Peaceweight Cooking

[Swordnboard]

I was reading Tachy's old thread explaining various mechanics, and stumbled on an interesting fact about AI peaceweight. The value we understand to be a leader's peaceweight is actually variable, and can deviate significantly (plus or minus 3) from game to game, but remains constant within a given game. This has no impact on intrinsic AI-Human attitudes, but can alter AI-AI attitudes significantly, (up to about 3 points).

As it turns out, testing indicates this deviation from normal is also unchanged when regenerating the map. This means that, through cunning use of map-finder and a single save to collect data, we can hugely impact the AI-AI attitudes within our HOF games by cherry-picking a game with the correct mix of peaceweights.

Some ideas I've brainstormed for this technique:
-Reduce declared war on friend penalties with a target AI (for example, play with a version of Mansa who dislikes even the high peaceweight AI)
-Prevent peaceful lovefests from breaking out across the ocean when you are playing a continents-type mapscript, thus stunting AI development and helping you achieve key wonders or religions. This one is especially good for hemispheres since you can effectively choose which AI will be on your continent.
-Cause extra hatred between two AI so that you do not reach worst enemy status with either of them.

I'm sure there are more as well. This can get a little bit tedious for larger maps, especially since determining an AI's peaceweight requires a lot of data without using the world builder, but can pay dividends for longer games and games which will be played many times with the same kind of map. On the other hand, this technique doesn't seem to really be in the spirit of the game since you are getting "free" information, but I guess it's sort of like handpicking leaders or checking early turn demographics screens until a favorable result is achieved. Curious to hear anyone's thoughts as far as additional ideas and uses or whether this technique feels unethical to use. Also, does anyone know of any other random factors which may not be re-randomized on regeneration?

 

[Noble Zarkon]

Interesting stuff - can you link to that thread please?

 

[Swordnboard]

Here's the particular post the original tip came from:
My personal thread: Clarification of certain game mechanics

Also, I think this trick could also work with random personalities, but haven't tested it yet. Entire personalities are a little harder to figure out than peaceweight.

 

[Kaitzilla]

I was reading Tachy's old thread explaining various mechanics, and stumbled on an interesting fact about AI peaceweight. The value we understand to be a leader's peaceweight is actually variable, and can deviate significantly (plus or minus 3) from game to game, but remains constant within a given game. This has no impact on intrinsic AI-Human attitudes, but can alter AI-AI attitudes significantly, (up to about 3 points).

As it turns out, testing indicates this deviation from normal is also unchanged when regenerating the map. This means that, through cunning use of map-finder and a single save to collect data, we can hugely impact the AI-AI attitudes within our HOF games by cherry-picking a game with the correct mix of peaceweights.

Some ideas I've brainstormed for this technique:
-Reduce declared war on friend penalties with a target AI (for example, play with a version of Mansa who dislikes even the high peaceweight AI)
-Prevent peaceful lovefests from breaking out across the ocean when you are playing a continents-type mapscript, thus stunting AI development and helping you achieve key wonders or religions. This one is especially good for hemispheres since you can effectively choose which AI will be on your continent.
-Cause extra hatred between two AI so that you do not reach worst enemy status with either of them.

I'm sure there are more as well. This can get a little bit tedious for larger maps, especially since determining an AI's peaceweight requires a lot of data without using the world builder, but can pay dividends for longer games and games which will be played many times with the same kind of map. On the other hand, this technique doesn't seem to really be in the spirit of the game since you are getting "free" information, but I guess it's sort of like handpicking leaders or checking early turn demographics screens until a favorable result is achieved. Curious to hear anyone's thoughts as far as additional ideas and uses or whether this technique feels unethical to use. Also, does anyone know of any other random factors which may not be re-randomized on regeneration?

Oh wow, that's crazy.
Having a Mansa that hated everyone would be great!

Might have to change HOF games to have "Random Seed on Reload" perhaps?
Maybe that would fix cherry picking.

 

[elitetroops]

Also, I think this trick could also work with random personalities, but haven't tested it yet.

Yes, it works with random personalities. They are also kept when regenerating map.

To some extent it can work to pick opponents on your landmass on a continents map. I've noticed when regenerating manually for NC games that you tend to mostly have the same opponents on your landmass after regeneration, but this is not 100% accurate. Number of players on the continent can vary, and sometimes out of the blue the distribution of AI between the continents can be completely different.