Well this seems to be getting long. I was originally going to only mention one or two rather than a grand survey. Hopefully, one or two of the ideas here are useful. If I made a factual error, please post a correction and I will try to fix this post.
Links that I used as reference in an attempt to be more accurate:
Realms beyond exploit list
gotm exploit list
==================================
One thing I am hoping the civ team can do for the expansion i
s fixing of some of the exploits that occur in the game. Some are hard to fix and some it is really hard to determine if they are strategy or exploit.
On the other hand, I have real hopes for some of the seemingly easier ones:
One exploit that I find that bothers me is the fact that the AI will make a deal where it takes gold/turn when it plans to attack on the very next turn. I am not sure why it bugs me - perhaps its simply because it is hard not to take advantage of it, but when I do it feels dirty. I would imagine this one is not too hard to fix. Easiest would be to not make per turn deals with someone the AI has recently decided it is going to declare war on. More complex would be to base the decision to attack on your outstanding deals.
--------------------------------------------------------------------------
Palace jumping is another exploit that bugs me. The best fix to my mind is to eliminate the abilty to abandon your capitol. If I remember correctly you can't choose "abandon city" with your capitol, but you can if you build a settler. Another idea would be to put the capitol in a completely random city or the closest city to the previous. The final idea would be to have some kind of penalty for losing your capitol (going into anarchy would be one idea, though it doesn't make the jump less exploitive - you just do it when you want to change goverments).
------------------------------------------------------------------------------
Using a spy to get someone to declare war on you, if I remember correctly, is still an exploit that should also be not to hard to prevent. One simply counts only the first failure for your rep hit and not later ones in the same turn.
-----------------------------------------------------------------------------
I agree with the realms beyond site on the issue of peace treaties. Hopefully, breaking a peace treaty will mean you can't then extort the AI for more stuff in exchange for peace later on.
----------------------------------------------------------------------------
There are a number of "bug" exploits I would imagine being fixed either in the expansion or the next patch. The mobility bug being one of them. Another would the following that I took from the Realms Beyond site (assuming it still exists - I haven't tested it)
"Free Healing": Normally, if a unit moves, it is not supposed to heal at the end of that turn. However, when the game is saved and loaded, the "has moved" flag is not maintained. It is redetermined from movement remaining. If a damaged unit can make it to a barracks city without using any movement -- via rails and "sailing port to port" -- and you save and then load the game, it will heal at the end of that turn. Saving and reloading to deliberately gain free healing is prohibited.
--------------------------------------------------------------------------------
It would be also nice if the game was a bit more intelligent about deal breaking. I think this one was the hardest on new players, and somewhat discouraging. Too many things can happen to break a deal that are outside your control which means your rep is trashed without you knowing why. It would be really neat if blame could be assigned to whatever action causes the problem.
For example, if the human breaks the road between him and a trading partner, then I think he should take the blame. On the other hand if the trading partner loses his last port city due to a war with a third civ, then the player should not be blamed. If A and C are trading through player B's land, then I think an A vs B war should give A the rep hit and a B vs C war should give C the rep hit.
Even if none of this is doable, it would at least be nice to inform a player whenever he takes a hit to his reputation. I found it VERY difficult to improve my relations with the AIs because I had no clue that my rep was being damaged or why as there is no feedback. I still do not know if the following are bad or not:
a) having an alliance to kill a civ and taking its last city.
b) hitting 0 gold when in anarchy with gpt owed to another civ.
c) having a trade with another civ that is destroyed.
d) having a 20 year peace treaty with another civ that is destroyed.
Actually, I think if a civ is destroyed then all contracts that involve that civ (trades with it, or alliances against it) should have no rep hit. I have heard this is not the case, though I don't know for sure.
---------------------------------------------------------------
INTERFACE EXPLOITS
The following are also from the realms beyond site. They mostly involve doing things at the "wrong time". I kind of like being able to do things like handling disorder (due to lost luxury for example) so it only hits 1 town instead of all of them. I think some of these would be fixable without ruining that functionality. I would even be ok if you couldn't go to other cities at all if these were all fixed as well, though I would prefer a better solution.
"Free Wealth": If you set a city to wealth, you get the benefit of the cash it generates before the game goes through the city production rotation. Thus, if you zoom a city earlier in the queue, and "scroll ahead" to the wealth city, you can get double production on a given turn by assigning a project to the city after the wealth has already been tabulated, as the wealth benefit registers before the city's actual production turn comes up. All use of scroll ahead to get "free wealth" is prohibited.
"Shared Tile": You can use "scroll ahead" after a particular city has registered use of a tile in the production queue between turns, to reach a city that shares that tile but has not taken its turn yet, to "work a tile" on the same turn by more than one city. Any deliberate use of scroll ahead to share tiles is prohibited.
"The Big Picture": On any turn after discovering new tech, you can click "What's the Big Picture", and from there get to any of the operational menus with the Function keys. This opens all manner of between-turn access, even before you get into the production queue, with a number of loopholes and exploits opened up. ALL use of clicking "The Big Picture" between turns to gain deliberate access to menus is now prohibited. Any use of scroll ahead for legitimate purposes must come through zooming during the production cycle.
Grand Theft": If you steal tech and you grab the one you are currently researching, you get the select next tech popup. If you then click "The Big Picture", you don't go to the tech screen but to the steal tech screen, where you can steal another. This is strictly prohibited.
==============================================
Then there are some exploits that would be harder to fix:
ROP rape is an interesting one. It is already somewhat balanced by the reputation hit you take when you do it. Still it is a bit silly you can move huge numbers of troops next to someones capitol even with an ROP. The only way to "solve" it that I can think of would be to teleport all units out when war is declared. It would make for a very different game. I am not sure which I feel is better.
-----------------------------------------------------
I am not sure the best way to fix the "pop rushing" exploit that has been disallowed by most sites This involves adding workers to a town and rushing units. The lone towns rep becomes terrible, but since this only affects that town, it doesn't really matter. Perhaps one way to handle this is that if a city reaches a certain amount of unhappiness due to rushing, it starts getting applied to the whole civ. Conversely, maybe when the sum of unhappiness in all cities combined reaches a certain amount, then it acts on every city.
------------------------------------------------------------------------------
The "scout resource" denial is probably fixable. Probably, the best way would be for the AI to evaluate how important the resource is for it and factor that into the equation on whether war is worth it. Or perhaps just make it better at kicking out the scout and grabbing the resource itself.
The following description was grabbed from the GOTM NO-NO page.
Place a scout on a square where there is a resource and in (future) enemy territory. As long as you have peace with them you can leave him there undisturbed and the other will never be able to build a road to it.
============================================
There are others listed on the Realms beyond and GOTM page. Thing like "Infinite City Sprawl" didn't really seem to be a bug to me but more a strategy and others such as pulling the AI this way and that are ones that require different game dynamics or generic improvements to the AI that are outside the scope of the "easier" fixes. Like others, I hope the AI is improved, but I feel the AI right now if pretty strong for a game of this complexity (PTW 1.21). As someone who almost went for a PhD in AI until I realized just how difficult even the seemingly simplest things can be. Better troop coordination would probably be the single biggest improvement one could make to the AI, but also extremely non-trivial. So I will oppose a number of previous posts and state that I think they have done well already. The fact that they can keep such highly skilled players amused is a a real accomplishment. Hopefully, there will be time to make things even better than they are now.
I am getting off topic. Perhaps I should end this post.
PS: I read a random comment that the A* algorithm is used for computing shortest path. If I remember correctly, A* breaks down if you have 0 length paths as the standard heuristic (distance) is no longer an accurate measure. I think this is why the shortest path is not alway found when rails become available. A breadth first search would work, but that is a lot more computationally expensive. Now add in the fact that the tracks can be blocked by other units. So the seemingly simple job of getting from one point of the map to the another becomes a really difficult problem. In other words, telling how far apart two things are is a major computationally expensive and/or inaccurate calculation. This is the fundamental problem of AI. Even things that seem trivial often are very complex. Now I am really off topic. I will end now.
Links that I used as reference in an attempt to be more accurate:
Realms beyond exploit list
gotm exploit list
==================================
One thing I am hoping the civ team can do for the expansion i
s fixing of some of the exploits that occur in the game. Some are hard to fix and some it is really hard to determine if they are strategy or exploit.
On the other hand, I have real hopes for some of the seemingly easier ones:
One exploit that I find that bothers me is the fact that the AI will make a deal where it takes gold/turn when it plans to attack on the very next turn. I am not sure why it bugs me - perhaps its simply because it is hard not to take advantage of it, but when I do it feels dirty. I would imagine this one is not too hard to fix. Easiest would be to not make per turn deals with someone the AI has recently decided it is going to declare war on. More complex would be to base the decision to attack on your outstanding deals.
--------------------------------------------------------------------------
Palace jumping is another exploit that bugs me. The best fix to my mind is to eliminate the abilty to abandon your capitol. If I remember correctly you can't choose "abandon city" with your capitol, but you can if you build a settler. Another idea would be to put the capitol in a completely random city or the closest city to the previous. The final idea would be to have some kind of penalty for losing your capitol (going into anarchy would be one idea, though it doesn't make the jump less exploitive - you just do it when you want to change goverments).
------------------------------------------------------------------------------
Using a spy to get someone to declare war on you, if I remember correctly, is still an exploit that should also be not to hard to prevent. One simply counts only the first failure for your rep hit and not later ones in the same turn.
-----------------------------------------------------------------------------
I agree with the realms beyond site on the issue of peace treaties. Hopefully, breaking a peace treaty will mean you can't then extort the AI for more stuff in exchange for peace later on.
----------------------------------------------------------------------------
There are a number of "bug" exploits I would imagine being fixed either in the expansion or the next patch. The mobility bug being one of them. Another would the following that I took from the Realms Beyond site (assuming it still exists - I haven't tested it)
"Free Healing": Normally, if a unit moves, it is not supposed to heal at the end of that turn. However, when the game is saved and loaded, the "has moved" flag is not maintained. It is redetermined from movement remaining. If a damaged unit can make it to a barracks city without using any movement -- via rails and "sailing port to port" -- and you save and then load the game, it will heal at the end of that turn. Saving and reloading to deliberately gain free healing is prohibited.
--------------------------------------------------------------------------------
It would be also nice if the game was a bit more intelligent about deal breaking. I think this one was the hardest on new players, and somewhat discouraging. Too many things can happen to break a deal that are outside your control which means your rep is trashed without you knowing why. It would be really neat if blame could be assigned to whatever action causes the problem.
For example, if the human breaks the road between him and a trading partner, then I think he should take the blame. On the other hand if the trading partner loses his last port city due to a war with a third civ, then the player should not be blamed. If A and C are trading through player B's land, then I think an A vs B war should give A the rep hit and a B vs C war should give C the rep hit.
Even if none of this is doable, it would at least be nice to inform a player whenever he takes a hit to his reputation. I found it VERY difficult to improve my relations with the AIs because I had no clue that my rep was being damaged or why as there is no feedback. I still do not know if the following are bad or not:
a) having an alliance to kill a civ and taking its last city.
b) hitting 0 gold when in anarchy with gpt owed to another civ.
c) having a trade with another civ that is destroyed.
d) having a 20 year peace treaty with another civ that is destroyed.
Actually, I think if a civ is destroyed then all contracts that involve that civ (trades with it, or alliances against it) should have no rep hit. I have heard this is not the case, though I don't know for sure.
---------------------------------------------------------------
INTERFACE EXPLOITS
The following are also from the realms beyond site. They mostly involve doing things at the "wrong time". I kind of like being able to do things like handling disorder (due to lost luxury for example) so it only hits 1 town instead of all of them. I think some of these would be fixable without ruining that functionality. I would even be ok if you couldn't go to other cities at all if these were all fixed as well, though I would prefer a better solution.
"Free Wealth": If you set a city to wealth, you get the benefit of the cash it generates before the game goes through the city production rotation. Thus, if you zoom a city earlier in the queue, and "scroll ahead" to the wealth city, you can get double production on a given turn by assigning a project to the city after the wealth has already been tabulated, as the wealth benefit registers before the city's actual production turn comes up. All use of scroll ahead to get "free wealth" is prohibited.
"Shared Tile": You can use "scroll ahead" after a particular city has registered use of a tile in the production queue between turns, to reach a city that shares that tile but has not taken its turn yet, to "work a tile" on the same turn by more than one city. Any deliberate use of scroll ahead to share tiles is prohibited.
"The Big Picture": On any turn after discovering new tech, you can click "What's the Big Picture", and from there get to any of the operational menus with the Function keys. This opens all manner of between-turn access, even before you get into the production queue, with a number of loopholes and exploits opened up. ALL use of clicking "The Big Picture" between turns to gain deliberate access to menus is now prohibited. Any use of scroll ahead for legitimate purposes must come through zooming during the production cycle.
Grand Theft": If you steal tech and you grab the one you are currently researching, you get the select next tech popup. If you then click "The Big Picture", you don't go to the tech screen but to the steal tech screen, where you can steal another. This is strictly prohibited.
==============================================
Then there are some exploits that would be harder to fix:
ROP rape is an interesting one. It is already somewhat balanced by the reputation hit you take when you do it. Still it is a bit silly you can move huge numbers of troops next to someones capitol even with an ROP. The only way to "solve" it that I can think of would be to teleport all units out when war is declared. It would make for a very different game. I am not sure which I feel is better.
-----------------------------------------------------
I am not sure the best way to fix the "pop rushing" exploit that has been disallowed by most sites This involves adding workers to a town and rushing units. The lone towns rep becomes terrible, but since this only affects that town, it doesn't really matter. Perhaps one way to handle this is that if a city reaches a certain amount of unhappiness due to rushing, it starts getting applied to the whole civ. Conversely, maybe when the sum of unhappiness in all cities combined reaches a certain amount, then it acts on every city.
------------------------------------------------------------------------------
The "scout resource" denial is probably fixable. Probably, the best way would be for the AI to evaluate how important the resource is for it and factor that into the equation on whether war is worth it. Or perhaps just make it better at kicking out the scout and grabbing the resource itself.
The following description was grabbed from the GOTM NO-NO page.
Place a scout on a square where there is a resource and in (future) enemy territory. As long as you have peace with them you can leave him there undisturbed and the other will never be able to build a road to it.
============================================
There are others listed on the Realms beyond and GOTM page. Thing like "Infinite City Sprawl" didn't really seem to be a bug to me but more a strategy and others such as pulling the AI this way and that are ones that require different game dynamics or generic improvements to the AI that are outside the scope of the "easier" fixes. Like others, I hope the AI is improved, but I feel the AI right now if pretty strong for a game of this complexity (PTW 1.21). As someone who almost went for a PhD in AI until I realized just how difficult even the seemingly simplest things can be. Better troop coordination would probably be the single biggest improvement one could make to the AI, but also extremely non-trivial. So I will oppose a number of previous posts and state that I think they have done well already. The fact that they can keep such highly skilled players amused is a a real accomplishment. Hopefully, there will be time to make things even better than they are now.
I am getting off topic. Perhaps I should end this post.PS: I read a random comment that the A* algorithm is used for computing shortest path. If I remember correctly, A* breaks down if you have 0 length paths as the standard heuristic (distance) is no longer an accurate measure. I think this is why the shortest path is not alway found when rails become available. A breadth first search would work, but that is a lot more computationally expensive. Now add in the fact that the tracks can be blocked by other units. So the seemingly simple job of getting from one point of the map to the another becomes a really difficult problem. In other words, telling how far apart two things are is a major computationally expensive and/or inaccurate calculation. This is the fundamental problem of AI. Even things that seem trivial often are very complex. Now I am really off topic. I will end now.
