How do you play Intelligence Agents?

[RemoteLeg]

Okay, here's a question for the players who are good at espionage.
How do I effectively play my Intelligence Agents?

When they first arrive they have some basic skills, but these never improve unless I send them on missions.
So I send my agent on his first mission.
- I always pick the easiest one at first: take one year to steal the target's territory map.
- If he fails he usually dies.
- If he succeeds, his skills improve a little.

Now I can either put him back on the safe option, stay home and prevent enemy intelligence missions, or I can send him on another mission and roll the dice again.
No matter how good he gets, there's always a small chance of failure and sooner or later the dice come up craps and he's gone.

Is this all there is to espionage, or am I missing something?
I would like to level up some good spies and protect them so they at least have a decent chance of coming home if they fail.

[RogerBacon]

I've heard that sabotaging construction in progress is actually the easiest mission. I always go for one month missions because 3 month missions only double the chance of success and 12 month missions only triple (or quadruple it, I forget).

[Retreat1970]

Sabotage, destroy, research, and maps are my primary missions.

Honestly though I play auto-spies, unless I really need something done.

[Aeson]

Is this all there is to espionage, or am I missing something?

That's basically all there is to espionage, so no, you're not really missing anything.

I would like to level up some good spies and protect them so they at least have a decent chance of coming home if they fail.

If you happen to have at least one character with Inspiring Presence, you can safely train your spies by putting them in the same location as the character or characters who have the trait. Be aware that training via Inspiring Presence can be somewhat slow by comparison to pulling off a few missions. Note that if training via mission, it would appear that the amount of new information provided by information-gathering missions (steal territory/galaxy map, steal research) affects the gain in agent skill and so it is a bad idea to attempt training intelligence agents by repeatedly stealing the same territory maps or doing other things like that. It is possible that there may be some similar mechanic in play for the other mission types. Also, while I expect you've noticed this already, the difficulty of stealing a research project seems to be correlated to the base cost (i.e. before research cost modification due to game settings) or tier of the project, though if I recall correctly a few technologies are more difficult to steal than would be expected given their tier (first two hyperdrive techs and species-specific techs) or their cost (species-specific techs).

As far as protecting your agents goes, the only direct control you have is over the missions you send them on. Counterintelligence agents basically never die; agents that get sent out on the other missions have rather more limited lifespans even if they're extremely highly skilled. If you happen to have an agent with a high assassination skill and intelligence agents can be selected as targets of assassinations, you might consider trying to kill off the intelligence agents of other empires before sending in agents on other missions.

A few things that seem to me as though they might be true:
- Agents with higher Concealment skill seem to live longer than agents with lower Concealment skill, regardless of assigned mission type (other than Counterintelligence, of course).
- The empire-level espionage bonus from species within the empire (and probably also from the faction leader's espionage skill, assuming that that skill does anything) seems to affect the relative likelihood of being undetected, being detected but evading capture, and being captured for national agents conducting missions elsewhere and for foreign agents conducting missions against the empire. In other words, it's riskier to send your own agents against empires with higher espionage bonuses than against empires with lower espionage bonuses.
- Pirate factions seem in general more likely to detect your agents than standard empires.
- It seems as though there are ongoing checks for agent detection, so it might be the case that agents are more likely to survive shorter missions than longer ones. Obviously, you'll need to make some judgement calls here, and I'm not going to recommend taking a 3-month mission that has only a 30% chance of success over a 1-year mission that has a 70% chance of success or anything like that, but if the reported success chance with a lower mission duration isn't much less than the reported success chance with a higher mission duration, it's certainly something to think about.

Note that I am NOT certain that any of the above is true. I haven't done any actual testing or any file-diving to try to confirm these observations. Regardless, if the above is true, there are a few more things you can do to try to protect your agents. Namely, try to avoid sending agents against empires with espionage bonuses or pirate factions, try to get at least some Concealment skill on all agents sent on missions against other empires and pirate factions (good luck), and start as or integrate a reasonably large population of a species with an espionage bonus.

Honestly though I play auto-spies, unless I really need something done.

Personally, I tend to control the agents myself and use agents with high espionage skill to aid my empire's research department; any other agents engage in counterintelligence missions. Occasionally I'll also assign agents to assist in hunting down pirates (operational maps can be a godsend when trying to hunt down a pirate faction's construction ship) or in gathering information useful for operational planning, but I tend to prefer more overt means of gathering such information. I don't really think sabotage construction, destroy base, or assassination missions are useful enough in general to bother with, especially against developed empires.

3 month missions only double the chance of success and 12 month missions only triple (or quadruple it, I forget).

Looking at the success chances reported in the game, I don't think that this is exactly accurate. There may very well be something in the game code that gets doubled or tripled/quadrupled when you go from a 1 month mission to a 3 month or a 12 month mission, but it's not the success chance, at least not as reported by the game.

Against the same target faction, the game reported the following success chances for a Steal Territory Map mission:
- With an agent with a 10% espionage skill: 25% for one month, 51% for three months, and 79% for one year.
- With an agent with a 30% espionage skill: 37% for one month, 72% for three months, and 86% for one year.
- With an agent with a 50% espionage skill: 49% for one month, 79% for three months, and 89% for one year.
All of these were checked at the same time, so there is no chance that the mission became more difficult between checks. Given the pattern, I'd be willing to accept that there is some part of the success/failure formula that is doubled for the three-month and tripled for the twelve-month mission relative to the one-month mission, but whatever that variable is, it is not the effective (or, at any rate, reported) success chance. It may also not be the dominant variable determining mission success or failure, especially for more skilled agents.

[RemoteLeg]

Thanks everybody.
It's a shame the espionage side of things isn't better fleshed-out, but I guess this is a strategic space 4X and the developers had to curb their enthusiasm somewhere.
I think I'll try the different mission types and not just steal territory maps so much, but it's difficult when the "map" mission has an 80% chance of success and stealing tech has only a 55% chance. I must also try playing a race that is good at espionage to see what sort of difference that makes.

[Shark7]

If you have any characters with the 'Inspriring Presence' trait, be sure to base your spies where that character is. It really helps them level up more quickly.

Other than that, my only advise is to repeat sabotage construction and steal territory map missions. You'll eventually get them up there. In one game, I had a spy with espionage skill of 120 when it was all done and over.

[johanwanderer]

I always send the agents out (a few at a time) on easy missions first to figure out what they are good at. Sometimes the skill are negative, so it is good to find out up front. Then, I tend to keep a few back for counter, and send the rest out to do various different things. I also turn on "suggest mission" so I know when I forget to send them out.