficulty triggers. Instead, the AI gains a Golden Age cost reduction based on difficulty.

Rationale:

The main aim of this proposal is to address a balance issue regarding AI Brazil.

At the moment, AI players gain Golden Age Points on city founding and on most types of historical events, as detailed in the wiki and in Recursive's proposal last month. For reference, here's how much an AI player gains based on difficulty and era as of patch 2.7.3:

​	settler​	chieftain​	warlord​	prince​	king​	emperor​	immortal​	deity​
ancient / classical​	54​	57.5​	61​	64.5​	69​	72.5​	76​	81.5​
medieval​	92​	101​	110​	119​	130​	139​	148​	161.5​
renaissance​	144​	161.5​	179​	196.5​	217​	234.5​	252​	277.5​
industrial​	210​	239​	268​	297​	330​	359​	388​	429.5​
modern​	290​	333.5​	377​	420.5​	469​	512.5​	556​	617.5​
atomic​	384​	445​	506​	567​	634​	695​	756​	841.5​
information​	492​	573.5​	655​	736.5​	825​	906.5​	988​	1101.5​

And the GAP triggers are as follows:

• Enters a new era (3x normal bonus; Food, Production, Gold, Golden Age Points, Science, Culture)
• Founds its original capital (Gold, Golden Age Points)
• Founds a new city, other than its capital (Food, Production, Gold, Golden Age Points, Science, Culture)
• Wins a war (warscore 25+) (Food, Production, Gold, Golden Age Points, Science, Culture)
• Constructs a World Wonder (Gold, Golden Age Points)
• Generates a Great Person (Gold, Golden Age Points)

This creates a considerable imbalance in favor of AI Brazil. This civ's UA converts GAP into Gold and Tourism, currently at a 30% efficiency, making these AI procs far more powerful for AI Brazil than for any other AI civ. It is as if AI Brazil had an Arabia-like UA on top of Brazil's usual capabilities. As such, you are bound to face AI Brazil as a noticeably more powerful civ than usual among AIs, and having a stronger performance relative to a human Brazil player.

Replacing the GAP AI difficulty procs with a GA cost reduction addresses this issue. Brazil's GAP conversion in the UA doesn't benefit from a GA cost reduction effect. This should put AI Brazil's performance much closer to how the civ performs in human hands, and to other AIs as well. And the discussion of Brazil's balance should be much closer between human based and AI based feedback.

There are some other expected observable effects on the AI. Their golden ages should stop proccing unusually more often in earlier eras relative to a human player, and their GA patterns should more closely resemble the GA patterns of human players. Also, this is likely to prevent other GA civs to be able to use their GA bonuses in the early game when in the hands of an AI. For instance, AI Ethiopia would not benefit from Stele's +25% faith during GAs in Ancient era far more frequently than a human Ethiopia does.

 

[ma_kuh]

This makes me curious about whether giving these GAP triggers to human players would solve the problem as well, with a buff to the player rather than a nerf to AI. Then Brazil could be balanced around the GAP procs. I'm not sure I feel strongly either way.

Other notable AI nerf is Persia, who can have some scary early timings, presumably as a side-effect of these triggers.

 

[Recursive]

There will be a built-in option for this in the next version, along with numerous other difficulty-related options.

Edit: It's AIGoldenAgePercent.

 

[Recursive]

Proposal vetoed.

Reason: Proposal doesn't list specific values. How much should the cost be lowered by on each difficulty level?