Observations from Spying

[JerichoHill]

Prince Level, Terra Map, Epic Speed

I experimented in this game by planting a spy in each civ's capital, plus one on the other continent to understand barbarian activity.

The following is a brief summary.

1) Hatty loves to switch to producing research after the territory expansion phase is over with, and she has enough units to defend her territory. Mansa Musa does the same thing.

2) Bismarck tends to build more military than the average civ. Obviously Monty and Toko build the most.

3) No Civ prioritizes Wonder building appropriately. Cyrus does the best at this.

4) Tech choice path appears to be random.

5) Barbarians research techs.

6) Barbarians always attack with the countering units to a city's defensive force. They never use (or rarely) units of the same type, unless its early on when counters are rarer.
They do not build siege weapons.

7) Tokogowa will kamikaze if you're close to beating him. Charge of the Light Brigade.

8) The AI almost never has enough defenders in their non-important cities.

9) AI rule of thumb is to improve terrain seemingly for growth, except it almost always cottages plains.

 

[petertr2000]

Good post mate

 

[emills]

Though spying, not only will Monty always enough forces to invade, he will also get himself into a production trap using Hereditary Rule (i.e. he grows to his military happiness but then get massive unhappiness when switches out).

 

[JerichoHill]

Yes, I suppose thats why his cities had unhappiness.

I should mention that Hatty and Cyrus both grew their cities to unhealthy levels, and Hatty WILL STARVE her city rather than lose production on a tile!

 

[Toshiro126]

Very useful post. Thanks for the info! I have a feeling I haven't been making as much use of my spies as I should be...

 

[Viper Daimao]

perhaps it would be useful (if tedious) to make a mod that allows you to have more than 3 spys and put spys in all the AI cities to see how they act.

 

[JerichoHill]

4 spies is the limit, at least for that game.

ill do another game next week with different ai and take very detailed notes

 

[ToV]

Thanks for the information. This can be very helpful, perhaps vital, to know. Entire strategies will be based on your information. I too noticed Motezuma's massive armies> I used a spy to raid his plans, just before a war. At least too of his cities were stacked with millitary units, and I needed a little help in that war.

It's nice to now that the AI doesn't prioritize wonders. Explains my consistent wonder-building sprees on Noble. Some of these quirks have exploitive potential as well.

 

[Yzman]

From a breakdown of the leader stats that someone actually posted here, it looks like Louis IV is really the biggest wonder builder when it comes to AI. Montezuma is the lowest.

 

[atreas]

From a breakdown of the leader stats that someone actually posted here, it looks like Louis IV is really the biggest wonder builder when it comes to AI. Montezuma is the lowest.

Correct. It's the iWonderConstructRand property in the XML files, and the best (and worse) are
Good Builders: LOUIS_XIV, ASOKA, CYRUS, HATSHEPSUT
Bad Builders: MONTEZUMA (with 0 value), GANDHI, GENGHIS_KHAN MAO_ZEDONG, NAPOLEON, SALADIN

The number of units is also reflected in the XML file, in the property iBuildUnitProb. First is (surprise) NAPOLEON followed by ALEXANDER, GENGHIS_KHAN, and MONTEZUMA. On the other side, the less army should be expected (of course) on GANDHI, followed by ASOKA, ELIZABETH, FREDERICK, HATSHEPSUT, QIN_SHI_HUANG, and FRANKLIN_ROOSEVELT. The ratio of army between NAPOLEON and GHANDI is 40/15 (i.e. almost 3 times more units).