Best Way To Steal The Entire Tech Tree?

[Bandobras Took]

Since stealing techs is not based on what on opponent has researched, and considering the +1 spy from Command Centers and Feedsite Hubs, what's the most optimal way to steal the entire tech tree? What's the probability of tech steal failure? How helpful is it to enhance success by leaving agents at HQ?

 

[Cai30]

I've only quickly examined the XML for this stuff, so it's possible I'm misinterpreting. If someone can confirm I'm reading correctly, I'd appreciate it.

In the covertoperations XML file, Establish Network has a 'basedifficulty' of 10. From the context of the file, it seems the success rate is 100-basedifficulty expressed in a percentage. So that would mean Establish Network has a 90% success rate.

Success rates seem to be:
*Establish Network: 90%
*Siphon Energy: 65%
*Steal Research: 65%
*Steal Technology: 50%
*Recruit Defectors: 50%
*Hack Satellites: 40%
*Coup D'etat: 40%
*Affinity Operations: 30%

Interestingly enough, the quest operations Deliver Dossier and Extract Operative (something I've never seen in game) have a basedifficulty of 0, meaning they should have 100% success rates. My limited experience is I have not seen Deliver Dossier fail yet.

As for Ops Support and Ops Recon: Ops Support shows an attribute called 'agentsurvivalpercent' of 1. Ops Recon has the attribute operationsuccesspercent at 1 as well. So it seems that ops support and ops recon don't really do much for spying.

Haven't found the counter-intel stuff yet.

Edit: Started looking in the LUA, which is unfortunately not something I'm proficient in reading. It might have more answers. What little my minimal programming expertise self can figure out right now is it seems the tech one receives from steal tech is weighted towards cheaper less expensive technologies based on your own research rates. Which means to get a better chance at stealing high end techs one needs to have a decent science rate already, making it difficult to simply 'steal the tech web.'

 

[Bandobras Took]

Or just stagger the finish dates of missions. There are only 90 techs, I think, and ARC can run tech steal every 15 turns. Depending on how quickly you get the Agency up, your initial three could have time to steal every tech if success rate were 100%.

Unfortunately, it isn't. Thanks for the info; now I'll work on getting a solid 10 command centers and feedsite hubs early.

 

[Toranth]

Edit: Started looking in the LUA, which is unfortunately not something I'm proficient in reading. It might have more answers. What little my minimal programming expertise self can figure out right now is it seems the tech one receives from steal tech is weighted towards cheaper less expensive technologies based on your own research rates. Which means to get a better chance at stealing high end techs one needs to have a decent science rate already, making it difficult to simply 'steal the tech web.'

This is correct.
The game goes through every tech you could research.
1) The base "weight" of each tech is the number of beakers it requires.
2) It compares the number of turns it would take you to research the tech vs the number of turns it would take to steal. If the tech takes longer to research than steal, halve the weight, otherwise double it.
3) If the tech gives Affinity of the same Dominant Affinity as the defending player, add 50% to the adjusted weight (aka, it is now 75% or 300%, not 50% or 200%). This is the only time the defending player affects what technology you steal.

Then randomly selects one of those techs, based on the combined weighting.

So, if your research is better, more expensive techs get the x2 weighting bonus. But since you always get a tech anyways, researching the cheap ones yourself means fewer cheap techs to steal, and better probability of getting an expensive tech.

 

[alpaca]

In the covertoperations XML file, Establish Network has a 'basedifficulty' of 10. From the context of the file, it seems the success rate is 100-basedifficulty expressed in a percentage. So that would mean Establish Network has a 90% success rate.

The base difficulty in the XML is modified by

- 5% * your agent's rank + 5% * (defending agent's rank + 1)

then clamped between 10% and 90%

 

[Mercade]

In my experience, leaving agents in HQ is pretty useless. Those modifiers aren't worth the success chance another agent in the field can have (or doing counter-espionage in your cities).

 

[player1 fanatic]

What those modifiers do anyway?

 

[Cai30]

In my experience, leaving agents in HQ is pretty useless. Those modifiers aren't worth the success chance another agent in the field can have (or doing counter-espionage in your cities).

That's what I experienced from the 1% detection reduction or 1% success rate.

A scenario: Two agents out, one at HQ, with +success rate project. Let's use tech steal since it's a 50% success rate normally. Now you have 2 agents with 51% chance of stealing a tech each, or a roughly 76% chance for at least one of them to steal a tech each time.

Scenario two: Three agents out, 0 at HQ. 50% chance each for tech steal. 87.5% chance for at least one of them to steal a tech.

Whether you want to use the other projects for Wonder build speed or Health is another matter that is outside the scope of 'how to spy better.' Even then I'd rather have them doing counter-intel than at HQ, personally.

 

[Kaigen]

Propaganda (the health bonus) is only really noticeable if you're going super wide, since you need a sizable pool of health from buildings/biowells for it to be more than a point or two with such a small percentage. Even then, it'll only matter if you're hovering around one of the breakpoints, which is pretty rare.