Espionage Risk vs Espionage Reward

[StormingKiwi]

It is intended as Espionage works right now that you use your Ambassadors in conjunction with your Spies for the best results and that the most difficult missions should be difficult high risk/high reward endeavours that could cost you a good spy. As was pointed out, there are many possible ways to fail not all of which result in the loss of your spy.

It's also pretty clear that the current policy risk threshholds are not properly adjusted to the success chances we have at this time.

With the intention of keeping a tight focus on improvements to the espionage system, I would like to discuss these two points made by Erik in a recent thread.

I disagree with the conclusions reached by Erik after interpreting the information presented.

1. "The most difficult missions should be difficult high risk/high reward endeavours that could cost you a good spy"
The reward for the most difficult missions is low, and the missions are not worth the potential of losing a good spy.

For example:
Kygnos Consciousness has:

• 40 colonies
• 19 spaceports
• 60 research stations.
• 334 mining stations.

Destroying a station or sabotaging construction is not worth potentially losing any spy for. The spy is 10% of my total. 1 mining station is 0.3% of their total.
The loss of a spy is significant. The loss of the mining station is a rounding error.

Proposed solution: Increase the value of a successful mission.


2. It's also pretty clear that the current policy risk threshholds are not properly adjusted to the success chances we have at this time.
Currently, the policy risk thresholds are pretty much right (if 200,000,000 players each send spies on 10 missions with an 85% chance of success, you'd expect one player would never experience a failure)

The success chances are much too low. The risk thresholds are not too high.

If risk thresholds are to decrease, then spies need to be much more disposable.

[Nightskies]

I had gathered a bunch of semi-obsolete data through a Spy Master play style from before the 1.1.1.1 beta patch, most of which looked like this...

chartsnip.png (16.15 KiB) Viewed 1882 times

...including a chart in which I periodically checked the chance of success of every action if it was above 15%...

chart2snip.png (11.8 KiB) Viewed 1882 times

...and notable events.

chart3snip.png (13.12 KiB) Viewed 1882 times

If you want to look at that save game for whatever reason, here it is:

Yirphunea Technocracy - Zenox - 2838-05-07.rar

(6.47 MiB) Downloaded 36 times

I did a handful of additional tests in the same galaxy dozens of years later into it after letting it run on auto for a while, just to be sure that even larger, more powerful empires didn't change anything up.

Unfortunately, it only shows what we already know concerning the topic. After some 50 lines of entries, I realized it wouldn't help, but I kept checking the numbers anyway in case anything telling happened (nothing really). However, I did figure out a few things while testing this. The big one is that the chance of success is inversely proportional to the target empire's population. Revealed by game editing, I deleted the capital of the #2 empire and checked how that affected success estimations, then reloaded and, subsequently, their largest population world otherwise. That removed their leader, all spies, and their spy facility for the capital. The chance of success was bumped up about the same in both cases.

I suspect, but have not confirmed, that the spy's host empire population contributes, meaning that it could be proportional population is the significant factor. Not essential to verify, IMO, but worth noting. Other possible contributors could be the number of colonies in the target empire, income, the total amount of corruption $ lost, etc... in any case, all those possibilities correlate to population. And, of course, this doesn't apply to pirates. Other factors like total technology may matter, but that was beyond the scope of my tests.

Since Aurora, I've only played the Haakonish, whose spies have been ... utterly useless ... they haven't succeeded in a single mission against an empire worth anything, including territory map stealing. Since it is on very difficult settings, this isn't a surprise. Anyway, the beta patch made low-success-chance missions significantly more likely to succeed, such that despite having the worst spies I remember having, they still have close to the same chances of success as the Zenox empire I tested before the 1.1.1.1 beta when it comes to those 'high-value' missions.

Now, despite not getting much out of the data (sigh), spending all that time with the Spy Master run did force me to reflect on espionage, made further evident in the subsequent Haakonish run with some hindsight.

• Spies are inherently free, and their actions are too.
• Their value comes from scarcity and the potential for successful actions.
• As the chance of success is inversely proportional to population, the value of spies against high-population empires is lower due to the much smaller potential for success not keeping up with the desirability of success.
• The cost of getting caught is only a tiny reputation hit and some diplomatic ire from the target empire. Concealment seemingly makes getting caught less likely. Getting caught leads to capture far less frequently as of 1.1.1.1.
• The cost of a spy getting captured is that the spy can no longer take action; in other words, they lose their value until freed. There is also a significant chance they'll develop a very bad trait.
• The cost of freeing a skilled spy is almost always worth it, as their value far exceeds the price. Bribing a skilled spy is undoubtedly worth it unless they are patriotic.
• The fewer you have, the more likely you'll acquire a new spy. It seems the bigger population you have also affects how likely you'll get another.
• A captured spy seems to count toward your total spy number, making the appearance of further spies less likely. However... with automated Character Locations, your AI will periodically dismiss bad spies that are captured. You can't do that. This is probably not intended. Perhaps it's a bug, though I've been able to duplicate it easily.
• Successful technology and map stealing directly benefit the empire. The value of these missions is clear.
• Counter-espionage effectively slightly hurts the reputation of the offending party and has a chance to net you a captured spy, and as aforementioned, is likely worth bribing or selling back.
• Other actions (sans Deep Cover) damage other empires, the value of which can't be easily measured. The value of these missions is largely dependent on the user.
• The value of numerous successful direct-value missions outweighs those of others- an espionage spy is far more valuable than a saboteur/assassin/psy-op. This is amplified by Ambassadors' aid to espionage actions, which doesn't contribute to other missions. They can have a counter-espionage bonus, but I can't tell how impactful that is or how it works on an ambassador.
• Although higher technology is more difficult to steal, they are worth multiple times as much research as earlier tech (never mind the weird values coming from diplomatic trade right now), so they do not gain or lose value but are more likely to cost in failure. They are still worth a net gain until the population becomes a prohibitive factor.
• Counter-espionage already had significant diminishing returns for multiple agents performing it. It is even more so now, so a few spies good at counter-espionage similarly reduce other spies' value at it.
• Excess spies only have significant value in the potential of successful missions in assassination, psy-ops, and sabotage.

Yeah, that's a lot of points, but they're all relevant. Something very, very important about all that?

Spies aren't as precious as they feel.
They are all expendable. The fewer you have, the more you get. The bigger you are, the faster they come. You can get them back diplomatically with location information, pirate contacts, or just throw money at the problem. Dismiss them without a second thought if they don't have a good trait or desirable skills. Hope they get killed instead of being captured when they don't have great skills. Don't feel bad about that. They know what they're doing is a top-tier dangerous activity.

There are a lot of hold-ups because espionage gives so much direct value; the other options seem to be a waste by comparison. Yes, espionage is great, but it tapers off, and it's only good if you can get good spies. That's not always a viable option. Meanwhile, the other missions don't need much to get a lot out of them. Just takes a bit of luck and some persistence.

All you need is 1-3 counterespionage buffs and a few (or no) espionage experts, then use the rest on more risky missions. If you lose an espionage expert because they've gone traitor or got selected as Boskara dinner, train a new one, and don't sweat the loss. With that, I think it's worth continuing into the value of sabotage, psy-ops, and assassination.

Sabotaging destroys whatever is being built or a station. Pretty straightforward. In the save above, I saw an agent with a mere 5% sabotage skill from a trait alone target the #2 empire with a 22% chance to sabotage construction and a 37% to get a research base. During a war, if you claim their capital, you regain your captured spies if you don't glass the planet and aren't unlucky in the invasion. So why not send your spies that don't have value otherwise to try to blow up some ships or stations? Just don't throw out your espionage or counterintelligence experts at this.

Psy-ops: Inciting a rebellion just applies a negative modifier to approval for a single colony. The chance of success shoots up the closer the target colony is to rebellion. This is useful when fighting an enemy that refuses to give up in a long war. Instead of committing to an invasion of a world, just blockade it, then after a while, send in your PsyOps agent(s) to incite a little chaos. It might just be a 30% chance to work, and you might make it happen without a spy's help, and it might just be a -4 to -12 (or higher) bump to approval. If it works, though, they will join at 100% assimilation. That will save you a lot. Very worthwhile.

The revolution? If you have five spies with a 10% chance to incite a revolution against a warring Military Dictatorship, that's a good deal. A coup d'etat eats up an admiral/general, maybe even 2 or 3, and imposes a -18 approval. That's not even taking into account whatever bonuses the leader had. Your spies are worth far less than that.

Assassination seems to be the least impactful of these. You can't see enemy character traits, just their skills. It also seems to be the least likely to succeed overall, so I have to wonder if there's a trick to using assassins like there is for inciting a rebellion. I have no idea. Nonetheless, a spy with the right trait and skill in assassination could be more useful doing just that instead of the other missions.

A tip. A spy with the Inspiring Presence trait instantly makes them the best waifu writer for your spies, and you should never let them leave home. Being a Demoralizing Spy makes them worse than a mutant in the Imperium of Man, which is grounds for instant execution. Both these continue to affect spies so long as they aren't captured.

[Erik Rutins]

• Spies are inherently free, and their actions are too.
• Their value comes from scarcity and the potential for successful actions.
• As the chance of success is inversely proportional to population, the value of spies against high-population empires is lower due to the much smaller potential for success not keeping up with the desirability of success.
• The cost of getting caught is only a tiny reputation hit and some diplomatic ire from the target empire. Concealment seemingly makes getting caught less likely. Getting caught leads to capture far less frequently as of 1.1.1.1.
• The cost of a spy getting captured is that the spy can no longer take action; in other words, they lose their value until freed. There is also a significant chance they'll develop a very bad trait.
• The cost of freeing a skilled spy is almost always worth it, as their value far exceeds the price. Bribing a skilled spy is undoubtedly worth it unless they are patriotic.
• The fewer you have, the more likely you'll acquire a new spy. It seems the bigger population you have also affects how likely you'll get another.
• A captured spy seems to count toward your total spy number, making the appearance of further spies less likely. However... with automated Character Locations, your AI will periodically dismiss bad spies that are captured. You can't do that. This is probably not intended. Perhaps it's a bug, though I've been able to duplicate it easily.
• Successful technology and map stealing directly benefit the empire. The value of these missions is clear.
• Counter-espionage effectively slightly hurts the reputation of the offending party and has a chance to net you a captured spy, and as aforementioned, is likely worth bribing or selling back.
• Other actions (sans Deep Cover) damage other empires, the value of which can't be easily measured. The value of these missions is largely dependent on the user.
• The value of numerous successful direct-value missions outweighs those of others- an espionage spy is far more valuable than a saboteur/assassin/psy-op. This is amplified by Ambassadors' aid to espionage actions, which doesn't contribute to other missions. They can have a counter-espionage bonus, but I can't tell how impactful that is or how it works on an ambassador.
• Although higher technology is more difficult to steal, they are worth multiple times as much research as earlier tech (never mind the weird values coming from diplomatic trade right now), so they do not gain or lose value but are more likely to cost in failure. They are still worth a net gain until the population becomes a prohibitive factor.
• Counter-espionage already had significant diminishing returns for multiple agents performing it. It is even more so now, so a few spies good at counter-espionage similarly reduce other spies' value at it.
• Excess spies only have significant value in the potential of successful missions in assassination, psy-ops, and sabotage.

Yeah, that's a lot of points, but they're all relevant. Something very, very important about all that?

Spies aren't as precious as they feel.
They are all expendable. The fewer you have, the more you get. The bigger you are, the faster they come. You can get them back diplomatically with location information, pirate contacts, or just throw money at the problem. Dismiss them without a second thought if they don't have a good trait or desirable skills. Hope they get killed instead of being captured when they don't have great skills. Don't feel bad about that. They know what they're doing is a top-tier dangerous activity.

Thanks for this Nightskies, it's good advice and you have indeed figured out the basic philosophy in terms of how we look at the spy system. They are expendable, they are recoverable, they can be honed into superspies but not all need to be. If you use them up, you will get more to try again and the costs to using them are limited. When you need to use them, don't hesitate and the low risk missions aren't the only useful ones.

We have some more changes to make over time to get the balance to exactly where we want it, but I'm very glad that the principles of the system came through clearly in your play. I take that as meaning that we're getting fairly close.

Regards,

- Erik

[AKicebear]

Great post. The tidbit about no assimilation following a successful rebellion mission is especially useful!

Your comments on tolerable risk vs reward of some missions, especially sabotage and rebellion, seem to hint at new or adjusted thresholds for automatic spies to take on those missions - even at relatively low chance of success.

[Jorgen_CAB]

Destroying space stations is more of a strategic or tactical tool than an economical one... so you should not dismiss the value of a timely destroyed space station in a war or just before a war.

[frankycl]

@Nightskies: Thank you very much for this detailed report and sooo many nice hints and explanations.

But why do you write: "You can't see enemy character traits, just their skills." ?
While this is certainly true for the normal game, you could easily see their traits (for testing purpose) when using the editor and switching the factions, no ?
And I think this would help with the impact of ambassadors, too, don't you think ?

[frankycl]

Thanks for this Nightskies, it's good advice and you have indeed figured out the basic philosophy in terms of how we look at the spy system.
(...
I take that as meaning that we're getting fairly close.

Regards,

- Erik

Well, it's very nice to get this confimation from you, too - but please don't forget to put this (in a more brief way) into the game (-texts), too - like e.eg. the Galactopedia or even some tutorial-texts, because many gamers (on GOG and/or Steam) don't even read in those forums (let alone to join the discussions), but would be very happy about such info, too, I think !

[Nightskies]

@frankycl: Indeed, that is strictly from a normal playing perspective. I saw no reason to check traits for testing purposes, but perhaps you're on to something. I still think there is a trick to using assassins, perhaps it has something to do with traits... OTH, if it *does*, how would a normal player exploit it without using the editor?

Any ideas regarding assassins would be great... *eyes narrow and goes to Discord*

[Nightskies]

@Franky, AK, Erik: Thank you~

It might seem out of the blue, but I'd like to use this to apologize for my general insensitivity. I really like the game and hope that contributions like this will make the community better.

[frankycl]

It might seem out of the blue, but I'd like to use this to apologize for my general insensitivity. I really like the game and hope that contributions like this will make the community better.

There's nothing to apologize, Nightsk - I always found your input/reactions/wording, etc. very interesting and not in any way insensitive. (although I fear/must admit, I didn't understand everything sometimes )


Ok, back to the topic:

@frankycl: Indeed, that is strictly from a normal playing perspective. I saw no reason to check traits for testing purposes, but perhaps you're on to something.

I meant that getting to know these things (via the editor) would give further info on how the system works (= if traits are of importance at all or in what way, etc.).

I still think there is a trick to using assassins, perhaps it has something to do with traits... OTH, if it *does*, how would a normal player exploit it without using the editor?

Yes, that's an example for what the additional info (e.g. for traits) might be very interesting, because of two main reasons:

1. You will get the info (= if traits are of importance at all or in what way, etc.) in order to get to know what additional values might be the reason for increasing/decreasing success-chances in general.

2. You/others (as a player) will know if/how their traits (that they see on their own characters) might have an effect on the intelligence-actions of other AI-factions (if any) and might take this into account when deciding if the characters are to be dismissed or if they shouldn't.

However, all of this is only needed/of importance if the DEVs won't provide us with this info in the first place (because they should know all of this in detail) - hint at Erik (at least to put those info in the official modding guides) !