timerover51
Deity
Last edited:
Resource icon
- Thread starter Flintlock
- Start date
Flintlock
Emperor
- Joined
- Sep 25, 2004
- Messages
- 1,249
Looking at the file sizes, those EXEs are based on the PCGames.de EXE, which is compatible with C3X. I added support for that EXE back in mod version 12. I don't know if the edits for no unit limit, etc. would conflict with the edits done by C3X. Probably not, but I can imagine they might depending on what specifically they change.
Oni Ryuu
King
Hi, Tsubasanut
I hoped it was going to be that simple but my other PC seems to have died in storage. I might pick up a charity shop computer to see it will work before I get another new PC.
Thank you guys.
I hoped it was going to be that simple but my other PC seems to have died in storage. I might pick up a charity shop computer to see it will work before I get another new PC.
Thank you guys.
Would it be possible to limit the number of Units in a Tile? I'd love to try playing the game without Stacks Of Doom.
Predator145
King
- Joined
- May 22, 2020
- Messages
- 710
A feature of disabling the construction of a city improvement if another improvement is present would be very nice to boost the AI in tandem with the perfume. The presence of a palace should disable courthouses and police stations. Shakespears should disable hospitals and aqueducts. Right now the AI still builds perfumed courthouses in their capital.
Sutsuj
Chieftain
- Joined
- Apr 30, 2020
- Messages
- 43
I would love this feature too!
I see two possible issues and maybe some fixes:
1. Would this limit apply to cities and if so, does that stop unit production or auto-production?
2. The trickiest part might be AI path-planning. Can they plan for the fact that something *will* be open even though their own units are currently blocking their way? One simple way around this issue might be to make the "is-spot-available" function always say yes even if it's full of your units *unless* you're right next to the tile. That way you can't actually break the rule, but you can head in the right direction, and as long as your other units move in the meantime, there won't be many issues in practice. Regardless, for multiplayer purposes this limit would be really cool even if the AI is somewhat weakened.
Brenshar
Chieftain
I always had this idea in my head that the civ3 army function could in someway be used to limit units per tile. Of course such a solution runs into a lot of problems, but like if a produced unit could automatically be created into the nearest army, and if units could also be killed out of an army.
Good point: it almost certainly should not apply in cities.
I'd love first just to see what the AI does first - Emergent behavior? (
)I'm not certain that it would weaken the AI. If Flintlock chooses to, and can, the next step (
) might be to force the AI to balance each stack into some (on land) ratio of A
:Artillery.
Brenshar
Chieftain
I'd love first just to see what the AI does first - Emergent behavior? (
I imagine it would behave as the ai currently does when its path is blocked by other players - it would start a new pathing movement around the blocked tiles
My point precisely: the game cannot violate its own algorithms, so we'd hypothetically have, e.g., not one stack of 20 Units advancing, but two stacks of 10. I've not had the time to think this through all the way yet
help: @Civinator? @Quintillus? @WildWeazel?) but I suspect that it could be a genuine game changer.
- Joined
- Mar 17, 2007
- Messages
- 9,326
Generally speaking (I am not a pathing algorithm expert yet, maybe an enthusiast?), if a path isn't available due to enemy unit/enemy city/impassible terrain, its distance is going to be counted as infinity. In a straight-up A* or Dijkstra approach, when trying to find the shortest path, you're just looking at available routes. So right now, if you wanted to travel from Moscow to Rostov, Google Maps wouldn't route you on the M4 highway by Voronezh because it's closed.
Sutsuj is hinting at a couple possible ways around that. You could say, "but I want to leave on Sunday, the military convoy by Voronezh should be gone by then." Does Google Maps take into account projected road re-openings? I don't know. But one can imagine a timekeeping layer where you can say, "considering all current orders, once the unit gets here, the other unit will already be gone so the path will be open." As humans, we kind of consider that when moving units, especially in Civ V with its 1 UPT - sure there's a unit at the chokepoint there, but I'm not going to leave it there.
Or, we can say, "Okay, there's traffic at Voronezh, let's just route two-thirds of the way there are re-evaluate later." This is certainly a possibility, but you run the risk of longer-term obstacles, optimizing for what is actually a local (rather than overall) minimum distance. What if part of the Pacific Coast Highway slid into the ocean? You'd have been better off going around in the first place. I'm not sure what exactly an equivalent to this is in Civ terms; with 1 UPT, you run into the problem of maybe the units that are blocking aren't really going to move anywhere very fast, but that's less of a risk with a 10-unit cap. Assuming that the algorithm can distinguish that the obstacle is your mobile units, rather than enemy units, or something else blocking the way.
One of the key implications, though, is that while you could still use standard algorithms as a structure, there are enough rules introduced that you're going to have to be making your own routing rules as well, which means the problem is not easy. That, "how do we implement routing rules" is, IMO, one of the reasons Civ V/VI AI combat is so poor - especially with 1 UPT as the cap, the AI just didn't know how to handle traffic jams very well. Old World does a much better job, IMO both due to its less-crowded maps, and due to the skills of its developers (Soren Johnson in particular but I'm sure he's not the only one).
It certainly does fall into what I consider the "interesting problem" space. And I would make no guarantees that what I'd first implement would in fact work well, even if it sounded like it might before it was put into practice.
Sutsuj is hinting at a couple possible ways around that. You could say, "but I want to leave on Sunday, the military convoy by Voronezh should be gone by then." Does Google Maps take into account projected road re-openings? I don't know. But one can imagine a timekeeping layer where you can say, "considering all current orders, once the unit gets here, the other unit will already be gone so the path will be open." As humans, we kind of consider that when moving units, especially in Civ V with its 1 UPT - sure there's a unit at the chokepoint there, but I'm not going to leave it there.
Or, we can say, "Okay, there's traffic at Voronezh, let's just route two-thirds of the way there are re-evaluate later." This is certainly a possibility, but you run the risk of longer-term obstacles, optimizing for what is actually a local (rather than overall) minimum distance. What if part of the Pacific Coast Highway slid into the ocean? You'd have been better off going around in the first place. I'm not sure what exactly an equivalent to this is in Civ terms; with 1 UPT, you run into the problem of maybe the units that are blocking aren't really going to move anywhere very fast, but that's less of a risk with a 10-unit cap. Assuming that the algorithm can distinguish that the obstacle is your mobile units, rather than enemy units, or something else blocking the way.
One of the key implications, though, is that while you could still use standard algorithms as a structure, there are enough rules introduced that you're going to have to be making your own routing rules as well, which means the problem is not easy. That, "how do we implement routing rules" is, IMO, one of the reasons Civ V/VI AI combat is so poor - especially with 1 UPT as the cap, the AI just didn't know how to handle traffic jams very well. Old World does a much better job, IMO both due to its less-crowded maps, and due to the skills of its developers (Soren Johnson in particular but I'm sure he's not the only one).
It certainly does fall into what I consider the "interesting problem" space. And I would make no guarantees that what I'd first implement would in fact work well, even if it sounded like it might before it was put into practice.
- Joined
- May 5, 2005
- Messages
- 10,300
Quintillus explained it much better than I could do it.
But I can confirm it with some observations I especially had with the SOE WW 2 scenario. Here the Hungarian, Romanian, Bulgarian and Italian forces were mostly traffic obstacles for the more powerful German forces. They blocked the path or at least slowed down the momentum for many of the important operations and could not be attacked, as they are in an alliance with the more important forces. So one of the most urgent tasks in the settings of that scenario was to set traffic obstacles against these traffic obstacles, when those allied civs should stay as independent civs in that scenario (what in SOE should be the case due to the great work jorde has done by creating some special leaderheads for those civs). The same happened with the US and UK forces (here I decided to melt them together as it was not possible to set working traffic obstacles against those traffic obstacles).
All in all in my eyes these splittings should be avoided if there are no more relevant reasons for creating them. C3C in combination with the Flintlock mod offers some very appropriate different methods to deal effectively with the so-called "stacks of doom":
The stealth attack settings allow to have a "first strike" against any defending unit in a stack - and this eliminates the main argument in using stacks of doom (to have always the best unit available for defending against attacks to that stack). Of course this means a lot of work for the modder to create those stealth attack target settings for may be thousands of units in a mod or scenario. In SOE and CCM this work was done. The Flintlock mod greatly allows to set the stealth attack option even if there is only one target unit in the stack.
Additionally the charm attack settings allow to dim down the defense quality of complete stacks of land units and there is also the ultimate option of nuclear attacks.
So in my eyes there is no need for limiting the number of units in a tile in C3C (even without taking into account the consequences of pathfinding and production for the AI). It should not be forgotten, that the feeling to have to deal with too many units in a battle mostly is a question of production costs and in scenarios by preset units.
But I can confirm it with some observations I especially had with the SOE WW 2 scenario. Here the Hungarian, Romanian, Bulgarian and Italian forces were mostly traffic obstacles for the more powerful German forces. They blocked the path or at least slowed down the momentum for many of the important operations and could not be attacked, as they are in an alliance with the more important forces. So one of the most urgent tasks in the settings of that scenario was to set traffic obstacles against these traffic obstacles, when those allied civs should stay as independent civs in that scenario (what in SOE should be the case due to the great work jorde has done by creating some special leaderheads for those civs). The same happened with the US and UK forces (here I decided to melt them together as it was not possible to set working traffic obstacles against those traffic obstacles).
All in all in my eyes these splittings should be avoided if there are no more relevant reasons for creating them. C3C in combination with the Flintlock mod offers some very appropriate different methods to deal effectively with the so-called "stacks of doom":
The stealth attack settings allow to have a "first strike" against any defending unit in a stack - and this eliminates the main argument in using stacks of doom (to have always the best unit available for defending against attacks to that stack). Of course this means a lot of work for the modder to create those stealth attack target settings for may be thousands of units in a mod or scenario. In SOE and CCM this work was done. The Flintlock mod greatly allows to set the stealth attack option even if there is only one target unit in the stack.
Additionally the charm attack settings allow to dim down the defense quality of complete stacks of land units and there is also the ultimate option of nuclear attacks.
So in my eyes there is no need for limiting the number of units in a tile in C3C (even without taking into account the consequences of pathfinding and production for the AI). It should not be forgotten, that the feeling to have to deal with too many units in a battle mostly is a question of production costs and in scenarios by preset units.
I am intrigued by Charm. Although I have tom2050's study on it bookmarked, I have never used it, although I believe I recall encountering it in your creations & @AnthonyBoscia's (R/L far too often gets in the way of my "One More Turn" time) - Rather than continuing to Thread Jack, do you gents think it's worth another thread?
Moff Jerjerrod
Deity
I like 'charm' but the only thing is you can't use it against units that are in a city. Also, artillery units don't seem to damage hit points. They just reduce the defense of every unit in a stack that is out in the field by half.
Chalito
Prince
- Joined
- Mar 26, 2022
- Messages
- 319
New requests for Mr Flintlock:
Is there a way to make the Landmark Terrains be able to be "impassable to all units", so Mountains could be passable, but Landmark Terrain Mountains couldn't?
Is there a way to make civs be at "Always Peace", the opposite of "Always War" that is currently in the scenario properties? Also, a way to create more than the 4 Groupings in the Editor?
No changes to the actual Editor, but just something we could modify on the scenario.config text file.
Thank you!
Is there a way to make the Landmark Terrains be able to be "impassable to all units", so Mountains could be passable, but Landmark Terrain Mountains couldn't?
Is there a way to make civs be at "Always Peace", the opposite of "Always War" that is currently in the scenario properties? Also, a way to create more than the 4 Groupings in the Editor?
No changes to the actual Editor, but just something we could modify on the scenario.config text file.
Thank you!
Chalito
Prince
- Joined
- Mar 26, 2022
- Messages
- 319
I've noticed this too, I just stick with the text in the config file. The only issue is there are some scenarios with tons of Arty and Naval units, but you just gotta write them all out.
Chalito
Prince
- Joined
- Mar 26, 2022
- Messages
- 319
The Firaxis boys were probably planning on something unique with this feature, but ran out of time, or abandoned it all together when it failed to materialize. Too bad they didn't have Mr. Flintlock on staff to fix all of their mess!!
Sutsuj
Chieftain
- Joined
- Apr 30, 2020
- Messages
- 43
I largely agree with Quintillus. Like he said, my suggestion was intended as a way to avoid routing issues. My goal was to find a simple change that would require as little coding effort as possible, yet fix most of the potential issues. The entire feature could be put into the "is-tile-open" function.
The biggest problem I can foresee with stack limits that don't also include a change such as the one I suggested would be chokepoints on a map. If the map is broad enough, then it's easy for an AI to circumnavigate itself, but if it's narrow, it will get in its own way and decide there is no reason to even walk toward the chokepoint.
In regard to Civinator's point that stack limits are not needed to deal with stacks of doom, I mostly agree but think that you're missing three reasons that tile limits could still be very valuable:
1. Not all mods have the features you mentioned implemented, and thus it would be nice to get to import stack limits into pre-existing mods.
2. More importantly, it would be fun to have battles spread out over a larger area regardless of whether or not we can deal with doom stacks. It would be fun to be forced to do this as a human.
3. Lastly, I don't know if charm works in live multiplayer, but I know from personal experience that teleport does not. Based on what I've seen, Flintlock's engine changes work in live multiplayer; specifically, I've tested that land-sea intersections work.
Similarly, I agree with Vuldacon that the stacks make thematic sense (though so could stack limits). I think Ozymandias and I are focused on gameplay rather than strict historical plausibility.
The biggest problem I can foresee with stack limits that don't also include a change such as the one I suggested would be chokepoints on a map. If the map is broad enough, then it's easy for an AI to circumnavigate itself, but if it's narrow, it will get in its own way and decide there is no reason to even walk toward the chokepoint.
In regard to Civinator's point that stack limits are not needed to deal with stacks of doom, I mostly agree but think that you're missing three reasons that tile limits could still be very valuable:
1. Not all mods have the features you mentioned implemented, and thus it would be nice to get to import stack limits into pre-existing mods.
2. More importantly, it would be fun to have battles spread out over a larger area regardless of whether or not we can deal with doom stacks. It would be fun to be forced to do this as a human.
3. Lastly, I don't know if charm works in live multiplayer, but I know from personal experience that teleport does not. Based on what I've seen, Flintlock's engine changes work in live multiplayer; specifically, I've tested that land-sea intersections work.
Similarly, I agree with Vuldacon that the stacks make thematic sense (though so could stack limits). I think Ozymandias and I are focused on gameplay rather than strict historical plausibility.
- Joined
- May 5, 2005
- Messages
- 10,300
Sutsuj, I think you see this all from the perspective of a multiplayer game, where many of the problems that strike the AI, can be corrected by the human player, but are staying big problems for the AI civs. I see these problems from the perspective of a single player game, where the problems of the AI must be taken into consideration. So some features of the Flintlock mod may be working in multiplayer games, too, until now the Flintlock mod is a single player mod.
It is not only the traffic problem of the blockade of units. There are also AI pathfinding errors of units appearing, that lead to freezes of the game. The release of the mod CCM had an delay of more than one year by these problems.
What should be the solution for the production of units in a city when the tile of the city reaches the tile limit ? What units should the AI throw out of the city and on what (neighboring) tile ? These decisions can be done by the human player when having a look at the specific situation on the map and a facing opponent, but the AI ??
The battles are not "spreading out" by such a setting, as neighboring tiles are not connected in a C3C battle. Additionally in my eyes it is no fun, to handicap the AI even more with such limitations and even more it is no fun, when your game is closed by a freeze triggered by an AI path finding error.
I cannot say if charm attack works in multiplayer games, but this has nothing to do with tile limitations.
All in all, of course it is the decision of Flintlock what to do here (and this could trigger a lot of additional work), but as the things are standing now, for me the most important would be a button to skip that option.
If such stack limits should be so nice, why are you not using the simplest solution for multiplayer games to set a limit for units in a stack between the human players ? If you play your multiplayer games with a mixture of human and AI players, you should not completely ignore the big problems the AI has with such limitations in C3C. The existing traffic problems for the AI triggered by such tile limitations were already posted above. The argument to make the maps much bigger doesn´t solve this problem, at least if you want still to have an interesting playable map for your game.
It is not only the traffic problem of the blockade of units. There are also AI pathfinding errors of units appearing, that lead to freezes of the game. The release of the mod CCM had an delay of more than one year by these problems.
What should be the solution for the production of units in a city when the tile of the city reaches the tile limit ? What units should the AI throw out of the city and on what (neighboring) tile ? These decisions can be done by the human player when having a look at the specific situation on the map and a facing opponent, but the AI ??
The battles are not "spreading out" by such a setting, as neighboring tiles are not connected in a C3C battle. Additionally in my eyes it is no fun, to handicap the AI even more with such limitations and even more it is no fun, when your game is closed by a freeze triggered by an AI path finding error.
I cannot say if charm attack works in multiplayer games, but this has nothing to do with tile limitations.
All in all, of course it is the decision of Flintlock what to do here (and this could trigger a lot of additional work), but as the things are standing now, for me the most important would be a button to skip that option.
Similar threads
