Palace Corruption Exploit

Originally posted by AlanH
... To your previous point about this being six months late: This rule would only have been required six months ago if earlier games had invited the use of this exploit, or if players had demonstrated that they needed such guidance. ...

As I recall in GOTM26 Bremp did it and honestly confessed and some other players did it in GOTM24 though did not benefit substantially to influence the score. However, indeed in GOTM23 and 25 there was no real benefit of doing this.

This is more of a problem of starting location. Or whether it is centered towards the major part of the estate. It was more or less centered in GOTM25 and 23 and even 24 and 26. It was not centered in GOTM22 (that was apparently how it was discovered) and 27. That is a clear reason to move the Palace. And why not move it so that the benefit is maximal... Lets imagine Europe which has been conquered by Portugal with Palace in Lisbon or by Germany with Palace in Munich. Certainly, temptation to move the Palace in the first case is very high and it might be allowed. But move it to Munich, not to Iceland. Moving to Iceland is an exploit and this is what is actually in the spirit of the "red rule". Because it makes no sense why your empire gets less corruption with Palace in Iceland than in Munich.
 
Originally posted by Justus II
Maybe I'm missing the point, but I don't think so. Any palace move will improve corruption near the new location. As I understand it that is a perfectly acceptable reason for moving the palace, to create a new productive area, assuming that you have or build a new group of cities to take advantage of that. ...

Certainly it all makes sense but not with the FP/rank corruption exploit. Having good intentions does not make a player less guilty. Easy or difficult, it is FP that has to be moved not the Palace. And probably the wording you suggested is better than what it is now although it is unclear how to evaluate the actual game. The only real solution is to design the gotm maps so that starting location is more or less central at least during early expansion. It does not sound good at all.

Another solution is switch to C3C or have some scoring/evaluation system compensating for differences between C3C and PTW/Civ3. Or just different player ranking. It might take a while to do so and by the time the whole thing is settled down, it is quite possible that Civ4 will be around to play with.
 
Originally posted by Justus II


Maybe I'm missing the point, but I don't think so. Any palace move will improve corruption near the new location. As I understand it that is a perfectly acceptable reason for moving the palace, to create a new productive area, assuming that you have or build a new group of cities to take advantage of that. The bug, and exploit, is a result of the effect on corruption in the "old" palace area, which presumably now has the FP. Moving the Palace far away suddenly reduces corruption near the FP, when it should not have any effect on it. The exploit is moving the palace where it would reduce corruption in the original core, rather than to start a new productive core. Perhaps the wording should be adjusted to reflect that. "remote from the majority of your empire in order to gain a significant corruption advantage in your previous holdings" or something like that. And I do think that is the legitimate use for the Palace, since the FP cannot be readily moved.

Qitai discovered that there is a bug in the rank calculations for cities closer to the Forbidden Palace than the Capital. In short, for such a city, the rank is given by the number of cities that are closer to the Capital than that city is to the Forbidden Palace.
The Exploit
To exploit the above, basically, all you need to do is to put your palace as far away as possible without any of your own cities near it. This way, all your cities near the FP will be considered as Rank 1! Try this on a recent game you finish by abandoning all the cities near the palace and see how your corruptions near the FP reduces like magic(You may need to remove and put back the population to see the changes). So if the nearest city to the capital is distance 20. Then all cities less than 20 distance from the FP is considered to have a rank of 1. Imaging that!! Even those cities that are considerable distance from the FP now become very useful. I tested it on a recently finish game and I see my previously fully corrupted cities in the outer ring of my FP turning into low corruption power cities.

In short: If you move your capital somewhere else within your empire, to (for example) reduce the likelihood of culture flips recently conquered cities, or to reduce corruption in cities in that part of your empire, then that is fine: You are moving the capital to another area within your empire. You are moving it close to where other cities are!

The most common implimentation of this exploit would require a great leader, generated by combat. If you are invading a far-off island and you capture a city and get a great leader, you may choose to rush a palace in the newly conquered territories. This may, or may not be exploitive.

If the far-off island was tiny and could only support a couple of cities, then I suggest that this is exploitive. If the island is large, but you only bother to capture one or two cities and never develop a core of cities there, then that would be exploitive. If the continent is large and you capture/settle a reasonable number of cities, (which I would as a rough rule presume to be at a bare minim OCN/2), then this is NOT exploitive.
 
Has anybody tested the effect of taking the palace from a highly productive core of cities when they also have a highly productive core of cities around their FP and moving the palace to some remote location that increases the prductivity around their FP? What will be the overall effect on the whole empire as the old highly productive core around the old palace drops to highly corrupted?
 
Originally posted by Svar
What will be the overall effect on the whole empire as the old highly productive core around the old palace drops to highly corrupted?

I believe that is the bug, what should happen does not. And then, the cities around the new palace become productive thus trigger the exploit. Correct me if I'm wrong.

-Layman ;)
 
Originally posted by Svar
Has anybody tested the effect of taking the palace from a highly productive core of cities when they also have a highly productive core of cities around their FP and moving the palace to some remote location that increases the prductivity around their FP? What will be the overall effect on the whole empire as the old highly productive core around the old palace drops to highly corrupted?

That indeed is the problem, the "old highly productive core around the old palace" does NOT become highly corrupted. As long as your old core is closer to the FP than the new palace, (even if it is 10-20 or more tiles away) corruption improves. The only cities adversely affected are the capital itself, and cities immediately nearby (former Rank 1 cities) and they can easily build a courthouse prior to the jump, if needed.

As an example, I created the following spreadsheet, that uses common references (I tried to use the former capitals, as they would be the same location for everyone, except York which was the closest to an equidistant location, 3 tiles SE of London).
This gives corruption/waste data from 1010AD, just prior to rushing the Palace on the Oda continent, and 1020AD after. Note that the effect would have been more dramatic if Washington had originally been more centrally located, but since I settled Washington near the lambs on the East coast, most cities were already closer to the FP than the palace. As I stated earlier, this palace jump was not intended to take advantage of the exploit, as I wanted to develop the Oda lands, and had 10 cities there that I captured from the Oda, and I eventually built 3-4 more there, but it does give an example of how it affects the old core.


CorruptCalc.jpg


You can see that corruption dropped everywhere except the former first-ring cities, and with a courthouse even they get down below 15% corrupt. (Washington was 1 turn from a courthouse, and I think I rushed one in New York after the jump). It makes sense that the Oda should jump up in productivity, but cities near the old core (Madrid, York, even Nidaros) should get worse, not better, but they don't. You can also see the overall effect on the empire, cutting the corruption rate nearly in half, and a similar effect on production (sorry, I didn't count all the sheilds!).
 
I fear I am going to open a another hot topic. But for me RCP is just as exploitative as Palace move in terms of corruption. It is why is has been (badly) patched in C3C. So in the same topic. Should we ban RCP ?

My view are : We all know we can do this. It is up to players to use it or not... Banning RCP would be banning AI as I have seen AI using it (Spanish in GOTM 27)

In conclusion, I would allow Palace move & RCP
 
I see your point CdB, but i disagree, i dont think RCP should be banned. RCP is a well known stragtegy used a lot (which the rank-corruption-exploit is not). Also the latter is a lot more powerful and really unbalances the game. RCP give u 5-6 highly productive core cities, the rank-corruption-exploit might give u 20-30 (or even more) highly productive cities.
 
Originally posted by Ainwood

Many of you are aware of an exploit in PTW and vanilla Civ, where the corruption of your empire can be greatly decreased if you move your palace/capital to a remote location, well away from your core empire.

There have been a few cases where this has been used in GOTM games. The staff have discussed it, and agreed that it is outside the spirit of the game, offering an massively unfair advantage to those that use it.

We therefore have decided to ban it.

I think this was a good decision. :)

I must admit that I've used this exploit in GOTM26, but it was part of my strategy and I don't fell bad about it now.
 
@Salte
Point Taken

Nevertheless, I think the rule "No Palace Move" is too harsh
 
I don't think the rule IS "No Palace Move", but "No Palace Move to some remote location whose only purpose is to take advantage of the Rank Corruption Exploit". See in this post from Ainwood some examples of allowed and not allowed moves.

Originally posted by ainwood


In short: If you move your capital somewhere else within your empire, to (for example) reduce the likelihood of culture flips recently conquered cities, or to reduce corruption in cities in that part of your empire, then that is fine: You are moving the capital to another area within your empire. You are moving it close to where other cities are!

The most common implimentation of this exploit would require a great leader, generated by combat. If you are invading a far-off island and you capture a city and get a great leader, you may choose to rush a palace in the newly conquered territories. This may, or may not be exploitive.

If the far-off island was tiny and could only support a couple of cities, then I suggest that this is exploitive. If the island is large, but you only bother to capture one or two cities and never develop a core of cities there, then that would be exploitive. If the continent is large and you capture/settle a reasonable number of cities, (which I would as a rough rule presume to be at a bare minim OCN/2), then this is NOT exploitive.
 
Originally posted by CdB
I fear I am going to open a another hot topic. But for me RCP is just as exploitative as Palace move in terms of corruption.

Banning RCP would be banning AI as I have seen AI using it (Spanish in GOTM 27)

I've had similar thoughts regarding RCP as being just as exploitative, mainly because it is easily done and more common. I also know that Cracker took at least some effort to try and make RCP around the starting position less than ideal through placement of mountains, etc. I would like to suggest that in GOTM27, the opposite was done - placing AI starts surrounded by "best settlement" tiles in an RCP to better emulate the same advantage that the players can take advantage of. This could be the prime reason why the Spaniards were such tough foes early on.

If you don't believe that the AI works this way, take time to review different players' minimaps of the available QSCs. You'll notice that the AI settles in the same places in each game. It turns out that AI settlement is controlled by best potential output for a tile. Granted that the AI inherently knows this even though it is "under the fog", but the algorithm must include a distance factor as well to make settlement logical. I stumbled upon it by accident in exactly this fashion, and had a short conversation with Cracker about it.

In short - the AI is not programmed to use RCP. It was made to use it by tight control of the terrain surrounding their start positions. I would only be convinced otherwise if you could prove that the AI didn't settle at start, moving first, so as to get a better RCP.

Banning RCP would be counterproductive to keeping players involved in GOTM. As stated by a previous writer, RCP is an exploit, but without great effort the effect will be small compared to the palace jump exploit. Equally, banning palace moves could also drive people out, and is counterintuitive. I think that Ainwood, Alan, etc. have a good handle on it.
 
GOTM is competition and there are judges. Like in every game. To make the competition fare, the rules have to be clear. What has been phrased as the "red rule" is obviously not clear enough. The explanation of ainwood is also obscure enough. It is not possible to consider the goal of a player or intention of a player when moving Palace. And intentions are not relevant. Only the result counts like in any other game. So the rule has to be clarified in terms of OCN and distance from Palace/FP. For example, number of cities located within a certain radius from new Palace should not be less than OCN/2 or the number of cities located within the same radius from FP. Radius can be defined depending on the map size or a fraction of the area corresponding to each player. For example, radius=(MapHeight+MapWidth)/8 which is a very reasonable number. Or, radius=square root of (total map area/number of players starting the game) where total map area is a product of MapHeight and MapWidth. It is just a matter of taste because values should be very similar for most cases. Even with OCP, this area should be enough to put OCN/2 cities in it and leave some space for more cities. It is complicated and certainly requires both OCN and map size tabulated before the start of the game. Hmm, the player should also be able to use a calculator, and pen, and piece of paper, or Excel. Otherwise, it cannot be considered as a rule just a vague guideline which cannot be used to throw players from the scoring list. The use of Palace jump can be only qualified as "not so honorable gameplay style" which does belong to GOTM by default. Because if one wants to play an "honorable gameplay style" then it is the RBC code. Otherwise, it might be possible to ban RoP rape, Peace treaty violation, resource connection/disconnection, and many-many other tricks used by the players against AI. On the other hand, Palace jump is probably worse than worker dogpiling and resource denial in terms of global effect on the gameplay, so it should be banned. The problem is how to keep the Palace jump in the "allowed" list. Just for reminder, the GOTM rules are listed here.

For me personally, there is no problem at all. If a player wants to do these "not so honorable" things and is doing them and I do not do, and that other player has a higher score because of this, it is fine with me. Well, if I want to do them also, I can. But if I don't want, its my problem. And this is just a game and not even for money and not with a real name.
 
Originally posted by RowAndLive

In short - the AI is not programmed to use RCP. It was made to use it by tight control of the terrain surrounding their start positions. I would only be convinced otherwise if you could prove that the AI didn't settle at start, moving first, so as to get a better RCP.
I doubt the AI considers an initial settler move at all, regardless of its city placement programming. Has anyone *ever* seen the AI move its settler before founding the capital? I'm quite resigned to seeing that early score/power discrepency in the histograms as all the ai civs found at 4000 BC if I move and start scoring one turn later. If it did ever move its first settler I think it would be difficult to prove it did so specifically for RCP reasons.
 
@AlanH, I'm not concerned _why_ an AI starting settler might move before founding, as I'm quite thoroughly convinced that it doesn't / can't happen unless the scenario were set up so that all the start spots were on unbuildable terrain. As such, there is no way that the AI can be programmed for RCP. Aside from the fact that the game code was written long before RCP was accidentally discovered. Any apparent RCP by the AI is incidental from the terrain set-up of the scenario.

@akots, just to clarify, please define RBC, OCN and OCP. Also, if investigated, I think that you might find many specific tools used by the players here are exploits of the way the code works. An excellent example is using multiple ships to transport troops long distances over water simply by transferring them between ships. You've given other examples above: ROP rape, RCP, resource denial (during peacetime). I don't think of resource connection / disconnection in this way, although I'd guess that the designers didn't think of people using their own troops for this purpose. Then again, I don't use that tool either.
 
Originally posted by RowAndLive
... just to clarify, please define RBC, OCN and OCP. ...

OCN is Optimal City Number as found in the editor. OCN is used for rank corruption calculation. It depends on map size (not map area) and difficulty level. Rank of each city determines rank corruption. The higher is the rank, the higher is rank corruption. On the other hand, the higher is OCN, the lower rank corruption is. Well, you can read everything in alexman's post. Palace jump exploit was discovered by Qitai and details you can find in another thread. In C3C RCP and Palace jump exploit are eliminated as described by alexman.

OCP is optimal city placement strategy. AI usually tries to use OCP. OCP means that the number of overlapping tiles which belong within worked borders of several cities is minimized.

Finally, RBC is Realms Beyond Civilization community with very noble gameplay rules strictly forbidding any dishonorable tactics/strategy. It is so much more difficult to play by the RBC rules...
 
Originally posted by akots
To make the competition fare, the rules have to be clear. What has been phrased as the "red rule" is obviously not clear enough. The explanation of ainwood is also obscure enough. It is not possible to consider the goal of a player or intention of a player when moving Palace. And intentions are not relevant. Only the result counts like in any other game. So the rule has to be clarified in terms of OCN and distance from Palace/FP.

akots, perhaps the rule could be more clear. However, it is extremely unlikely that it will be clarified in terms of OCN and distance, because it turns the game into an equation that takes the fun out of the game, and most players will refuse to use.

As a result, we will have to accept that this is a judgment call, and as such will be made very conservatively. As soon as even a few cities are placed near the new palace, the power of the exploit is dramatically reduced. To truly benefit from this exploit, the palace has to be placed in a location that, in my mind, leaves no doubt as to the player's exploitive intention. To use a recent example, putting the palace all by itself on the barb island in GOTM27 is exploitive, because it helps only by isolating the palace, rather than creating the center of a new core.
 
Brace yourselves, here comes another on my long-winded posts :lol:

Power of the Rank Exploit

First, some comments about the power of the Palace rank bug. I consider this exploit to be in a league of its own. When deliberately exploited, on a map where it can have a lot of impact, its effect is hugely greater than RCP or any other semi-exploit mentioned in this thread.

As an example of how powerful this exploit can be note that the current top three Deity games in the HOF (click here) are all based on the exploit. Moonsinger and I each reached over 3 million gold. Moonsinger set the income record, making over 16,000gpt at the end of her game :) This kind of result is obviously not quite right :lol:

All three of those games (SirPleb, Moonsinger, Mazarin) have writeups in the HOF discussion area. If you want to know about the mechanics of exploiting the rank bug, my writeup here is fairly detailed.

Minor vs. Major Effects

The bug can have a wide range of impact depending on how heavily it is used. At one extreme when used on an appropriate map and fully exploited, the impact is huge. At the other extreme, the bug may have only a slight effect, depending just on a small difference in city density around the Palace and around the FP.

At the small effect extreme note that the bug can even have a negative impact. Some games have undoubtedly suffered an accidental penalty due to the bug, getting worse city rank numbers in their FP region than they should have.

Impact on GOTM Play

Overall I think this bug has had little impact on GOTM play so far. The majority of players have avoided thinking about it (and that's a good thing! :) ) Most players have unknowingly gained or lost a bit due to the bug but at this level (non deliberate use) the effect of the bug is, I think, in the same league as RCP. I.e. it can have a noticeable effect but not greatly imbalancing. Good play without gaining an advantage from it will win over mediocre play with the advantage every time. The effect is large enough that it would be nice if we could get rid of it. But we can't. And the effect is not so large that it wrecks the competition. So we can just live with it.

For some players, particularly the top scorers, the effect of this bug has already been stronger. When I first understood this bug, I found that I couldn't bring myself to build an FP at all. I felt that I'd be watching every town's placement to avoid unfairly gaining or losing a bit due to the bug and it would take the fun out of the game. So I played one GOTM without an FP at all. Then I played for a while by deliberately underusing my FPs to avoid the bug. In the meantime I played that HOF game where I flagrantly exploited the bug and that got it out of my system :) After that I was finally able to ignore it again and play normally (but making sure never to gain much from the bug.)

I think that the bug may have been a contributing factor to losing other player(s) from GOTM. Maybe there are other factors too, but we haven't seen Qitai here much since he discovered it. It sure seems sad that this bug may have been a Pandora's box for its discoverer, who did an admirable job in chasing down some anomalies in the game and publishing the results instead of using them for personal gain.

When Should a Game be Disallowed?

I think the GOTM staff have taken an excellent approach to the problem in this thread.

Although it can be argued that a very specific rule is required, a rule which can be measured in black and white for any particular game, I think that:

1) Such a rule is almost impossible to define.

2) Even if a specific technical cutoff rule could be defined, it would be bad for the spirit of GOTM to do so. The point here should not be to draw a line which many players would then try to get as close to as possible. It should be to establish the spirit of the game, and thus to discourage people from even thinking in those terms.

3) In practice there isn't a lot of grey area which matters. When the bug is exploited for large gain, it will be fairly obvious that the player set out to exploit it. If it is exploited for small gain (roughly equivalent to RCP gain) we can't distinguish that from some kinds of normal play. But at this level, it isn't earth-shattering. At this level it is like cheating other GOTM rules where cheats can't be detected. It may allow someone to gain a few steps up in the standings but it won't gain them a lot. The vast majority of people here won't sacrifice the spirit of the game and their integrity to gain a few steps in the rankings. So as regards minor use of this exploit, I'd say we don't need to worry about the grey area, just assume that minor gains from the bug are side effects of legitimate play, as they almost always will be.

So it seems to me that the rule the GOTM staff have worked out is very good. (Well done AlanH :) ) If in their judgement a particular game has used the exploit deliberately, that's that, the game gets excluded. I don't think anyone needs to worry about being falsely accused. Non-deliberate gains from the bug won't be mistaken for exploiting it for large gains.

Personally, I'd be more ruthless than the staff have been so far. I'd retroactively ban games which have exploited it in the past, at least in the unreported games GOTM26 and GOTM27. If anyone has used it in those games they might of course be upset by such a ban. My reply to them would be that I don't believe they had any doubt when they used it that they were using an exploit. They've taken an approach of doing something which we all know is wrong (using an exploit for advantage in competition) and gambling on asking for forgiveness afterward instead of asking for permission in advance. This is supposed to be a friendly level playing field competition - most players here are careful to ask permission in advance for actions they think might be questionable. So why reward the players who do the opposite? (My apologies to those players, particularly Bremp, for being so harsh here. But I stand by what I'm saying. To Bremp in particular: You play so well that I'm sad sad to see you using this for competetive advantage. Finding that you will use in competition something which you know is an exploit and which you know your competitors have avoided using, undermines your standings in my view.)

I see this exploit as a case where it makes sense for GOTM and HOF to take different paths. GOTM is (I think :) ) about being a friendly competition where many kinds of players compete in a "normal" Civilization III environment and discuss their approaches to a comparable game. HOF is for extreme games which don't allow for many different approaches - there's a focus there on particular kinds of maps and on score as an ultimate goal.

The Future of Moving the Palace

It is interesting (and sad) to note that moving one's Palace will almost always be a bad thing to do in Conquests. At least, that's how it stands with the 1.15 patch.

A few people have mentioned legitimate uses for moving the Palace in this thread. E.g. strengthening one's cultural hold in new regions while taking them over. It looks like this technique will be a thing of the past when we eventually get to Conquests.

In the meantime, I think it is great that the GOTM staff to continue to allow those techniques here. They could have made an easier decision - the rank exploit's power would be greatly reduced just by saying that moving the Palace is not allowed. The decision they've made leaves the game richer, at the expense of more work on their part to ensure that the rules are followed.

I do think that if this decision creates too much work for the staff, e.g. if a lot of people were to start pushing the spirit of the rule, changing the rule in future to say "no moving the Palace" would be an acceptable alternative. Because that's what we'll probably be faced with in the long run anyway in Conquests.
 
Originally posted by Txurce
... As a result, we will have to accept that this is a judgment call, and as such will be made very conservatively. ...

There is little doubt that every player will get fair judgement from you and ainwood or other current gotm staff. However, the rules are important and not the discretion of the judge.

Two stupid questions... First, why did you decide to do it now? Does it have to do something with current GOTM? Second, what comes first, Palace jump or core of cities around it? It makes little difference indeed unless these events (Palace jump and core build) are 20-50 turns apart. But just curious of your and ainwood's opinion on the matter.

Originally posted by SirPleb
Brace yourselves, here comes another on my long-winded posts ...

Not so long and brilliant as usual! The only thing that I would disagree is evaluation of Bremp's action in GOTM26. Bremp did what was allowed. And if you and others did not do it, that is not a problem of Bremp. For example, many players never connect/disconnect resources every turn. Or never browse through cities in between turns or go somewhere from "What's the big picture". It might be, some of them don't know how to do this but for many who know it just does not feel right. Well, who can tell right from wrong ...

So, lets try play the new GOTM, and move these Palaces while it is still possible! But do it by the spirit of fair competition.
 
Originally posted by akots


First, why did you decide to do it now? Does it have to do something with current GOTM? Second, what comes first, Palace jump or core of cities around it? It makes little difference indeed unless these events (Palace jump and core build) are 20-50 turns apart. But just curious of your and ainwood's opinion on the matter.

akots, this came up now because some players made it clear in their posts that they were using the exploit solely as an exploit. Until then, it did not seem to be an issue that needed addressing, because everyone seemed to be avoiding the exploit.

As to your second question, I agree that it makes little difference which comes first, unless the cities follow the palace (too) many turns later. In normal game play, by the time a palace shift is considered, there aren't too many empty areas left - which is another example of why this exploit is easy to spot, when it is being used strictly as an exploit.
 
Back
Top Bottom