Are Espionage Success Percentages Known to Be Inaccurate?

[steveg700]

Not to sound bitter or nothin', but espionage missions with an 84 or 90 percent chance of success seem to fail an awful lot, specifically the Fabricate Scandal mission that's supposed to flip envoys out of the city-state.

Maybe I'm just unlucky, but I'd have to be *really* unlucky three 84% missions in a row.

 

[Datian]

I've come to suppose that it had something to do with the mission duration - "fabricate scandal" is a very long one, 12 turns if I remember right.
If your spy can get caught each turn (and the RNG checks for that), then 84% during 12 turns isn't so good.

 

[Chefofrats]

The percentages do seem off. I have not tracked the exact statistics, but for missions with 50-70% chance of success, I get many more failures than successes, and my spies almost always get killed, even though the chance of that happening is listed as something ca. 10%.

 

[_hero_]

I have suspected for a while now that the listed success rate doesn't factor in any anti-spy modifiers the defending nation has.

 

[kaspergm]

I have suspected for a while now that the listed success rate doesn't factor in any anti-spy modifiers the defending nation has.

That's my assumption as well, which would be logical.

 

[Tech Osen]

I have suspected for a while now that the listed success rate doesn't factor in any anti-spy modifiers the defending nation has.

This. So even when it says 90% without Gain Sources it's wise to use it anyway if the mission is important.

 

[Victoria]

espionage missions with an 84 or 90 percent chance of success seem to fail an awful lot.

This is known. (Well this part)
It took many many hours of testing because I had to play hundreds of missions manually to get the stats.
The irony of testing is this is explained in the texts of the mission result but we tend not to read well.

So a 90/7/3 indicates it has a 90% chance of mission success, this is not a lie, nor is the chance of failure but the 3% chance of being caught captured is a significant understatement.

The 3% itself is the guaranteed caught/captured stat and is strictly correct but what is not well explained is that both a successful and unsuccessful mission have a chance of detection based on level.

An example.
Your spy is fabricating a scandal and successfully does so but your spy is detected doing so (someone Heard then doing so in the toilets and told the CS ruler) your spy hears of this and has to escape... and the choices are by foot or by sea (because the city has a harbour).
The spy (you) chooses to escape by sea which is riskier but faster to get home and another roll occurs which I believe is along the same chances but modified by escape route level, enemy spy level) and this time a bad roll means you are caught. Because it is a CS that catches you, you cannot be sold back so they kill the spy.
If you now look at the resultant screen it screams at you that the spy is killed but if you read the text, the mission was successful.

Read the second message as an example, how the artwork got home, I do not know but it did.
Which is better for you as the aggressor, the first or second message?

What I do not know is exactly how the chance of escape changes with differing routes and the roll... because the roll is based on the routes. Also things like how the enemy can counter a fabricate scandal. Too much testing. I’ll drag out the original testing thread where I posted my results to show I’m not just making it up.

3 84% fails is unlucky but not very badly so. I did check rolls a little in my testing

The first promotion I get myself is ace driver if I can. I may fail some missions but my survivability rate seems better, cat burglar may help you succeed but not escape. It is a personal preference and others are not ‘wrong’ choosing +2 for a mission as it increases level growth.
Intelligence agency gives your spies +1 level as does quartermaster. These 2 really help with escapes also... and quartermasters stack... I have managed to test that.

EDIT: Original investigation thread final post is most useful
How to use spy as France

 

[Rosty K]

Well, it's probability. The percentage is only exact if you have an infinite amount of attempts. So yes, they're off, and will always be, because that's how it works

 

[Victoria]

because that's how it works

But many do not realise the probability of capture/killed in higher than quoted and that’s more to do with it than RND

I've come to suppose that it had something to do with the mission duration - "fabricate scandal" is a very long one, 12 turns if I remember right.

16 turns, reduced to 12 due to machieavellianism or 8 also including a linguist. But chance not altered by duration nor is siphon gold duration.

 

[steveg700]

Well, it's probability. The percentage is only exact if you have an infinite amount of attempts. So yes, they're off, and will always be, because that's how it works

That seems to have gone a rather long way to around to not really contribute anything.

This is known. (Well this part)
It took many many hours of testing because I had to play hundreds of missions manually to get the stats.
The irony of testing is this is explained in the texts of the mission result but we tend not to read well.

So a 90/7/3 indicates it has a 90% chance of mission success, this is not a lie, nor is the chance of failure but the 3% chance of being caught captured is a significant understatement.

Well, that all has to do with the part that happens after a mission fails, right? It's not pertinent to the actual chance of failing. The chance of failing at something that has an 84% chance of success comes down to less than what, a half-percent?

Overall, spy failure seems to happen more than its share of times in my experience. Just wondering if it occurs with others, at least on certain types of missions.

 

[TheMeInTeam]

Well, it's probability. The percentage is only exact if you have an infinite amount of attempts. So yes, they're off, and will always be, because that's how it works

There have been cases where due to a bug the percentage isn't exact even with infinite attempts. Usually it's confirmation bias, but testing is always an option if one wants to demonstrate something beyond that.

Victoria's already done some testing, and displayed odds are misleading at best. An unspoiled player would not reasonably anticipate their real %chance of being captured based on what the game represents to them. Even if everything else works exactly as advertised, that should still be considered an issue.

 

[Victoria]

Well, that all has to do with the part that happens after a mission fails, right? I

No.. that is the whole point!
The 4 messages in the middle of my post, did you read the second?
The mission is stated in big letters as a failure because the spy was killed... you are probably seeing these in your game as the 3% death but the mission was successful

 

[steveg700]

No.. that is the whole point!
The 4 messages in the middle of my post, did you read the second?
The mission is stated in big letters as a failure because the spy was killed... you are probably seeing these in your game as the 3% death but the mission was successful

So I'll tell you what I did notice the last time that seemed odd to me.

There was no notification of success or failure at the start of the turn. Rather, the game eventually prompted me to select an escape route. The spy escaped and the mission failed...again.

 

[Linklite]

I've had issues with it as well. I've not done many spy missions, probably only single digits actually, and I've never done a mission that had less than 74% chance of success (I'm somewhat conservative, to be honest), if all the odds are less than u4% I'll do a round of gaining resources or setting up a listening post instead, which are 100%. 1 succeeded, all the others failed. By that I mean, one got to the escape screen where you choose how to get away, and succeeded. All the others were either captured or killed, failing the mission.

The problem is that there will be people like me who just hit bad luck and get bad rolls. With over 5 million copies sold and using steveg's stat of 3 times in a row, that's still 110,000 people who experience that stat, assuming people only play their copy once (and how many do buy Civ 6 and only it the once?). The 5.39 million people who had the rolls work as expected or better won't post, but the 110,000 are more likely to post something about it, complaining.

The issue is whether it is just that we're the ones who got the raw end of the deal, or if the code doesn't match the description.

Personally, I think games with percentage chances like these should have the percentages tapered; if it fails the first time, then the odds are invisibly stacked in your favour the second time and more so each time till you succeed, when it resets.

I say so because I'm now somewhat disillusioned with spies; I can't buy them for some reason which means I have to take time out of my production schedule to build them, so I don't really want to be building them for what seems a 10% chance of success with small benefits if I can be upgrading my harbour or velodrome districts instead. As a result, they only get used in endgame and I don't need to upgrade my cities and I don't need to build up my army (because the situation is under comtrol). For me, they become at that point a novelty to play around with rather than a tool to help me win.

If they fudged the stats so that I am either unlikely to or can't get such bad runs (at least when starting out playing Civ, like me atm), then I wouldn't get soured by it.

 

[Victoria]

There was no notification of success or failure at the start of the turn. Rather, the game eventually prompted me to select an escape route. The spy escaped and the mission failed...again.

the spy finishes it’s mission and a single roll is done to decide on its 90/7/3 fate. Within that single roll let’s say it rolls around 20%. This means the mission was successful but was detected. It is all decided in a single roll (check my original referenced post) then your turn starts and there is a lot of notifications and amongst those at some stage it will ask you how you want to escape because your spy was detected. If the spy had not been detected you would have been told at the beginning of your turn but as it was detected it cannot fully resolve your notification until it decides if you escaped or not. Your mission was successful and then you decide to escape on foot and succeed you get a notification immediately that you succeeded. It is not the next turn, it may feel like it because some turns take a while.

The issue is whether it is just that we're the ones who got the raw end of the deal, or if the code doesn't match the description.

the description for success or failure is correct, the description for capture death is incorrect in reality but strictly speaking by the first roll is accurate. Those that get the raw end of a deal complain, those that do not, don’t. I have been hit by hurricanes 3x in games and it really ruined the entire week and when I complained many responded “I have never been hit by a hurricane”. Does that change how I feel? I play a lot and have lost a lot of spies but get the mechanics so at least appreciate that while my spy died, at least it removed that governor and the city will now flip to me.
If you use spies without the intelligence agency you are likely to suffer more.

 

[TheMeInTeam]

If the game prompted you to choose an escape route in advance of selecting the mission attempt, it could give you accurate odds on capture risk. That would have been a cleaner implementation in that it both gives a more accurate representation of the game to the player and the player does not have to engage with a notification/pop-up as often.

Player could even have a default route checked to ease the UI burden further, if Firaxis cared about end user experience with UI enough to consider it routinely with Civ 6.

 

[steveg700]

Those that get the raw end of a deal complain, those that do not, don’t. I have been hit by hurricanes 3x in games and it really ruined the entire week and when I complained many responded “I have never been hit by a hurricane”. Does that change how I feel? I play a lot and have lost a lot of spies but get the mechanics so at least appreciate that while my spy died, at least it removed that governor and the city will now flip to me.
If you use spies without the intelligence agency you are likely to suffer more.

Mind you, I am not even talking about a scenario where I got to actually fabricate a scandal and flip some envoys at the expense of my master spy. I'm talking about having wasted the not-small amount of time it takes to fabricate a scandal based on odds that seemed highly favorable. It's not aggravating because a bad thing happened, it's aggravating because the information the game gave me seems to be faulty. And this seems to happen not just in this limited instance, but rather as the norm for the Fabricate Scandal mission in particular. I don't think it was just the bad luck of something with a half-percentile chance occurring.

The hurricane thing is obnoxious because, as you've called attention to, the repair times for a district are excessive relative to the ease and immediacy with which they can be pillaged. Other than that, nothing indicates a broken system.

 

[Linklite]

Just wanted to clarify something I said in my earlier post; 1 succeeded full stop (ie did the mission, attempted to escape, succeeded and got be reused). ALL the others never got to the escape stage; they all failed by being captured or killed before completing the assignment (ie the odds should have been 26% (or less) of that happening at that point).

I play a lot and have lost a lot of spies but get the mechanics so at least appreciate that while my spy died, at least it removed that governor and the city will now flip to me.
If you use spies without the intelligence agency you are likely to suffer more.

Unfortunately I'm playing just the base game (the Switch is getting the expansions later on), so I'm not sure of what you're referring to, probably R&F/GS mechanics.

I don't mind losing spies and even bad runs because that's part of the game, but because it's happened at the beginning of my time with Civ, it makes me suspicious. Part of it is that I'm aware that in games things that are presented as being pure chance or independent of certain inputs really aren't (unless my luck is just ridiculous).

I'm coming towards ths end of a DV game with Australia and Egypt still standing and a few spies hanging around; I'll pay attention to the rages off success and see if it averages better.

 

[Victoria]

ALL the others never got to the escape stage;

Sure, I just cribbed it off the original post but keep that eye out and you will see more tried to escape and failed than initially messed up.

probably R&F/GS mechanics.

Yes, the intelligence agency comes in with R&F... of course you could always play as France with +1 level spies.

 

[Lonecat Nekophrodite]

Well, it's probability. The percentage is only exact if you have an infinite amount of attempts. So yes, they're off, and will always be, because that's how it works

Even the near 100% Doesn't mean the 0.001% will not happen.

And Playing as Kate Medici permits proactive use of spies early on. That's quite a killer thing since she can have an ace spy lurking inside cities of other civs before others can train the first one.