1. We have added a Gift Upgrades feature that allows you to gift an account upgrade to another member, just in time for the holiday season. You can see the gift option when going to the Account Upgrades screen, or on any user profile screen.
    Dismiss Notice

A Rank Corruption Discovery and Exploit to negate rank corruption

Discussion in 'Civ3 Strategy Articles' started by Qitai, Aug 31, 2003.

  1. AlanH

    AlanH Mac addict, php monkey Moderator Hall of Fame Staff Supporter GOTM Staff

    Joined:
    Jan 9, 2003
    Messages:
    29,155
    Gender:
    Male
    Location:
    England
    I'm still trying to get my head around all this, but I*think* it means that I only have to change a few things about my favorite playing style. Does this seem like a plan?

    1. I like keeping my first core tight and having three tile movement between cities to allow rapid shifting of forces. I could never see the enthusiasm for OCS at that stage given the size 12 limit until late in the game. No change needed.

    2. I like to build the FP in my first core. That gets it started earlier and built faster than waiting for a foreign capital to fall vacant and building there rather slowly and/or hoping for a great leader. No change needed.

    3. I like to build a new Palace in a foreign capital once the FP is completed, preferably using a Great Leader. By this time GLs are more likely to arise, and I'm more likely to be in a position to take such a location.

    I can continue to do this, but now I'll try to ensure that it is far enough away from my first core that all my core cities, and any local ones I've captured, are closer to the FP than to the new palace. Sounds like I need to be earlier and more aggressive with long distance/intercontinental wars. But that's true of my game as a whole, anyway. Lighthouse and/or Navigation become more important along with earlier development of the navy.

    [I'd love to claim virtue in avoiding an exploit, but fact is the Great Palace Jump via an abandoned capital just seems too complicated for this bear of little brain. Also, I can seldom resist developing my capital, and I'd hate to destroy or sell off all those nice buildings and have to rebuild it from scratch.]

    4. Instead of keeping the captured cities round the new Palace I should be more prepared to abandon those closest to it so that my FP-centered core has a lot of low rank cities. Up to now I've kept and developed all the captured cities, and given the AI's low density city placements even sticking to that may still not be too bad.

    So maybe my "natural" playing style has actually exploited this bug inadvertently all along? If so the small changes I can make in future games would not be detectable as deliberate.

    @Txurce: You suggest that the FP/Palace concept in Civ3 does not reflect historical development. I would beg to differ. Many civilizations have relocated their capitals at some point, sometimes more than once, usually retaining the ancient capital city as a cultural or religious center. The main inaccuracy in Civ3 is using the modern-day capital name for the 4000 BC capital.
     
  2. Grille

    Grille panel insect

    Joined:
    Dec 19, 2002
    Messages:
    2,709
    Location:
    Kiel, Germany
    That's exactly why there's this grey area about this bug. You may already have unintentionally gained an advantage in some game, most likely when there was a more or less dense built in the initial core and your palace moved to a distant stand-alone city. Then, there'd be a noticed income boost which would not make any sense in the context the rank/distance system was formerly assumed to work.
     
  3. rabies

    rabies Happy Dad

    Joined:
    Nov 9, 2002
    Messages:
    268
    Location:
    Maryland, USA
    Dianthus/Qitai,
    Thanks for the explanation! Thanks to you, I now know what the fuss is all about. Pretty bad. I won't even bother trying to comprehend why.
     
  4. AlanH

    AlanH Mac addict, php monkey Moderator Hall of Fame Staff Supporter GOTM Staff

    Joined:
    Jan 9, 2003
    Messages:
    29,155
    Gender:
    Male
    Location:
    England
    I haven't played many games, but I found this approach worked for me for all the stated reasons. Qitai's discovery simply adds another great reason to use it. In practice the income boost would have been unnoticed because I would have already started to re-grow the second core around the new palace site, after the usual starvation process to minimise foreigners, and my development attention was focused there.

    I never made a conscious effort to note the change of income in the old core when I built my new palace, as I expected the FP to maintain performance there. I was just grateful for the new low corruption and high productivity in the new core, and got on with business.
     
  5. Svar

    Svar Chieftain

    Joined:
    Jun 10, 2003
    Messages:
    615
    Location:
    China Lake, Ca
    Here are the results of my test.

    The test was conducted on a huge map with an isolated small continent where I could build ring cities out about distance 11 from the palace. The palace city has all the resources needed to play the game to the end. All other tiles on the continent are bonus grassland with cows bordered with rivers. Every city works exactly the same type of tiles with the same population so should always get the same productivity if it is the same distance from the palace. This is what always happens so all I had to do was take one of these test cases for a palace/FP switch.

    The test case was a map that had two rings. One ring had 6 cities at distance 5 and another ring of 12 cities at distance 9. The test map also has two other large continents on it and on one of these I placed a new city labeled New Palace. New palace is over 50 tiles from the closest city in the outer ring. There are also 4 coast cities but even though they are up to 14.5 distance from the palace I decided not to use them in the test because many of their tiles are not bonus grassland with cows.

    The first test condition will be with the ring cities having the following improvements: marketplace, courthouse, bank, and factory. I only list those improvements that will affect the production of shields and gold. The population of the first test condition will be 12 and I will present data from both the palace center and the FP center as raw data and modified data. The raw data will only include the effects of the courthouse with the entertainment slider set to 100%. The modified data will include all the effects of all the listed improvements with the tax set to 100%.

    The 6 ring 5 cities show exactly the same results regardless of whether the palace or the FP is in the ring center. It is in the 9 ring cities where you see the difference.

    In the raw data the shield production for a palace centered 9 ring city is 51 with a loss of 5 for an efficiency of 90.2%. When you replace the palace with the FP the shield production is still 51 but the loss is only 2 for an efficiency of 96.1%. That is an improvement of almost 6%. When you include the effects of the city improvements the palace centered city produces 74 shields with the same loss of 5 for an efficiency of 93.2%. The FP centered city produces 75 shields with the same loss of 2 for an efficiency of 97.3%. Now the improvement is only about 4%.

    Also in the raw data the commerce generation for a palace centered 9 ring city is 42 gold with a loss of 6 for an efficiency of 85.7%. When you replace the palace with the FP the commerce generated is 42 gold with a loss of 3 for an efficiency of 92.9%. That is an improvement of over 7%. When you include the effects of the city improvements and the 100% tax rate the palace centered city generates 78 gold with a loss of 6 for an efficiency of 92.3%. The FP centered city generates 81 gold with a loss fo 3 for an efficiency of 96.3%. Now the improvement is only 4%.

    The other test condition will be with the ring cities having a population of 20 and the following improvements: marketplace, courthouse, bank, factory, hydro plant, police station, and stock exchange.

    In the raw data the shield production for a palace centered 9 ring city is 81 with a loss of 6 for an efficiency of 90.8%. When you replace the palace with the FP the shield production is still 81 but the loss is only 2 for an efficiency of 97.5%. That is an improvement of almost 7%. When you include the effects of the city improvements the palace centered city produces 156 shields with the same loss of 6 for an efficiency of 96.2%. The FP centered city produces 160 shields with the same loss of 2 for an efficiency of 98.8%. Now the improvement is less than 3%.

    Also in the raw data the commerce generation for a palace centered 9 ring city is 65 gold with a loss of 6 for an efficiency of 90.8%. When you replace the palace with the FP the commerce generated is 65 gold with a loss of 3 for an efficiency of 95.4%. That is an improvement of less than 5%. When you include the effects of the city improvements and the 100% tax rate the palace centered city generates 153 gold with a loss of 6 for an efficiency of 96.1%. The FP centered city generates 158 gold with a loss fo 3 for an efficiency of 98.1%. Now the improvement is only 2%

    One odd thing that surfaced in these tests was the fact that distance affects the calculation for commerce but had no effect of shield production when the FP is the centered city and there are no cities closer to the palace. In other words the shield production and losses were the same for all cities centered on the FP no matter if they were in the 5 ring or the 9 ring.

    My conclusions are the effects are more significient if you play the game with little on no improvements to your cities. If you are a builder who builds most of the improvements in your core cities the effects are minimal no matter how you arrange your cities.
     
  6. Qitai

    Qitai .

    Joined:
    Jan 22, 2002
    Messages:
    1,177
    Location:
    SG.MY.TW.US
    Svar. For the effect of the improvements on corruption, please refer to Alex's thread on corruption. His formulas clearly explain the effect of the improvement on corruption.

    As for your observation, when you use the exploit, rank corruption is minimized or constant. Distance corruption can be drastically reduce with government types, improvements and WLTKD. The effects of all these in combination can make distance corruption almost neligible when the distance is short (i.e. in your test case, distance 5 and 9). So, esstentially, what you see is that the 2 corruption is probably due to rank corruption only. I cannot verify your figures since you did not specify the government type and difficulty level.
     
  7. DaveMcW

    DaveMcW Chieftain

    Joined:
    Oct 8, 2002
    Messages:
    6,489
    This is from the Conquests beta test channel. So they will at least be fixing it in the expansion.

    <Ed-Firaxis> good news... Ring build, palace corruption exploit, and mobilization exploit fixed
    <Ed-Firaxis> Yes
    <Ed-Firaxis> I Suxor
    <Friedrich> PM me a link to an explanation of it?
    <Chieftess> There's a couple of threads on CFC, but the search has been disabled.
    <Ed-Firaxis> Basicly if you have a forbidden palace, with a city closer to it then the capital it determines its rank by.....
    <Ed-Firaxis> wait.....
    <Ed-Firaxis> this is hard to explain
    <Friedrich> LOL
    <Friedrich> No need, Ed, just say "it is fixed"
    <Chieftess> But, I have wondered why corruption seemed to do strange things at times
    <Ed-Firaxis> it is fixed
    <Ed-Firaxis> The dude that found this has way to much free time :)
     
  8. Svar

    Svar Chieftain

    Joined:
    Jun 10, 2003
    Messages:
    615
    Location:
    China Lake, Ca
    I knew I left something out. From an earlier post "I play all test city placements as Carthaginians at Monarch difficulty level and conquest as the only victory condition so I really don't have to pay any attention to the other 7 civilizations."

    Also the government type was always Republic and the cities were always in WLTKD. I hope this clarifies the test better. You are welcome to the 4 save games that were used to generate this data.
     
  9. anarres

    anarres anarchist revolutionary

    Joined:
    Apr 22, 2002
    Messages:
    6,069
    Location:
    www.civ3duelzone.com
    That's pretty amazing that they've fixed RCP! I wonder how they will cope with lots of cities all at the same distance - I can't see how they could fix it without changing from integer calculations. :confused:

    I *hope* the mobilisation exploit refered to is the new one and not just the old one (i.e. a fix to mob as it is in 1.27, not just the 1.27 fix).
     
  10. Qitai

    Qitai .

    Joined:
    Jan 22, 2002
    Messages:
    1,177
    Location:
    SG.MY.TW.US
    LOL. I just spend 1 hour to find this out, although this troubled me for a long time. Though, I have to say I was damn lucky to just have tested it out in a way which makes it quite obvious to suspect the result.

    I did spend quite a bit of time creating my own corruption calculator with Alex's formula though.

    Anyway, glad to hear they are going to fix this. :)
     
  11. AlanH

    AlanH Mac addict, php monkey Moderator Hall of Fame Staff Supporter GOTM Staff

    Joined:
    Jan 9, 2003
    Messages:
    29,155
    Gender:
    Male
    Location:
    England
    They would probably assign the same rank value for corruption/waste to all those equidistant cities and make it the average corruption/waste value for the range of ranks n, n+1, n+2, n+3 ......
     
  12. Plux

    Plux Emperor

    Joined:
    Mar 3, 2002
    Messages:
    373
    Location:
    the Netherlands
    I don't think he means that RCP is fixed. I don't see that possible without introducing (a lot)more variables to make pure distance be less important.

    I think only the far-away palace build exploit is fixed, maybe by introducing two distance rank counters, one for palace and one for FP, as was initially supposed by alexman (?) to be the case.
     
  13. AlanH

    AlanH Mac addict, php monkey Moderator Hall of Fame Staff Supporter GOTM Staff

    Joined:
    Jan 9, 2003
    Messages:
    29,155
    Gender:
    Male
    Location:
    England
    Why?

    If you have four cities all at the same distance, with four cities closer to the palace then if you just look at Fc in alexman's formula you can calculate four values of Fc with Ncity = 5, 6, 7, 8, add them together and divide by 4. Use the resulting average value as Fc in the Corruption% calculation for each of the four cities. Sure, there are rounding errors in the averaging, but so what? Civ3 is full of rounding errors.

    If the sequence of Ncity values straddles the Fn*Nopt threshold then that just modifies the separate Fc values either side of the threshold before they are averaged.

     
  14. rabies

    rabies Happy Dad

    Joined:
    Nov 9, 2002
    Messages:
    268
    Location:
    Maryland, USA
    Having it get fixed in the expansion will be great...but I wonder what the result will be for the GOTM? The GOTM is all about letting people play whatever version they want...clearly, people using the expansion (when it comes out) will be at a disadvantage to those who decide to use it (even in the slightest manner) in PTW and vanilla.
     
  15. AlanH

    AlanH Mac addict, php monkey Moderator Hall of Fame Staff Supporter GOTM Staff

    Joined:
    Jan 9, 2003
    Messages:
    29,155
    Gender:
    Male
    Location:
    England
    It seems to me there will be a strong incentive to move GOTM towards Conquests-only at some point in the future, otherwise it will get left behind in terms of the majority player population. Not to mention that it removes the need to police exploits using these bugs.

    1.29 is already "obsolete" as far as Atari/Firaxis is concerned. PTW is likely to go the same way once Conquests is out and selling. Cracker and the team have done wonders to allow 1.29 and PTW players equal access, but there's no way they are going to be able to create a level playing field across all three platforms in the likely event that these bug fixes are not retro-fitted.

    My concern would be that the Conquests upgrade could become a price of admission to GOTM at some point, and that Conquests may not even be an option for Mac players.

    WARNING: The above is all purely personal speculation. It might never happen. I'm sure the GOTM team are going to take a long hard look at Conquests and test it thoroughly before they ever allow it as a competition option.
     
  16. Plux

    Plux Emperor

    Joined:
    Mar 3, 2002
    Messages:
    373
    Location:
    the Netherlands
    So you're saying that cities with the same distance should get an average rank instead of all getting the highest rank?? I guess that should straighten things out better, yes :)
     
  17. Cartouche Bee

    Cartouche Bee Appropriations Consultant

    Joined:
    Jan 20, 2002
    Messages:
    1,731
    Location:
    Multiple Victory Avenue
    Cities that are the same distance should be ranked by the date they were founded.
     
  18. AlanH

    AlanH Mac addict, php monkey Moderator Hall of Fame Staff Supporter GOTM Staff

    Joined:
    Jan 9, 2003
    Messages:
    29,155
    Gender:
    Male
    Location:
    England
    There's another easy solution. Nice one, CB.
     
  19. ptclouds

    ptclouds Chieftain

    Joined:
    Sep 7, 2003
    Messages:
    8
    Location:
    Sydney, Australia
    Bizarre.

    I accidentally found this "exploit", although not in such detail. And I haven't even played a game to its conclusion yet. I was moving my palace all over the place, and it seemed to pay dividends. Its such an obvious thing to do. Build up your improvements, increase culture value, then move on.

    Now I wonder if I shall ever bother playing again. It seems to make things too easy.

    Only joking. I might just buy the latest version instead. Gonna cost me though.

    This ring city thing, does anyone actually use it? It seems too much hassle to even bother. Plus what happens if you are set on the coast to start with, does it work then too? Or there are mountains where you want to set your cities?
     
  20. Roland Johansen

    Roland Johansen Chieftain

    Joined:
    Apr 29, 2003
    Messages:
    4,292
    Location:
    the Netherlands
    Then it wouldn't take people here long to find out that founding cities in the same year using RCP would give the old advantage of RCP. ;) Maybe the order in which cities were founded if this is also registered by the game somehow.

    I'm happy that I modded corruption a long time ago so that I don't have these problems.
     

Share This Page