Tech Theft

[starlifter]

by Slowthinker:

contact Mark (Apolyton admin) directly by an e-mail. He is awaiting you and prepared to help.

OK, I'll do that later tonight. Thanks.

By lafayette:

Sorry, this is not in accordance with the results of my testing when we studied diplomats and spies, SlowThinker and I. Here are some results:
1) Inciting revolts: vet and non-vet succeed, but the cost is lower with vet (if 1 is the bribing cost when using a diplo, it becomes 5/6 with non-vet spy and 2/3 with vet spy))
2) Industrial sabotage:
Method #1 ('use her judgment'): losses = 25% with vet, 50% with non-vet
Method #2 ('choose primary target'): losses = 55 vet after 400 attacks (different kinds: capital, non-capital, walls, no-walls,...), 136 non-vet after 400 attacks (same attacks as those with vets)
3) Poisoning water supply:
Capital city, non vet spies, losses = 2/3
Non capital, non vet spies, losses = 1/2
Vet spies = 1/3

Outstanding!! Thanks for posting those numbers... I love the spy unit, but have not done similar tests, but have been meaning to for over 2 years. Was the above at Deity? Have you or anyone else noticed any changes in odds at different levels, or at different points in the Tech Tree? Just curious, no data myself.

The Capital difference with Water Poisoning is something I had not considered. Now that I see your numbers above, Capital Nuking makes the top of my nuke test list for one day looking into Nuking odds.

About #1, yup, that one is exact, due to the game algorithm for computing the gold. Did you find any difference due to Walls? I'd never noticed or considered that one, but someone on ST's site said his son had noticed it. The problem with just noticing stuff in a game (and not testing for it like you have) is that one's "perception" of odds in the heat of battle can often greatly affected.

A thing I've noticed, and am 100% convinced about (but have not tested for the exact numbers, soooo) is that going after City Walls with a diplomat will result in very low odds of actaully getting those walls. That is, if there are 4 improvements, a Dip will not have a true 1 in 4 chance of nailing the walls (or a 1 in 5 if you consider production). Instead, the odds are reduced (maybe by a factor of 2 or 3) for Walls (e.g., making the odds more like 1 in 8 or 1 in 12, using the 1st example above). These odds are jjust a hypothesis to point me to the fact I ought to look into it more sometime, however. Some people I known in real life that play an early conquer style like to use dips to strip a city of walls (in early game) and I watched some of their gameplay and they seemed to be right.

Have you (or anyone else) done any tests for Nuking? That was the one at the top of my Spy list (I actually have a list of stuff to find out more about, but not nearly as organized as Slow Thinker's and it's pretty much stuff that occurs in gameplay in my own games). I have not considered the effect of a capitol on nuking, but now suspect (based on your results and a new gut feel) that nuking a capitol will turn out to have higher odds, too.

Thanks for posting those test results!! Those are good details for determining just what is necessary, but not overkill, for the way I normally plan large invasion forces (which consist largely of spies, both vet and non-vet).

PS, maybe both of you can have a look at a new set of Civ 2 1-page summary sheets that nethog is preparing & feel free to scrutinize it with your knowledge & experience . This is the link to the thread

here.

Thanks!



EDIT: Fix typos, add link.

 

[la fayette]

Originally posted by starlifter


By lafayette:
Was the above at Deity? .

A thing I've noticed, and am 100% convinced about (but have not tested for the exact numbers, soooo) is that going after City Walls with a diplomat will result in very low odds of actaully getting those walls. That is, if there are 4 improvements, a Dip will not have a true 1 in 4 chance of nailing the walls (or a 1 in 5 if you consider production). Instead, the odds are reduced (maybe by a factor of 2 or 3) for Walls (e.g., making the odds more like 1 in 8 or 1 in 12, using the 1st example above

Have you (or anyone else) done any tests for Nuking?

PS, maybe both of you can have a look at a new set of Civ 2 1-page summary sheets that nethog is preparing & feel free to scrutinize it with your knowledge & experience . This is the link to the thread

here.



EDIT: Fix typos, add link.

1) Yes, all tests at Deity level.

2) Attack of city walls with diplomats (industrial sabotage, method #1)
Here is how it works: there is a pattern (that I might call 'normal results') which is as follows:
Dip #1 destroys unit or improvement under building
Dip #2 destroys first improvement in the city (according to the list in the city screen)
Dip #3 destroys improvement #2... and so on until
Dip #(n+1) destroys city walls (n = number of improvements inside the city)
The trick (one of the genial ideas of the programmers of this great game) is that there is a slight random factor: the result of any attack has 90% chance to be 'normal' and 10% chance to be different. This is the reason why most players (including you
) think that the odds are higher.
The truth is that, if there are n improvements inside the city, you are 100% SURE that you will destroy the walls IF YOU ATTACK WITH (n+1) DIPLOMATS (but there is a fair chance that the walls tumble down sooner).
P.S. Attacking with spies works exactly the same, but losses are lower (100% with dips, 50% with nonvet spies, 25% with vet).

3) Tests for nuking:
I didn't do any thorough testing (capital or not, walls or not). My biggest sample up to now has been the endgame of the ww79 scenario. Several hundred nuclear missiles previously used by the AI and the strong American civ (20 cities average size 20) in front of me. I decide to use my vet spies to make it short: attack with 30 vet spies, 10 spies lost, 20 cities destroyed, game over...
One important point: when one of your spies plants a nuke in a foreign city, there is a fair chance that all nations ally against you (better be prepared )

4) Summary sheet:
I shall check and let you know.

 

[starlifter]

The trick (one of the genial ideas of the programmers of this great game) is that there is a slight random factor: the result of any attack has 90% chance to be 'normal' and 10% chance to be different. This is the reason why most players (including you

Ahhhh.... this 90%/10% issue is most illuminating!!!! I did not realize how they introduced the random element...

Also, I had not noticed that the progression was in the city list order. In hindsight it should have been obvious . The practical solution we've used for critical cities (like capitals) was indeed the n+1 method, but because when the improvements are gone and a sabotage was conducted, only the walls were left.... which means a horde of early game dips to get into place in a era that is lucky just to have roads. As a result, the strategy was kill the capital, then bribe the remainder (hopefully with leftover dips).

So for a 3 improvement city, there is about a 30% chance of getting the walls before wiping out all the impovements.

This is quite enlightening.


about Nethog's summaries:

I shall check and let you know.

If you just post in the Nethog's thread, he'll see it. Thanks for your help!

 

[SlowThinker]

La Fayette,
you were faster than me.

But I plan to post full threads in html quoted in Info:dip/spy to Starlifter by an e-mail if he agrees and if Apolyton won't be still unavailable.

Originally posted by starlifter
For 100% sure things, like making Embassies, of course, the effectiveness does not change. To me )depending on the exact action a spy does), the effectiveness is measured by either/or :

1. Did the spy get caught?
2. Did the spy succeed & return?

I would suppose that steps in 7.1 are independent and so their probabilities should be gauged separately...

Edited after La Fayette rioted

 

[la fayette]

I have not checked the value of the 90/10% trick. It might just as well be 80/20% (and testing it seriously would need a very large sample). But we don't care, do we?
Which is to be remembered is:
1) One peeping dip (to know how many improvements there are in the city)
2) Then (n+1) dips (on the same turn of course, because otherwise you might be destroying the unit under building, turn after turn) to make sure the walls tumble down. And, if you are lucky, the walls might be destroyed at your first or second attack.

 

[funxus]

SlowThinker, I play as Deity, but I just started to play it a couple of months ago. That might be why I haven't noticed.

 

[la fayette]

Originally posted by SlowThinker
La Fayette,
you were faster than me.

But I plan to post full threads in html to Starlifter if he agrees and if Apolyton won't be still unavailable.
I would suppose that steps in 7.1 are independent and so their probabilities should be gauged separately...

SlowThinker
Sorry I don't understand you.
Do you mean you are willing to transfer Apolyton's GL?
...and what are the steps in 7.1?

AFAIK any thread at CF or Apolyton is free to reading by anyone. Aren't they?

 

[SlowThinker]

La Fayette,
I repaired my post about sending threads to starlifter.

starlifter has no access to Apolyton, pages don't load.

Originally posted by la fayette
...and what are the steps in 7.1?

The quote from Info: dip/spy:

7.1 It is possible to divide spy's mission into three parts:
Firstly, a spy may be or may not to be captured before completing her mission.
Secondly, her attempt may be foiled by enemy dip/spy positioned in the city for steal technology/"any will do" mission (see 11.2) ; she completes mission for all other missions
Thirdly, she may be or may not to be captured after completing her mission.

I meant you shouldn't mix the capture in step 1 and step 3.

 

[starlifter]

starlifter has no access to Apolyton, pages don't load.

I've sent MarkG the technical info, node owner address, and trace plots... but even last fall, it was very very difficult to connect to them except about 2 am. And now, their pages never load up. The problem is at a node in California, and the user (me) has no control over Internet routing. But maybe the Apolyton folks can get it resolved.

Trace.

 

[la fayette]

Originally posted by SlowThinker
La Fayette,
I meant you shouldn't mix the capture in step 1 and step 3.

Now I understand what you mean.
But I don't think I ever mixed step 1 and step 3, because step 1 (capture before she accomplishes her mission) mostly occurs in case of industrial sabotaging, method #2 ('choose primary target').
IIRC the only detailed results concerning that method are in my thread 'poisoning and sabotaging' at Apolyton (IIRC you made a nice table with them, didn't you?).

Step 1 also occurs in case of planting a nuke, but AFAIK no thorough research has been made in that field, and there is no step 3 , since the spy always escapes the city after having successfully nuked the city. The only results that I know of are those I published recently and they are easy to sum up: 30 spies attack, 10 are captured (step 1), 20 succeed and escape (no step 3).

 

[Old n Slow]

Originally posted by funxus

I've never seen it. Does it depend on the version you play, I play 2.42. I'm not sure if the AI has ever bribed my units ever?

IIRC, I've seen it at the King level.

 

[la fayette]

BTW SlowThinker
I don't find my results concerning losses when stealing techs and sabotaging units. Have you got results somewhere or are we going to be compelled to reopen our labs?

 

[floppa21]

Wow I think my playstation version is all different...

 

[starlifter]

by Slow Thinker:

starlifter has no access to Apolyton, pages don't load.

In my e-mailing back and forth with Markos Giannopoulos, he thinks the Domain Name Server has old Apolyton.net info from last year.

He gave me a dot address, which is working, albiet slowly. That means I can now get to the Apolyton fourms, but I usually have time to cook something in the Microwave between each page load .

If anyone else has troubles with lots of page load timeouts, try this dot address:

http://64.246.32.51

 

[SlowThinker]

Originally posted by la fayette
But I don't think I ever mixed step 1 and step 3, because step 1 (capture before she accomplishes her mission) mostly occurs in case of industrial sabotaging, method #2 ('choose primary target').

I was talking to starlifter, not to you, see my post #24 here again.

Originally posted by la fayette
I don't find my results concerning losses when stealing techs and sabotaging units. Have you got results somewhere or are we going to be compelled to reopen our labs?

Do you mean In response to the sabotage thing and
Poisoning and sabotaging (both referred by yellow links from my Info:dip/spy) or something other?

Originally posted by starlifter
In my experience, veteran spies are 50% more efficient in doing "The 6 Dirty Deeds: Sabotage, Stealing Technology, Industrial Sabotage, Inciting Revolts, Poisoning Water Supplies, and Planting Nukes.

Are you able to load In response to the sabotage thing and
Poisoning and sabotaging (both referred by yellow links from my Info:dip/spy)? It looks that things are more complicated.

What I mean is that they can get the job done with 50% more success, and not get causght with 50% more success.

Again, see my post #24 here. I would suppose you cannot use this approach, but who knows? Maybe the odds of step 3 depends on the result of step 1 (and maybe this would explain strange results of LaFayette's and mine testing) .

 

[CaptBoulanger]

You guys seem to be the experts on spies, so let me ask you this -

If I wanted to rig the rules file in such a way that spies could be carried on submarines, do I just need to make them a missile (ie set "destroyed after attacking"), or do they also need to be coded as an air unit?

 

[duke o' york]

I think that an easier way to do this would be to give the submarine the flag that lets them carry ground units. You can either change the submarine so that it can do this, or change the transport so that it has the sub invisibility tag. Making the spy into a missile unit would mean that it'd just keep running out of fuel all the time. I'm no expert, so ask the people in the Civ 2 Scenario Creation forum for more advice if you need it.

 

[CaptBoulanger]

unfortunately your way seems to be the only workable solution; I'd hoped to avoid it, since it allows all units (not just spies) to be carried. However, making the spy an air unit and/or a missile seems to be something someone tried that did not work.

 

[ElephantU]

To nuke a city with a spy, do you have to "build" the nuke, or does it just show up as an option when the spy enters the city (assuming you have Nuc)?

 

[CaptBoulanger]

It becomes an option when the spy enters the city, once the conditions are met. You have to have NucFission, and (I think) someone has to have built the Manhattan Project wonder.

And of course you have to have Espionage; I'm pretty sure Diplomats can't nuke cities.