Spy misfourtunes.

[Tlalynet]

Espianoge economy is great, but the transition from teking to spying is pretty hard and puts me behind in the begining.

Certinly I've played great EP games, but I've had some go really badly too. The obvious dumb mistake is to piss off the wrong person, I've done that, and thats my loss. But I've had some games where bad luck screwed me. Most recently I lost 13 of my first 15 spies trying to get tecks, different cities of course, and 70-75% chance of success... Spies aint cheap at that point of the game and the whole EP advantage is a big military.

So what do you do, then, say screw the game and restart? Or try to tough it out?

I lost to supirior milatary later game, but I think after that it would be a good idea to just quit.

 

[Dracjn]

tough it out to the end u'll learn more

 

[VirusMonster]

Hmmm :-/ why don't you try sending a stack of 5 spies to a target city that has your state religion and is closest to your borders? It is natural to lose spies on the mission, yet your losses should not cost more than 10% of the research cost of the tech you are stealing. Spy production and maintenance costs compose a significantly larger portion of the stealing cost in the early game compared to Medieval era. I suggest you don't depend much on espionage economy early game, but switch to it after you have built some Courthouses and assigned quite a number of spy specialists.

For more EE information read: EEvsManualResearch

 

[Tlalynet]

Stacking more than one spy in a tile almost doubles each spy's chance of being caught. No state religion that game for diplo reasons. Obviously the hammer cost far exceeded that 10% margin at that point. By the end success chances where 74% and most missions where less than 75% cost, leaving a beaker to EP ratio of about 50% cost per teck. Of couse I started after currency and Code of Laws, but it doesent matter that the EP economy was good , losing that many spies and that much time held me back too far, and it wound up being axemen vs macemen ect. I think the only thing to learn there is that you cant lose that many hammers early on. Losing that many spies to chance early on by bad luck ends the game much like a turn 20 barbarian archer rush ends the game.

There may be some ideal way to run an espianoge economy, but you could never, ever reach it, the % chance of losing spies is very real. Usually its a pithence, but if you have bad luck, well, its tough luck.

 

[VirusMonster]

Stacking more than one spy in a tile almost doubles each spy's chance of being caught. No state religion that game for diplo reasons. Obviously the hammer cost far exceeded that 10% margin at that point. By the end success chances where 74% and most missions where less than 75% cost, leaving a beaker to EP ratio of about 50% cost per teck. Of couse I started after currency and Code of Laws, but it doesent matter that the EP economy was good , losing that many spies and that much time held me back too far, and it wound up being axemen vs macemen ect. I think the only thing to learn there is that you cant lose that many hammers early on. Losing that many spies to chance early on by bad luck ends the game much like a turn 20 barbarian archer rush ends the game.

There may be some ideal way to run an espionage economy, but you could never, ever reach it, the % chance of losing spies is very real. Usually its a pithence, but if you have bad luck, well, its tough luck.

:-/ I trust your experience... obviously your game was ruined, because as you said, the chance for getting caught doubles if 2 of your spies are on the same tile... and if you can't steal on time, you might lose the maceman vs. swordsman wars... Also, the AI might be running counter espionage mission aganist your spies?! perhaps steal from a weaker AI?!

In the link I provided, I assumed 5 spies would be enough to capture a single tech, but I forgot to add that if spies travel together as a single stack, the chances for getting caught will double per each spy...

You have to wait 5 turns to get the -50% espionage reduction. I also suggest you switch to a religion by the Medieval era, because even on Deity difficulty, switching usually makes sense... Maybe traveling to the same city via different routes will cost you less spies?

My approximation was as follows: for a 3000 beaker tech, if you can get around 2.4-3.0 commerce to espionage points ratio, then you are much better off if you steal the tech instead of manually researching yourself. Then, assuming 5 spies would do the trick and assuming it will take them 20 turns to reach the target destination, (5 for unit costs +5 supply costs) x 20 = 200 beakers goes for spy production and their maintenance. Then, assuming 1.25-1.50 to convertion ratio, this extra 200 would translate into 300 lost It would only increase the total price you pay by 10%, a negligible amount.

so are you saying 5 spies are not enough to steal a tech for sure?

 

[Tlalynet]

That is a good question about about conterespianoge, but CE tripples EP cost of missions, I dont know if the AI knows how to run it or if it effects success prop.

I did switch to a weak AI after spy 7, at that point I'd only stolen one teck, but switching is equivalant to abandoning reserch if you cant get the last of your EP out of the target AI. Normally I produce 2 spies per teck I want to steal, and that usually leaves me with surplus spies in the long run, The overall odds of their survival even with the wait is usually about 55%. 5 spies would have exceedingly good odds of getting a teck for sure, but alas the RNG foiled me.

I dont think the game was cheating or anything, but it helps me understand how people feel when they say the computer cheasts with the RND generator, especialy becasue that usually comes from players whove never tried to win a 5% battle, and dont see how it balances out. I've probably played a dozen EP games, and only 3 have ended in disaster due to bad odds, mostly becasue later on 5 spies or even 15 is not hard to manufacture. The only big unkowen is where enemy spies are stationed, thats why i mentioned i targeted different cities, but I know the AI was making spies because my EP was good enough to see them doing it, so for all I knew they had a spy in each city. In that case my odds would be bad even with that many.

Its good overall math, have you run an EP economy yet? Its very fun especially if you're playing a levle higher than normal. I wouldnt send a spy on a mission it took 20 turns to get to though, at least not at that point of the game, usually they 10 or less moves, 20 or less tiles from the cap if at all possible. Distance really inflates you're costs.

Well, thanks to everyone who heard out my rant about bad luck.

 

[VirusMonster]

well, usually, I manually reseach a single tech that leads to a tech monopoly. At the same time, I am using spy specialists to generate enough to steal from my current or next enemy. It does work, because I would have produced a high number of spies(5-10) anyway for city revolt missions. I don't aim to steal the most expensive tech, but rather the cheaper ones. I haven't tried a full economy with no manual research yet tough :-/

The reason I don't abandon manual research is the obvious trade value of a new tech.

I know 20 turns is an overestimate for all spies to reach a possible target city. I forgot they start with commando and travel just as fast in enemy territory. Still, that only makes their costs cheaper

I'll dig further into spy survival math and update my article. There is a good article in the forums about spy survival ratios.

 

[VirusMonster]

I did switch to a weak AI after spy 7, at that point I'd only stolen one teck, but switching is equivalant to abandoning reserch if you cant get the last of your EP out of the target AI. Normally I produce 2 spies per teck I want to steal, and that usually leaves me with surplus spies in the long run, The overall odds of their survival even with the wait is usually about 55%. 5 spies would have exceedingly good odds of getting a teck for sure, but alas the RNG foiled me.

Alright, I looked up all articles on spy detection, espionage mission costs and updated my own article as well... You are very right that sending a sufficient number of spies to ensure the success of the espionage mission is crucial.

What I found is:

1) Always steal from an AI with whom you have Open Borders or the chances of your spies getting caught increase drastically.

2) Send at least 5, if not 6 spies, for each tech stealing mission. I think you have been sending too few number of spies; 2 is definitely not enough :-/

How did I find that 5 or 6 spies are necessary to ensure success of the espionage mission? Well, if you have Open Borders signed, then each turn the chance of one of your spies getting caught is roughly 6%. Assuming your spies must spend 3 turns to reach the target city and 5 extra turns for the -50% bonus to kick in, then the chance of survival at the end of the 8th turn becomes (1-0.06)^8=60%

The spies can also get caught during a mission. It turns out the chances for getting caught range between 0% and 75%. I assumes 50% success rate. Then, 60%*50%=30% success rate for each spy trying to accomplish the espionage mission. ceiling(100/3)=4, hence, at least 4 spies must be sent to secure an espionage mission. To avoid truly unlucky scenarios, I would suggest sending 5, if not 6, spies per each tech stealing mission.

For info where I got those ratios in my approximation, refer to my article I linked earlier. It has been updated now.

 

[Tlalynet]

Its been a while since I read the global defines, but I recall the chance per turn of geting caught being as low as 2.5 once you get the EP economy clicking, its directly related to spending, so a pure EP economy gets less and less caught spies, especially if you last untill late game, then two per mission will leave a big surplus. I thought the base per turn chance of getting caught was 5% modified by spending ratio, but I havent gone through the global defines nor read an article about it in a while, so thats off my hat. Closed boarders hurts a lot though. Thats one of the tricks with the EP economy, to spy on a friend with open boarders or spy on an enemy without?

When you run hybrid EP and Tecking you will lose more spies becasue of the less EP spending, its a good plan, but it is give and take but the benifits probably outweigh the spy costs.

I usually have 3-5 tecks I want to steal at any given time, so that usually amounts to at least 6 spies in the feild, cities allowing, and with a pure EP economy I can usually get 1 teck per 2 spies inconsistanly early, consistanly late. That could just be good luck though. I do target rare tecks when I see them being reserched, the AI usually wont trade tecks it just reserched, but you can steal them and trade them away the very turn they get them, so I go for that when possible.

I whent back to reviewing the XML, closed boarders adds 15% to you're chance of getting caught, as do most things like recent missions and an enemy spy at the target city, and having more than one of you're own spy in a single tile. The chance of success is limited to 25%-75% based on relitive spending. I cant find the per turn chance of getting caught again, but I remeber it being low as long as you have open boarders. Even with that bad luck streak I only lost 2\15 while outside of missions, its an anomoly to lose them to waiting, I've kept spies in cities for as many as a hundred turns before I lost them some games, though that is also luck. 50% on mission success is where you have the same amount of EP total in the whole game as the opponent, if you use partial EP you probably will wind up around 60%, and full ep gets to 70 and up to 75 pretty quickly. Alltogether a 50\50 chance of getting what you want with a full EP economy and -50% wait cost is not too unreasonable, and post communism I ususually have a surplus.

You're high score game seems like a very well played one, +1Str to Jags! I've never quecha rushed so I didnt look at the third article.

Edit:
I now see the resason for our discrepancy. Bhruics examples set out a 12.5% intecept chance per turn in enemy territory with an EP disadvantage and closed boarders, but a .6% chance with open boarders and an EP adavantage, .6% not 6%, and runing an EP game you would run closer to 2, then 1, then less than 1% chance per turn of geting caught as you got closer to the situation in Bhruics example.

 

[VirusMonster]

I usually have 3-5 tecks I want to steal at any given time, so that usually amounts to at least 6 spies in the feild, cities allowing, and with a pure EP economy I can usually get 1 teck per 2 spies inconsistanly early, consistanly late. That could just be good luck though. I do target rare tecks when I see them being reserched, the AI usually wont trade tecks it just reserched, but you can steal them and trade them away the very turn they get them, so I go for that when possible.

I have the same experience: usually the AI has 3-5 techs I want to steal as well, but since I rarely run a pure EE, I don't think I can steal the most expensive techs, so I don't even try. I go for the mediocre ones. As you suggested, I probably should try more I also realized that controlling the holy city of the religion present in the target city leads to a huge reduction.

I whent back to reviewing the XML, closed boarders adds 15% to you're chance of getting caught, as do most things like recent missions and an enemy spy at the target city, and having more than one of you're own spy in a single tile. The chance of success is limited to 25%-75% based on relitive spending. I cant find the per turn chance of getting caught again, but I remeber it being low as long as you have open boarders. Even with that bad luck streak I only lost 2\15 while outside of missions, its an anomoly to lose them to waiting, I've kept spies in cities for as many as a hundred turns before I lost them some games, though that is also luck. 50% on mission success is where you have the same amount of EP total in the whole game as the opponent, if you use partial EP you probably will wind up around 60%, and full ep gets to 70 and up to 75 pretty quickly. Alltogether a 50\50 chance of getting what you want with a full EP economy and -50% wait cost is not too unreasonable, and post communism I ususually have a surplus. .

For the example in my writeup, I used a %50 mission success chance, but you can easily adjust for 70-75% mission success rate. Just plug-in 70% into the math: 60%*70%=42%. Ceiling (100/42)=3 spies should be enough to secure a mission. I like to bring a few more just to be sure, so 5 would do the job for sure. In conclusion, in a pure EE 1-2 less spies per mission are required than a hybrid EE.

You're high score game seems like a very well played one, +1Str to Jags! I've never quecha rushed so I didnt look at the third article..

Well thank you mister I spent much time writing that high score game up, I hope you enjoyed reading it

Edit:
I now see the resason for our discrepancy. Bhruics examples set out a 12.5% intecept chance per turn in enemy territory with an EP disadvantage and closed boarders, but a .6% chance with open boarders and an EP adavantage, .6% not 6%, and runing an EP game you would run closer to 2, then 1, then less than 1% chance per turn of geting caught as you got closer to the situation in Bhruics example.

.6% is very hard to achieve... Under open borders, the chance of getting caught ranges from 0%-7.5% For the example scenario in my article, I wanted to put a realisticly hard spy detection rate, so 6% looked like a difficult scenario. Of course with less spy detection rates, at end of 8 turns inside enemy territory, the chances of getting caught = (1- spy detection rate)^8. 2% spy detection rate would correspond to .98^8=85% success rate, significantly higher than 60%. Then, with high mission success rate of around 70%, 85%*70%=59% total success rate. ceiling(100/59)=2, so 2 spies should theoretically be enough for securing the mission well you said you were using 2 spies successfully, right? so the math works out fine.

I have witnessed a few mistakes in the calculations of those spy game mechanic articles, and some of the examples they provide are far away from game reality. For example, according to detektyw's article steal stealing mission increase base tech cost by 25% According to Krikkitone, they increase by 50%. Who to believe now? You gotta test it yourself

 

[Tlalynet]

I worldbuildered a 1 on 1, and the teck base cost markup was very roughly 55%, but this surprised me as it does not seem to be the normal markup, I thought it was smaller in my actual EP games. The Espianoge Steal Teck Base Cost Modifier in the XML is 0, so I dont know what this is tied to, perhaps it has to do with the # of civs that have the teck, in the test case it is low.
I've been interested in the holy city thing, but I've never founded a religion in the same game I was running Espianoge, or else I've founded Confusiconism but never spread it properly. Taking a holy city by force seems to dangerous in a game where you already have bad diplo, and last time I tried that got me dogpiled 5vs1.
Longterm in a Pure EP economy I've left sleeper cells for dozens of turns in countries I wasnt activly spying on at the time, just to have them ready if that civ got to be trouble or got someting I wanted, mid game in EP and especially late .6 is realistic, but you're right, early on even pure ep would have a couple percent per turn of getting caught.

Hail the spies, they seem underated on forums.

 

[VirusMonster]

I worldbuildered a 1 on 1, and the teck base cost markup was very roughly 55%, but this surprised me as it does not seem to be the normal markup, I thought it was smaller in my actual EP games. The Espianoge Steal Teck Base Cost Modifier in the XML is 0, so I dont know what this is tied to, perhaps it has to do with the # of civs that have the teck, in the test case it is low.

Spy tech stealing mission cost does not have to do with how many civs have the tech, that only matters when you are manually researching it (max +30% bonus).

where did you read the espionage steal tech base cost modifier is 0? can you post the file name here so I can look up? I suggest you to load up one of your games and do some math to find out the tech cost under EE before the reductions kick in, then compare to the actual expected tech base cost value... but you probably would have to look to research cost articles to find out the real tech base cost. Tech base cost could depend on map size, difficulty, game speed, etc...

This discussion has been very educational, thank you for the points you brought up... I will do one final tuning to my article I want to add the following analysis:

say one of your spies have a pessimistic 30% success rate for a tech stealing mission. Then the expected number of spies to secure the mission becomes ceiling(100/30)=4.

But the reality is very different. There is 70% failure rate. With 2 spies failure rate becomens %49, with 3 spies 34,3%, with 4 spies 24%, with 5 spies 16,8, with 6 spies 11,7%, with 7 spies 8,2%, with 8 spies 5,7%, with 9 spies 4,0%, with 10 spies 2,8%, with 11 spies 1,9%, with 12 spies 1,3%, with 13 spies %0.9.....

I hope you can see my point... :-/ While on average you should expect to lose not more than 3-4 spies on the mission, you should still bring roughly 10-12 spies to secure the mission :-/ well 30% success rate was roughly the worst game scenario for a hybrid EE economy, in a pure EE success rate is significantly higher, thus you can get away with less spies. So my feeling is hybrid EE is difficult to manage in the early Medieval era. Perhaps in addition to all cities getting the courthouse spy specialists and after castles have been built in a few cities, only then hybrid EE becomes a viable option. Without the espionage spending superiority against the target enemy, the spy losses can ruin your game.

A lot also depends how large your empire is relative to theirs.. If you can get early lead in landmass, then you could support a hybrid EE and still beat the target enemy in espionage spending. A smaller empire would be better of with pure EE, since the spy losses of a hybrid EE could cause problems.

 

[VirusMonster]

hey tlalynet, I updated the EEvsManualResearch section of my article once again.... Look fantastic now, everything explained very clearly and with good suggestions based on your feedback.

 

[Tlalynet]

I thought I saw a steal teck base cost modifier in Global Defines, but I couldnt refind the esp mission modifier when I serched again today.
Either way, the basecost Markup is 50%, the rougly 5% discrepency was from the difference in displayed teck cost and actual base tech cost, which I did not factor when I first looked. The only other thing I can think of that I havent checked is differences based on difficulty level. It's a solid 50% on Monarch anyway.

The failure rate array is very good and much more accurate and informational than a hardline 100\30, nice calculations.

I agree that an EE hybrid economy becomes viable after Castles and spy specialists. If you got GW and Pyramids for a spy specialist and republic you would already have good % discount for total EP and 4EP\4Sci spies. That sounds like a very good plan after Constitution even without the early game wonders, I think I will try my next game like that. If I play on levels that I'm not really good enogh to win on I usually play EE, because I can keep up better and usually last longer and get closer to winning.

As a counternote, I just had a REALLY good EE game, where I started 5 tiles away from the Hindi founder (saladin) and poped Hunting from a goody hut as Mali. A good Skirm Skewer away I had an early game holy city, tecked meditation, and spread it to be the #1 world religion. (Judism and Buddism competed for a while, but in the end only their founders stayed true) -40% Base for cities with you're holy city religion and a different state religion! Wow, that was worthwhile, once I had a commerce lead I didnt bother waiting for key tecks, and when I did wait I got them for 33% the cost of reserching. I had strong enough commerce and good timing to steal education, swap to reserch, and get libralism by 2 turns. The -1 and -2 for spying didnt even hurt diplomacy enough to mater. It was a very very great game, but lucky in my favour as much as the last one was unlucky in my favour.

As a side note, planting spys in cities you want to do passive missions on actually helps them. The -% for stationary spy counts to the passive mission as well active ones. If you want to see what a city is doing (I wanted to know if they where building the Spiral Mirianet) you can set a spy there just to, well, spy. I never knew that before.

 

[VirusMonster]

Hi sup great post again... Let me add my final 2

I thought I saw a steal teck base cost modifier in Global Defines, but I couldnt refind the esp mission modifier when I serched again today.
Either way, the basecost Markup is 50%, the rougly 5% discrepency was from the difference in displayed teck cost and actual base tech cost, which I did not factor when I first looked. The only other thing I can think of that I havent checked is differences based on difficulty level. It's a solid 50% on Monarch anyway.

Well for my Immortal game I wrote the article for, the espionage base cost is calculated through multiplying the technology cost by 1.5. Check the article for the espionage base cost on Optics.

The technology cost depends on game difficulty (1.15 modifier for Monarch), game speed ( 3 for marathon), map size (1.5 for huge maps), and tech base cost. I would need to know the exact settings to figure out the reason for the 5% discrepancy.

The failure rate array is very good and much more accurate and informational than a hardline 100\30, nice calculations.

Yea, it is very important to bring enough number of spies to secure the mission. I did not realize that until you started this thread.

I agree that an EE hybrid economy becomes viable after Castles and spy specialists. If you got GW and Pyramids for a spy specialist and republic you would already have good % discount for total EP and 4EP\4Sci spies. That sounds like a very good plan after Constitution even without the early game wonders, I think I will try my next game like that. If I play on levels that I'm not really good enogh to win on I usually play EE, because I can keep up better and usually last longer and get closer to winning.

yea, 4/4 from a spy specialist under Representation is really strong, so constitution becomes really important especially if you haven't build or conquered the Pyramids. And I completely agree with you that EE is the way to go at higher difficulty levels, I will definitely try more EE when I am going to attempt a deity win soon EE all the way

As a counternote, I just had a REALLY good EE game, where I started 5 tiles away from the Hindi founder (saladin) and poped Hunting from a goody hut as Mali. A good Skirm Skewer away I had an early game holy city, tecked meditation, and spread it to be the #1 world religion. (Judism and Buddism competed for a while, but in the end only their founders stayed true) -40% Base for cities with you're holy city religion and a different state religion! Wow, that was worthwhile, once I had a commerce lead I didnt bother waiting for key tecks, and when I did wait I got them for 33% the cost of reserching. I had strong enough commerce and good timing to steal education, swap to reserch, and get libralism by 2 turns. The -1 and -2 for spying didnt even hurt diplomacy enough to mater. It was a very very great game, but lucky in my favour as much as the last one was unlucky in my favour.

that is the way to go congrats! Religious modifiers affect the effective to conversion rate of an EE immensely.

Multiply espionage base cost by 0.85% if target city contains your state religion and if AI has a different state religion than yours. Multiply the espionage base cost by a further 0.75% if you control the holy city of the religion present in the city. Then, total cost reduction due to religious modifiers becomes 0.85%*0.75%=63%75. (it is not 40% as you mentioned, check my article out to find how these modifiers affect the final espionage base cost)

edit: I checked the SDK again, and you were right. The religious modifiers are summed before getting multiplied. So (15%+25%)=40% off yea

One more observation on your game stlye is that you use burst manual research and burst espionage economy. I am thinking along the same lines that it probably optimal to focus either EE or manual research to discover the technology you want earliest. When you gain the tech the earliest you can, then you can also backfill through tech trading.

As a side note, planting spys in cities you want to do passive missions on actually helps them. The -% for stationary spy counts to the passive mission as well active ones. If you want to see what a city is doing (I wanted to know if they where building the Spiral Mirianet) you can set a spy there just to, well, spy. I never knew that before.

I did not know that, good tip, thank you Take care, this thread has been very educational to me...