OK, posted a screenshot to the bugs thread.
C2C: Promotions
- Thread starter ls612
- Start date
OK, posted a screenshot to the bugs thread.
Hydromancerx
C2C Modder
He is right all the Hunter promotion if you look in the pedia are named Hunter I (+1 First Strike) or Hunter II (+1 First Strike). Which looks weird.
Did anyone else notice that Battering Rams are able to be built at The Wheel, but aren't eligible for any promotions until Military Training? I have an Ancient-era game going where I am building Battering Rams with 2 XP (from Caste civic), but they have absolutely nothing to spend them on. This seems very strange to me.
Dancing Hoskuld
Deity
Clockwork Golems have th same problem. They can be built 2-3 eras before their promotions become available.
Well, there isn't really a nice solution to that, as setting them to UNITCOMBAT_MELEE causes all sorts of wierdness, such as them being able to take Martial Arts promotions, thus leading to "Kung-Fu" golems
.
Dancing Hoskuld
Deity
Well, there isn't really a nice solution to that, as setting them to UNITCOMBAT_MELEE causes all sorts of wierdness, such as them being able to take Martial Arts promotions, thus leading to "Kung-Fu" golems
Then they should be removed from the game.
Thunderbrd
C2C War Dog
Don't remove them. A solution is pending. As a side note, it wouldn't be difficult, even without any solutions I'm about to bring to the table here, to create a new combat class for the golem and perhaps others like it to come, that allows you to craft the way you want its promo access to go. I think its more than justified that it have its own anyhow.
strategyonly
C2C Supreme Commander
EDIT:
unit naming in the Config file needs attention also, the para that was here before was wrong, "my bad."
Thunderbrd
C2C War Dog
SO: PLEASE pm me fully detailed information on all python interactions with combat classes! I'm definitely going to need to understand how python plays into that!
Do you have the outcomes for this in the CIV4OutcomeInfos.xml yet? If you do I can make the promotions, if I can find the post saying what you wanted them to be.
Dancing Hoskuld
Deity
Do you have the outcomes for this in the CIV4OutcomeInfos.xml yet? If you do I can make the promotions, if I can find the post saying what you wanted them to be.
It is a bit of a "chicken and egg" problem at the moment. Each time I try and do stuff I find there is something missing elsewhere. I need three promotions all only for UNITCOMBAT_DIPLOMAT.
- PROMOTION_SILVERTONGUE_I Silver Tongue I available with oral tradition
- PROMOTION_SILVERTONGUE_II Silver Tongue II available with oratory
- PROMOTION_SILVERTONGUE_III Silver Tongue III available with ????
IronClaymore
Warlord
Probably part of my own idiosyncrasy, but I usually alter the XML of my own personal download to allow all units access to basic combatty promotions. Mainly drill and combat, but in AND/C2C it includes sensors, land warrior and stealth. Late game, I end up with units with over 100 experience, and I like to be able to look forward to more levels and more power. So, perhaps give robot units access to the basics. Late-game, the player and AI will certainly choose combatAI or sensors over these when the techs are acquired, but it'll allow the golems something in early renaissance.
My annoyance is that clockwork golems don't upgrade to anything. Maybe steam golems with steampunk? And further on into modern and transhuman constructs. I want the ancient senile AIs from my veteran iCombat 14 golems eventually transferred to those awesome 290 hi-tech robots, and embarrassing their brethren by using ye olde english and refusing to learn what diodes are
.
Thunderbrd
C2C War Dog
I'm thinking they should be renamed from the fantasy term 'golem' to the more sci-fi term 'automaton' and yes, should be a chain that starts and clockwork and gets a step in every punk tech.
I'd also like to see the punk tech cultures eliminated and just use the technology as the prereq BUT make the techs ONLY available with the use of a Great Scientist (or perhaps the Great Engineer could be used for this as well...) and their normal tech prereqs. Is this possible without any new tags? I suppose it could take a touch of AI work on the Scientist to enable this properly for the AI. This would allow more than one civ to reach them but would certainly take an exotic cost to achieve and would give another great person a bigger role in delivering an edge to those that birth them.
I'd be interested to see some suggestions for a unique promo line or two for just automaton units... I may have some ideas for this to come as well.
I'd also like to see the punk tech cultures eliminated and just use the technology as the prereq BUT make the techs ONLY available with the use of a Great Scientist (or perhaps the Great Engineer could be used for this as well...) and their normal tech prereqs. Is this possible without any new tags? I suppose it could take a touch of AI work on the Scientist to enable this properly for the AI. This would allow more than one civ to reach them but would certainly take an exotic cost to achieve and would give another great person a bigger role in delivering an edge to those that birth them.
I'd be interested to see some suggestions for a unique promo line or two for just automaton units... I may have some ideas for this to come as well.
I'll put this on my todo list, but that is quite long, so these may not be made for a week or so.
Nakkimunakas
Chieftain
AI promotion choises have become very poor with the addition of new promotions. I think there should be many times less of the promotions like barbar hunter and terrain based promotion like arctic combat (or not at all untill the AI can use them properly) since the AI doesn't seem know how to use them properly at the moment.
I don't have anything against those promotions but they seem to be a must have to almost every AI unit and as a side effect they take away more usefull promotion combinations. For example I rarely see any combat I + shock I and II combinations anymore. I remember those posed some real problems formed as big stacks back when I played RAND or BTS. Now it's just too easy to slap around AI stacks since so many of their promotions are useless in combat. Terrain based promotions might be usefull but it's very rare that they are actually on right kind of terrain.
I don't have anything against those promotions but they seem to be a must have to almost every AI unit and as a side effect they take away more usefull promotion combinations. For example I rarely see any combat I + shock I and II combinations anymore. I remember those posed some real problems formed as big stacks back when I played RAND or BTS. Now it's just too easy to slap around AI stacks since so many of their promotions are useless in combat. Terrain based promotions might be usefull but it's very rare that they are actually on right kind of terrain.
Thunderbrd
C2C War Dog
I suppose part of the issue there is that the terrain based promos are giving so many bonuses that are adding to the evaluation value for the ai. I'll see if I can tweak terrain benefit values to compensate a bit.
I have a bunch of changes planned for AI promotion evaluation.
On the terrain ones specifically, it needs two things:
1) To have some 'inertia' built in, so that a unit promoted along one terrain line prefers furtehr promotions along the same line to other terrain promotions (that way stacks will tend to get different terrain 'specialists' and have good defenders in more terrains, rather than several 'ok' ones)
2) Stacks need to evaluate the revealed terrain in the local (but quite wide interpretation of local) area, and bias their choices towards terrains that occur there
Both of these are pretty easy to do (if you want to do it Thunderbrd, please go ahead - I doubt it'll hit the top of my list for a while)
Thunderbrd
C2C War Dog
In some areas the first point you made there does appear to exist, places like city attack, city defense, withdrawal etc... To add some of this thinking into the terrain promotions is a very good idea.
The second point you made there is SOMEWHAT in play already if we simply reduce the current terrain eval values to allow for a terrain promo to balance out more with other options. The way this is currently in play is via combat promotions. The more the stack attacks and defends in a particular region, the more likely its going to end up with the first promo for that sort of terrain. If the above noted issue is implemented further, we'd see those units that inadvertently become terrain defenders begin to start focusing more down that line.
However, one thing I think could be done to greatly improve that structure all around is to add some further unit AI options entirely. I no longer think 'Counter' is enough to handle the variety of combat roles where specialized troops to counteract given threats is considered. Thus, there could/should be a fairly significant overhaul of the whole unit ai imho. This would probably be a good thing for us all to start discussing I think.
But for NOW, I'll just address a reduction in the values of those promo elements (the main thing to consider there is that with vanilla, each terrain promo really only applied to one terrain type or feature and now they are considering much more where terrains are similar, such as the bonus to forest & jungle defense AND the bonus to bamboo defense in the Forestry line. The effect is added to the unit to balance out the fact that where many forests were in vanilla maps have now been replaced by bamboo, an equivalent, but the evaluation considers it to be twice as valuable because its two sets of bonuses. So I'll try to figure out some kind of means to get that to self-balance. Additionally, the Inertia element would be really worthwhile to implement as well so I'll see what I can do there.
But I've really been thinking that the underlying structure is just too crude to really represent a means to GOOD ai strategies. I suspect you'd agree.
I've been having to interact with the promo evaluation section extensively for my current projects. I won't pretend to know for sure if my math is at all effective until some extended playtesting to come. In other words, what I've currently done has been a bit of throwing darts at a target based on other examples.
In the process, however, I've also found some really bad evaluatory values in other areas. For example, city attack units are FAR more likely to want to take withdraw promos rather than actual city raider promos... Thus one of the big reasons I feel the ai wastes a LOT of time on cheap withdrawal units when they could have had REAL power slamming into my cities. Not to underestimate the value of withdrawal, but I've always found it over-preferred by the ai and I've already tweaked it a bit to balance that evaluation a bit better.
So I guess what I'm trying to say is that its a region that could use some extensive editing and full re-evaluation and anything I do for it at the moment would be merely a to-be-reworked patch for now anyhow.
Now, I'm not sure what ALL of your plans are and I hope that I don't end up conflicting with them at all when I go to merge this fairly large project I've been working on here. Its hard to say, even if its not likely to hit the top of your list for a while, whether I'll be updating the dll before you get to that section of your plans or not. I'm still a bit out from that considering that I keep realizing I need to do just a bit more so as to introduce the modification's full intention without having to go back over things and re-edit them just to update to the next step in the 'plan'. (if that makes any sense) So nothing I'm going to do will see application in the svn anytime soon really. But I am making a lot of progress so far
.
If it wouldn't be too big a pain, I'd like you to send me a message or something to kind of give me an overview on those planned changes for AI promo evaluation and other unit AI issues just so I can get a feel for what you have in mind there... I may also have some suggestions to offer in that dept now that I've been taking a deeper look at that region of the code, but more importantly, it would be good for me to make sure I keep them in mind as I go here.
The second point you made there is SOMEWHAT in play already if we simply reduce the current terrain eval values to allow for a terrain promo to balance out more with other options. The way this is currently in play is via combat promotions. The more the stack attacks and defends in a particular region, the more likely its going to end up with the first promo for that sort of terrain. If the above noted issue is implemented further, we'd see those units that inadvertently become terrain defenders begin to start focusing more down that line.
However, one thing I think could be done to greatly improve that structure all around is to add some further unit AI options entirely. I no longer think 'Counter' is enough to handle the variety of combat roles where specialized troops to counteract given threats is considered. Thus, there could/should be a fairly significant overhaul of the whole unit ai imho. This would probably be a good thing for us all to start discussing I think.
But for NOW, I'll just address a reduction in the values of those promo elements (the main thing to consider there is that with vanilla, each terrain promo really only applied to one terrain type or feature and now they are considering much more where terrains are similar, such as the bonus to forest & jungle defense AND the bonus to bamboo defense in the Forestry line. The effect is added to the unit to balance out the fact that where many forests were in vanilla maps have now been replaced by bamboo, an equivalent, but the evaluation considers it to be twice as valuable because its two sets of bonuses. So I'll try to figure out some kind of means to get that to self-balance. Additionally, the Inertia element would be really worthwhile to implement as well so I'll see what I can do there.
But I've really been thinking that the underlying structure is just too crude to really represent a means to GOOD ai strategies. I suspect you'd agree.
I've been having to interact with the promo evaluation section extensively for my current projects. I won't pretend to know for sure if my math is at all effective until some extended playtesting to come. In other words, what I've currently done has been a bit of throwing darts at a target based on other examples.
In the process, however, I've also found some really bad evaluatory values in other areas. For example, city attack units are FAR more likely to want to take withdraw promos rather than actual city raider promos... Thus one of the big reasons I feel the ai wastes a LOT of time on cheap withdrawal units when they could have had REAL power slamming into my cities. Not to underestimate the value of withdrawal, but I've always found it over-preferred by the ai and I've already tweaked it a bit to balance that evaluation a bit better.
So I guess what I'm trying to say is that its a region that could use some extensive editing and full re-evaluation and anything I do for it at the moment would be merely a to-be-reworked patch for now anyhow.
Now, I'm not sure what ALL of your plans are and I hope that I don't end up conflicting with them at all when I go to merge this fairly large project I've been working on here. Its hard to say, even if its not likely to hit the top of your list for a while, whether I'll be updating the dll before you get to that section of your plans or not. I'm still a bit out from that considering that I keep realizing I need to do just a bit more so as to introduce the modification's full intention without having to go back over things and re-edit them just to update to the next step in the 'plan'. (if that makes any sense) So nothing I'm going to do will see application in the svn anytime soon really. But I am making a lot of progress so far
.If it wouldn't be too big a pain, I'd like you to send me a message or something to kind of give me an overview on those planned changes for AI promo evaluation and other unit AI issues just so I can get a feel for what you have in mind there... I may also have some suggestions to offer in that dept now that I've been taking a deeper look at that region of the code, but more importantly, it would be good for me to make sure I keep them in mind as I go here.
If possible can whomever is in charge of what promotions are given as free promotions via battle please remove the Fieldsman promotions from battle promotions?
It's a pain to get it every time and see a specialized Hills unit suddenly lose some of it's hill strength.
Cheers
It's a pain to get it every time and see a specialized Hills unit suddenly lose some of it's hill strength.
Cheers
massive undertaking, not to be entered into lightly. Especially if you want to retain save compatibility.
IMO just reducing them is the wrong approach. AI should evaluate the proportion of its revealed (localish) terrain that benefits from the promotion and weight it according to that proportion. Woodsman reall IS excellent if you have a lot of forest - if you just reduce its value the AI wil take a 10% general combat bonus over woodsman in a highly forested area, which would be a bad choice. If you want to go for a simplistic approach just maintain counts of revelaed terrain types and adjust them (globally) each time a new tile is revealed (revealed value goes from false to true, and evaluate starting state on load in the map load routine). That has tiny runtime overheads and although it loses the 'local' element it would still be way better than what we have now.
But I've really been thinking that the underlying structure is just too crude to really represent a means to GOOD ai strategies. I suspect you'd agree.
I've been having to interact with the promo evaluation section extensively for my current projects. I won't pretend to know for sure if my math is at all effective until some extended playtesting to come. In other words, what I've currently done has been a bit of throwing darts at a target based on other examples.
In the process, however, I've also found some really bad evaluatory values in other areas. For example, city attack units are FAR more likely to want to take withdraw promos rather than actual city raider promos... Thus one of the big reasons I feel the ai wastes a LOT of time on cheap withdrawal units when they could have had REAL power slamming into my cities. Not to underestimate the value of withdrawal, but I've always found it over-preferred by the ai and I've already tweaked it a bit to balance that evaluation a bit better.
So I guess what I'm trying to say is that its a region that could use some extensive editing and full re-evaluation and anything I do for it at the moment would be merely a to-be-reworked patch for now anyhow.
Now, I'm not sure what ALL of your plans are and I hope that I don't end up conflicting with them at all when I go to merge this fairly large project I've been working on here. Its hard to say, even if its not likely to hit the top of your list for a while, whether I'll be updating the dll before you get to that section of your plans or not. I'm still a bit out from that considering that I keep realizing I need to do just a bit more so as to introduce the modification's full intention without having to go back over things and re-edit them just to update to the next step in the 'plan'. (if that makes any sense) So nothing I'm going to do will see application in the svn anytime soon really. But I am making a lot of progress so far
If it wouldn't be too big a pain, I'd like you to send me a message or something to kind of give me an overview on those planned changes for AI promo evaluation and other unit AI issues just so I can get a feel for what you have in mind there... I may also have some suggestions to offer in that dept now that I've been taking a deeper look at that region of the code, but more importantly, it would be good for me to make sure I keep them in mind as I go here.
Similar threads
