Exact spying figures (success rates, etc)
- Thread starter Question
- Start date
whoward69
DLL Minion
CvEspionageClasses.cpp
The released source code is distributed as part of the Civ V SDK, which is free on Steam (go to your Steam Library, and select Tools, rather than Software).
Magma_Dragoon
Reploid
- Joined
- May 10, 2008
- Messages
- 2,354
I don't think I went over counter espionage, but coups, rigging and tech stealing are discussed
http://forums.civfanatics.com/showthread.php?t=509665
http://forums.civfanatics.com/showthread.php?t=509665
Ah that helps a lot, thanks.
Im guessing that since the file isnt a xml file, that its not editable? Or is it?
Does anyone know how the counter spy chances are calculated, as it seems shockingly high. I couldnt get a level 3 spy to steal anything successfully without getting save scumming more than a dozen times.
Also since its based on science output of one city only, that explains why it takes 100+ turns to steal tech in later eras....since later techs assume you have plenty of cities to research with. Badly balanced overall.
Its also a dissapointment that spy rank has little to no effect on CS rigging...
Im guessing that since the file isnt a xml file, that its not editable? Or is it?
Does anyone know how the counter spy chances are calculated, as it seems shockingly high. I couldnt get a level 3 spy to steal anything successfully without getting save scumming more than a dozen times.
Also since its based on science output of one city only, that explains why it takes 100+ turns to steal tech in later eras....since later techs assume you have plenty of cities to research with. Badly balanced overall.
Its also a dissapointment that spy rank has little to no effect on CS rigging...
Do you have the figures for that? If there is no competition, CS vote rigging is guaranteed, so no need for a high ranking spy. When a CS is contested though, it certainly seems like my experienced spies do very well (not 100% though) while lower ranking spies are only 50/50 or so. Also, spies in CS are suppose to help thwart coups, but it's is hardly a guarantee, and for this particular task, high ranking spies don't seem to me to be much better than raw recruits. I would appreciate knowing the numbers!
Correct. They are just more likely to rig the election in the first place.
Magma_Dragoon
Reploid
- Joined
- May 10, 2008
- Messages
- 2,354
To catch a thief
(random number between 0-299 + 30*spy rank) * (100+policy modifier catch spies) / 100
policy modifier catch spies is not used in the base game
spy ranks are 0, 1, 2 for recruit, agent, special agent
if the final number is greater than 199 the spy is killed
if the final number is greater than 99 the spy is identified
a special agent will almost always know who did it, and has a 19/30 chance to kill a spy
a recruit depends entirely on Jon Shafer to kill an enemy spy
(random number between 0-299 + 30*spy rank) * (100+policy modifier catch spies) / 100
policy modifier catch spies is not used in the base game
spy ranks are 0, 1, 2 for recruit, agent, special agent
if the final number is greater than 199 the spy is killed
if the final number is greater than 99 the spy is identified
a special agent will almost always know who did it, and has a 19/30 chance to kill a spy
a recruit depends entirely on Jon Shafer to kill an enemy spy
I think something is missing. I am pretty sure I have gotten notifications about who stole my tech without a spy in my capital, am I mistaken about that?
If I read your formula correctly, I think it is only 17/30 for Level 3 counter spy to kill. I put it all in table, rounding percentages:
|none|Level 1|Level 2|Level 3
Chance to ID:|?|67%|77%|87%
Chance to kill|0|33%|43%|57%
If I read your formula correctly, I think it is only 17/30 for Level 3 counter spy to kill. I put it all in table, rounding percentages:
Chance to ID:|?|67%|77%|87%
Chance to kill|0|33%|43%|57%
Thats because the spy mod is only 30 or 60 depending on rank of defending spy. You still have a 2/3rd chance to identify who stole your tech and a 1/3rd chance to kill a spy. But im not sure if this is correct because i dont think you can kill spies with no spies at all, even though the formulae should give you a 1/3rd chance to do it.
Magma_Dragoon
Reploid
- Joined
- May 10, 2008
- Messages
- 2,354
Yeah, you can identify a spy without a counter agent, but the threshold for identification is the same
So a raw spy helps not at all with identification?
Where is the code that lets an empty city ID but not kill?
reddishrecue
Some dude on civfans
- Joined
- Nov 16, 2009
- Messages
- 6,213
An 85% rating of coup success, which is pretty good, still somehow fails often for some reason.
Perhaps 15% of the time, on average?
Magma_Dragoon
Reploid
- Joined
- May 10, 2008
- Messages
- 2,354
CvEspianogeClasses.cpp lines 385 to about 485
I think the coup success rating capping at 85% is pretty lame. In my experience, you only get 85% with a level 3 spy and almost near-ally influence anyway, so instead of risking your spy, you could just throw the CS some gold, maybe wait a few turns for a election, and become the ally that way, which is much safer.
Sorry, I was really asking if you would excerpt the relevant bits and explain why no spy == change to catch, but not a chance to kill. Also please confirm that a raw recruit adds a chance to kill, but does not improve chance to ID. Certainly, with no spy at all, it seems like tech is stolen anonymously at least half the time, but your formula says it should be only a third of the time?
Similar threads
