bribing non-barb units - Page 2

SlowThinker · December 26, 2000, 00:42

Tonic presented a formula specifying the amount of gold required for bribing:

G = M*S
Where
G = gold required to bribe
S = shield cost of the unit,
and M is given by
M = 2.05 + 0.0028*C
where C is the amount of gold in the coffers.

I want to present a more precise (i hope so...) formula that includes the distance from a capital and to add a formula for bribing cities:

Formula for bribing units
G = DF*GF*S
Where
G = gold required to bribe
S = shield cost of the unit (settler and engineer considered as 80-shield unit)
DF = 3*12.5/(D+2) (distance factor)
GF = (1+4*EG/3000) (gold factor)
EG = enemy's gold
D = the distance (of enemy's unit)from the enemy's capital (maximum 16)

Formula for bribing cities
G = DF*GF*S
Where
S = size of the city
DF = 5*200*SF/(D+3) (distance factor)
SF = 5/6 for a spy; 1 for a diplomat (spy factor)
GF = (1+EG/1000) (gold factor)
EG = enemy's gold
D = the distance (of your dip/spy) from the enemy's capital (maximum 16); (bribing from distance 1 is not allowed)

Is it correct?

The distance from the capital is straight (not diagonal; moving just one direction; just one coordinate is changed). Does anybody know how "S" should be set when moving both ways?
[This message has been edited by SlowThinker (edited December 26, 2000).]
[This message has been edited by SlowThinker (edited December 26, 2000).]

Maelhavok · December 26, 2000, 11:27

Hello Civ2 fans,

debeest, I use dips early for emergency non suppported defence. Good as you say for barbs. Good as 2nd unit in fort(no maintenance,bribe defence\bribe offense) take in trierme to buy cheap unsupported triermes and caravels, AND every city that has a dip is more resistant to bribe/theft/whatever from enemy spies.
I do not make a dip the cornerstone of my defence as you suggest some people do, but if enemy attacks an unprepared city with excessive force, buying 1 or 2 unstacked units to kill the stacked units is quite a viabe tactic.
Other Ideas but I gotta go

Merry Holidays,

Maelhavok

Blaupanzer · December 26, 2000, 12:02

quote:

Originally posted by SlowThinker on 12-25-2000 11:42 PM

The distance from the capital is straight (not diagonal; moving just one direction; just one coordinate is changed). Does anybody know how "S" should be set when moving both ways?
[This message has been edited by SlowThinker (edited December 26, 2000).]

For units, distance from capital is to the UNIT being bribed. For cities, distance from the capital is to the DIPLOMAT doing the bribing. Using this with the inaptly-named SlowThinker's formulae seems to work correctly. Good work!

La Fayette · December 27, 2000, 06:27

Slowthinker
It seems that your formula for bribing cities is essentially the same as the one I have in mind:
S(T+1000)/(D+3), where
S=Size of the city
T=Treasury of the civ
D=Distance to the capital (measured from the dip who is attempting to bribe).
Have you tested it? Have you tested 33% discount for vet spies? and 50% discount for city in disorder or empty or previously owned by your civ?
What about Dmax? You wrote 16. I have 32 in mind (10 for communist cities).
OK, I'm not THAT keen on bribery. But it's part of the game unless one decides to play no bribe.

SlowThinker · December 27, 2000, 10:13

quote:

Originally posted by La Fayette on 12-27-2000 05:27 AM
Slowthinker
It seems that your formula for bribing cities is essentially the same as the one I have in mind:
S(T+1000)/(D+3), where
S=Size of the city
T=Treasury of the civ
D=Distance to the capital (measured from the dip who is attempting to bribe).
Have you tested it? Have you tested 33% discount for vet spies? and 50% discount for city in disorder or empty or previously owned by your civ?
What about Dmax? You wrote 16. I have 32 in mind (10 for communist cities).

I found the same information as you on Winkler's page "Scrolls of Ancient Wisdom"(http://ltswww.epfl.ch/~winkler/civ2.html). There was formula for bribing cities only ,not units. Do you have another source?

When I sought formulas (by testing through cheat menu), i didnt know about Winkler's link, so you may consider that formula for cities is confirmed from two sources.
I tested 50% discount for city in disorder or empty, its all.
I think that Dmax=32 is a mistake on Winkler's page. Only elucidation is it depends on the size of a map, but it would affect tonic's constant 2.05 in this thread, but everybody agreed with him and I agree too.
Or Winkler counts some jagged way?
I think definitely there is a mistake in "Scrolls of Ancient Wisdom". I tested communism now, Dmax is 10 then, indeed, so 32 for other governments should be a mistake.

Note: communism doesnt affect bribing of units.
[This message has been edited by SlowThinker (edited December 27, 2000).]

SlowThinker · December 27, 2000, 19:41

I tried to concentrate all questions about Diplomats and Spies into one site. Look at http://apolyton.net/forums/Forum1/HTML/001678.html?0#0
[This message has been edited by SlowThinker (edited December 27, 2000).]

debeest · December 27, 2000, 22:43

I'll repeat my oft-stated complaint: the Scrolls of Wisdom formulae are wrong. I've tested both the bribe and the caravan bonus and ongoing trade bonus formulae many times, and they reallly don't even come very close to being accurate. I finally got around to providing concrete examples a month or so ago in a thread called "So you wanted examples" or something feisty like that. Slowthinker, since your formula duplicates the Scrolls formula, I'm curious how you could have generated the same formula, which I'll swear is just plain wrong. Can you explain?

For what it's worth, my experience is based on PC version 2.42, medium-size worlds, deity, 7 civs, often raging hordes but not always, and under many different conditions of government, relative power, etc. Do you think world size matters in this formula?

SlowThinker · December 28, 2000, 04:44

quote:

Originally posted by debeest on 12-27-2000 09:43 PM
I'll repeat my oft-stated complaint: the Scrolls of Wisdom formulae are wrong. I've tested both the bribe and the caravan bonus and ongoing trade bonus formulae many times, and they reallly don't even come very close to being accurate. I finally got around to providing concrete examples a month or so ago in a thread called "So you wanted examples" or something feisty like that. Slowthinker, since your formula duplicates the Scrolls formula, I'm curious how you could have generated the same formula, which I'll swear is just plain wrong. Can you explain?

I tested it now this way: on PC version 2.42 and MGE, medium-size worlds, deity, 7 civs, raging hordes. It works fine (both 2.42 and MGE). How do you count distance?
[This message has been edited by SlowThinker (edited December 28, 2000).]

La Fayette · December 28, 2000, 10:37

ST (about bribing cities)
I did not remember where I read it, but you are right: it was "scrolls of ancient wisdom" (link to it in SG's GL, page 2). I have just reread it: Winkler states that Dmax=32 is when the civ has no capital.
Shall we say that Dmax is 10 in case of communist government, 16 in general and 32 in case of "no capital" (until someone tests that thoroughly or tells us about tests we haven't heard of yet).
I have no other source, but I would like to state from personal experience that the 50% discount for a city previously owned by you and the 33% discount for bribing with a vet spy both are true or not far from the truth (because I love doing it and often do it on purpose).
Thank you for your tests (all tests are precious to careful readers ... and tonic's charts are really smart).

SlowThinker · December 28, 2000, 16:02

quote:

Originally posted by La Fayette on 12-28-2000 09:37 AM
I have just reread it: Winkler states that Dmax=32 is when the civ has no capital.
Shall we say that Dmax is 10 in case of communist government, 16 in general and 32 in case of "no capital"

La Fayette, I repeated a test: I claim that distance with no capital is 10 for communism and 16 for other government.

[This message has been edited by SlowThinker (edited December 28, 2000).]

debeest · December 28, 2000, 17:39

Slowthinker, last night I did some testing of my own (bribing cities, not units -- hey, this is my own thread, I can hijack it if I want to

) and verified a number of things:
- spy bribe cost is 5/6 of diplomat bribe cost
- veteran spy cost is 4/5 of novice spy cost
- cost is directly proportional to (EnemyGold + 1000)
- cost is directly proportional to CitySize
- cost (for a diplomat) is an integer multiple of CitySize x (EnemyGold + 1000), with the integer ranging no higher than 19 (or 16+3, you might say).

I've previously verified that courthouses double the cost, and originally owning the city halves the cost.

But I can't figure out the distance thing. I've read here that in Civ, distance is measured as (longest dimension + 1/2 shortest dimension), and that's what I've tried to use. But when I divided BribeCost by [CitySize x (EnemyGold + 1000)], the integers I got did not consistently reflect that either distance or (that distance + 3). And anyway, anyone who's ever tried counting distance in Civ has to have noticed that you can't always just count sideways and then count vertically; more often than not, you wind up having to add in a diagonal move. Certainly I found last night that bribing from "equidistant" spaces can result in a different integer, and bribing from "different" distances can result in the same integer. Help!!

Still, it's refreshing to find that certain parts of the formula are founded in reality. But why has my game experience been so different? Maybe the formula only works in cheat mode....

debeest · December 28, 2000, 18:07

SlowThinker, great dip/spy thread. I couldn't figure out how to post into it, so here's my feedback on it:

No, a diplomat is not guaranteed to succeed in tech theft. A spy is guaranteed to succeed if you don't try for a specific tech and it's your first time into that city.

I've also verified 1/2 off city bribe price for disorder, 1/2 off for empty.

Interestingly, my tests last night indicate that a veteran diplomat does not get a city bribe price reduction the way a vet spy does. Go figure. Maybe they thought no one would ever have a veteran diplo.

I cannot say it with certainty, but I'm pretty sure that no attitude adjustment results from expelling a spy. It really is "expected."

Again I'm not sure, but I think partisans begin to boil out of captured cities when the victim has learned industrialization.

And finally, the gold reward you get for capturing or bribing a city is not 1/3 of the enemy's treasury. I haven't been able to figure out what it actually is; I think maybe it represents the fraction of their treasury equal to (citizens in city / total citizens in empire). Check it out.

SlowThinker · December 28, 2000, 18:19

quote:

Originally posted by debeest on 12-28-2000 04:39 PM
I've previously verified that courthouses double the cost, and originally owning the city halves the cost.

I will inconspicuously direct you to MY thread

I think that your test result must be in conflict with
3.7 If courthouse is in the city then distance "S" is halved.
from MY

thread http://apolyton.net/forums/Forum1/HTML/001678.html?5
(i didnt test it, it is a winkler's statement)

quote:

Originally posted by debeest on 12-28-2000 04:39 PM
But I can't figure out the distance thing. I've read here that in Civ, distance is measured as (longest dimension + 1/2 shortest dimension), and that's what I've tried to use.

I didnt know this formula

I consider Civ map as a grid of xy cartesian coordinates, turned by 45 degrees. So my distance is measured north-east direction or sud-east direction or...

"longest dimension + 1/2 shortest dimension" could be answer to 4.32 in my thread.

[This message has been edited by SlowThinker (edited December 28, 2000).]

SlowThinker · December 28, 2000, 18:53

quote:

Originally posted by debeest on 12-28-2000 05:07 PM
SlowThinker, great dip/spy thread. I couldn't figure out how to post into it, so here's my feedback on it:

You couldnt figure?? I would prefer that all discussion about dips/spies would be concentrated there. You know, it is MY thread...

It is a weakness of appolyton, many ideas, but very straggled.
Note: I didnt find any tool at Appolyton that would search threads by subject or content...

quote:

Originally posted by debeest on 12-28-2000 05:07 PM
No, a diplomat is not guaranteed to succeed in tech theft.

How frequent is this phenomenon?

quote:

Originally posted by debeest on 12-28-2000 05:07 PM
I've also verified 1/2 off city bribe price for disorder, 1/2 off for empty.

I hope that all theses reduction work together cumulatively in all combinations...

quote:

Originally posted by debeest on 12-28-2000 05:07 PM
Interestingly, my tests last night indicate that a veteran diplomat does not get a city bribe price reduction the way a vet spy does. Go figure. Maybe they thought no one would ever have a veteran diplo.

If 1.8 is true then i propose an explanation: they added Communist spies late in a line of developement of Civ.

1.6 Spy may be lost in some actions (sabotage unit, steal technology, industrial sabotage, incite a revolt, poison water supply). If she is not lost, then she is moved to her nearest city and get veteran status.
1.7 All dips/spies produced in Communism are veterans.
1.8 1.6 and 1.7 are only ways how to get dip/spy with veteran status.

quote:

Originally posted by debeest on 12-28-2000 05:07 PM
And finally, the gold reward you get for capturing or bribing a city is not 1/3 of the enemy's treasury. I haven't been able to figure out what it actually is; I think maybe it represents the fraction of their treasury equal to (citizens in city / total citizens in empire). Check it out.

You're right... 1/3 works for city size 1 only

[This message has been edited by SlowThinker (edited December 28, 2000).]

SlowThinker · December 28, 2000, 20:02

debeest, what mean "longest dimension" and "shortest dimension"?

La Fayette · December 29, 2000, 11:00

Debeest,
The results of the test I ran last night are OK with yours. I suppose that we might consider that: S(T+1000)/integer= cost (for a dip)
is the formula.
All discounts for empty city,...and so on, also seem to be OK.
There remains one point to be tested: exact relationship between integer and distance to the capital.
I got one precise result in 3 different cities with no capital: integer=11
If integer=D+3, as stated by Winkler, then Dmax in case of "no capital" would neither be 32 (as stated by Winkler), nor 16 (as stated by ST), but 8!
IMO some more testing is needed.

ST
The Italians say:
Chi va piano va sano
Chi va sano va lontano
It seems to me that the name you chose is going to lead you a long way ahead. Go on and don't care too much about the quality of your english. As long as you make yourself understood, I suppose it's OK for anyone.
Happy New Year!

SlowThinker · December 29, 2000, 14:15

quote:

Originally posted by La Fayette on 12-29-2000 10:00 AM
There remains one point to be tested: exact relationship between integer and distance to the capital.
I got one precise result in 3 different cities with no capital: integer=11
If integer=D+3, as stated by Winkler, then Dmax in case of "no capital" would neither be 32 (as stated by Winkler), nor 16 (as stated by ST), but 8!

Happy New Year!

La Fayette,
Did you read my post (posted December 28, 2000 17:19)? I think all problems with distance should be clevered:
I consider Civ map as a grid of xy cartesian coordinates, turned by 45 degrees. So my distance is measured north-east direction or sud-east direction or...
I remind: I don't know how to measure distance if both coordinates of city and unit are different!
Joyeux Annee 2001! (Mon francais n'est pas mieux que anglais, peut-etre c'est mal dit...)
[This message has been edited by SlowThinker (edited December 29, 2000).]

debeest · December 29, 2000, 15:58

SlowThinker, sorry, is your dip/spy thread just an ordinary post-to thread? Mea culpa. Please feel free to transfer my commentary to it in any way you can.

I haven't kept track of diplomat tech-theft failure rate, but I'd casually estimate it at around one-half to one-third failure.

My reference to longest dimension and shortest dimension means: count the number of vertical spaces and the number of horizontal spaces between points, as the map itself is oriented; the larger number is the longest dimension, and the smaller number is the shortest dimension, and supposedly you add the largest dimension plus 1/2 of the shortest dimension to get distance. Maybe if I orient myself to a 45-degree rotation as you suggested, everything will become clear to me. I'll try it.

LaFayette, yes, I think the most concise formula for city bribe is:

Cost = Size * (Treasury + 1000) / Distance integer

Maybe we can define the distance soon.

It seems unlikely to me that having no capital at all would set a maximum distance smaller than you can get with a capital.

SlowThinker, I'm fairly confident, though not absolutely certain, that a courthouse doubles cost. That would have the same effect as cutting the distance INTEGER in half. If, instead, it cuts the DISTANCE in half, it would be pretty close to the same thing at long distances, but have little effect at short distances. I guess I'll do a little testing to check this, too.

SlowThinker · December 29, 2000, 16:52

quote:

Originally posted by debeest on 12-29-2000 02:58 PM
SlowThinker, sorry, is your dip/spy thread just an ordinary post-to thread? Mea culpa. Please feel free to transfer my commentary to it in any way you can.

Yes, my idea was to concentrate all discussion about dip/spy there and to clear posts when concrete problem is solved and integrated to the first post (probably more people should have the right to edit first post so that integration is always fast). I think almost all Apolyton could be managed this way and then there wouldnt be so large redundancy. It is terrible to search in Apolyton's threads concrete information sometimes.

quote:

Originally posted by debeest on 12-29-2000 02:58 PM
I haven't kept track of diplomat tech-theft failure rate, but I'd casually estimate it at around one-half to one-third failure.

Look to my thread, new things arose (7.2-7.4).

quote:

Originally posted by debeest on 12-29-2000 02:58 PM
It seems unlikely to me that having no capital at all would set a maximum distance smaller than you can get with a capital.

My opinion is that no capital sets all distances to 16 resp. 10

quote:

Originally posted by debeest on 12-29-2000 02:58 PM
SlowThinker, I'm fairly confident, though not absolutely certain, that a courthouse doubles cost. That would have the same effect as cutting the distance INTEGER in half. If, instead, it cuts the DISTANCE in half, it would be pretty close to the same thing at long distances, but have little effect at short distances. I guess I'll do a little testing to check this, too.

The result of my courthouse test(just finished): First, distance is pruned to Dmax, if needed. Then, distance is halved and rounded down.

[This message has been edited by SlowThinker (edited December 29, 2000).]
[This message has been edited by SlowThinker (edited December 30, 2000).]

William Keenan · December 29, 2000, 20:18

Congradulations Tonic, that was some excellent formula cracking! Bravo!

I have an unknown variable to add to your formula.

At some point late in the game, M changes from 2.05 to 3.1. This change becomes apparent when the cost of bribing barbarian units changes. Musketeers go from 62 to 93, Partisans from 102 to 155, etc.

SlowThinker · December 29, 2000, 21:33

Debeest and La Fayette,
I think that I'v solved the problem how to measure distance for general case:
1. Divide the route from capital to unit to two parts:
straight(adjacent squares connected by lines; for example north-east before rotating) and diagonal (adjacent squares connected by edges; for example up, left, down before rotating).
2.distance=(straight part)+(1.5*diagonal part)
3.round distance down
It is all: use in formula

If it works (i tested briefly), then 4.32 in my thread is solved.

It is equivalent to debeest's

quote:

My reference to longest dimension and shortest dimension means: count the number of vertical spaces and the number of horizontal spaces between points, as the map itself is oriented; the larger number is the longest dimension, and the smaller number is the shortest dimension, and supposedly you add the largest dimension plus 1/2 of the shortest dimension to get distance.

but you must rotate the map by 45 degrees first.

Why both ways are equivalent? Since you may imagine that diagonal part as a hypotenuse of two non-diagonal parts:

A
OO
O_O
O__O
O___O
BOOOOCOOOOOD

A and D are unit and capital
AC is diagonal part (and hypotenuse)
CD is straight part
BD is longer dimension
AB is shorter dimension
AB and BC are "two non-diagonal parts"
_ is nothing
O is the route between A and D

[This message has been edited by SlowThinker (edited December 29, 2000).]

SlowThinker · December 29, 2000, 21:37

William Keenan, look at http://apolyton.net/forums/Forum1/HTML/001678.html?5
we are trying to concentrace dip/spy wisdom there

SlowThinker · December 30, 2000, 20:58

debeest,
Maybe, you could make some bold announcement here and refer people to my thread, so that discussion is concentraced. Or, accordingly to my "Post 2", you could suggest people to post here about strategies and to post facts to my thread.

debeest · January 1, 2001, 03:05

SlowThinker's thread on spies and diplomats (which is not on what I think of as the regular list of Civ 2 strategy topics, but has a link in this thread), is terrific. It's done in a way that allows for feedback and modification within the thread, comprehensively addressing one specific topic in one location, which I think would work very well for the Great Library. Check it out!!

La Fayette · January 8, 2001, 10:42

ST
I remember that I read somewhere that Xin Yu used to measure distances with exactly the same formula (*1,5 for "diagonal squares"). I suppose it would perhaps be a good idea to test *1,414 (diagonal of a square).
But I get bored testing bribery.
I HATE BRIBING.
I swear I am going to play at least 3 games "NO BRIBE" from now on. I leave it to Tonic and debeest and you. Good luck to you all.
(BTW I get repeatedly S*(T+1000)=11 when bribing cities with no capital, but I play version 1.024, perhaps that's why).

------------------
aux bords mystérieux du monde occidental

SlowThinker · January 8, 2001, 19:08

quote:

Originally posted by La Fayette on 01-08-2001 09:42 AM
ST
...the same formula (*1,5 for "diagonal squares"). I suppose it would perhaps be a good idea to test *1,414 (diagonal of a square).

No, it is 1.5 exactly.

debeest · January 10, 2001, 04:15

You're getting bribe costs = S * (T + 1000) * 11 because there's a courthouse in the city, so the maximum distance (16) is cut in half, making 8, before the three is added.