Game balance and AI

Zaga

Chieftain
Joined
Dec 1, 2011
Messages
8
Since it's a long post: the money issue is just a means to make a foundation and build an argument for the main point in the last half: game balance/AI and design. I tend to write long ones :S

A problem which Zeinul recently addressed: there is too much gold. Currently I'm playing a Deity game, and though I started outteched big time I could just buy my way up the tech tree as soon as I shared a religion with one of the techleaders and did some sucking up. Gold had become a more or less irrelevant factor - what's a mere 2000-3000 gold for a key tech when you have over 30.000 (and more coins keep flowing into your coffers)? And after having bought your way to the top, it's easy to modernize your own civ and become the powerhouse of your age. No fancy leader (played Hittites - I think its leader is expansionist/imperialistic while I'm a builder rather than a warmonger) or a sound plan of action other than building as much stuff in my cities as possible and keeping research at 100%. Which is not much of an impressive or coherent strategy, I guess.

Even though my gameplay sucked, the game nonetheless handed out enough cash to break core game aspects like keeping up with tech. But also aspects like crime (the extra costs for watchmen are irrelevant), many events (the costs to do anything, like rebuilding a burned down forest or building, are peanuts in comparison to income), the choice whether to upgrade obsolete units or not (with lots of cash it's close to a no-brainer, effectively eliminating the choice) and the building of wonders (if others get close to building a wonder I'm working on, I just shrug and finish it by paying) are heavily affected by more or less money. Of course one can solve that money-issue by either reducing the amount of gold handed out or by approaching the problem from the other direction and modify the costs of things - make all costs depend on the amount of in-game money, for example.

But maybe it's useful to step back and look at the bigger designing picture. Its clear that in civ4 every change to the game affects a lot of other aspects. Instead of that cash-issue, one can make a similar list of effects as the one above for mechanics like hunting, rogues or the caravans-which-can-be-converted-to-hammers. When reading the posts of the modding team on the forum, lots of enthousiasm to add more and more stuff is shown, and the following decision of whether to add it or not seems to be purely decided by the mechanical feasibility of implementing the idea and the personal vision of the one programming that idea. The issue of having to keep the game as a total balanced and challenging seems to be becoming an afterthought rather than a central consideration. In one of the "fresh idea"-threads regarding some new mechanic for prehistoric era, I remember Koshling replying to the OP/team member that one cannot make things which a big impact on the game modular, while the following reaction of the OP only considered the mechanical aspect without any reference to modularity in the sense of its effects on gameplay or AI. Whether or not Koshling actually meant that mechanical aspect is not really the point here (though given Koshlings dealings with AI, I personally strongly think he didn't), but the reply suggests something about the way which the OP approaches the adding of a new game-aspect.

I'm not in a position to tell the team how to work on their mod of course, but I can imagine how its current decentralized laissez-faire civic invites both innovation (good) and imbalance+AI weakness (bad). Which indeed has resulted in a current game on my harddisk with awesome mechanics and more novelties added every day, but an AI which gets more and more crippled every day since it can't deal with them properly and balance suffering from 'collateral damage'. If some dork (like me, in this case) playing without much skill or any strategy, without checking his leaders traits until midgame and playing on deity before the middle ages already feels like he's won, it testifies to that assertion of the AI having a hard time to cope with all the changes. And it's not an incidental fluke- at least since october there hasn't been a C2C game I didn't abandon between ancient and middle age-era due to having technically won with no remaining threats, and several other posters including Koshling himself mentioned the same issue of the game being not much of a challenge. Without downplaying Koshlings AI programming efforts by the way. With lots of commercial games programmed by paid, professional and full-time people still suffering from crappy AI, it seems a tad optimistic to keep adding new stuff and trust both balance and how the AI has to deal with it will be sorted out in a satisfying way later when the addition is already a fait accompli. Can the team please, please, pretty please with cherry on top consider redirecting their efforts to making the game deal better with the added stuff, instead of adding even more stuff? The doctrine of "more being better" doesn't automatically seem to apply anymore, and it even starts to be counterproductive. No matter how much neat mechanics are added, if a player gets his victory on a gold platter by just sitting, building and waiting it's not much of a mod which adds fun to the vanilla game. Since this is exactly how C2C now plays: a gold-plated automobile which has built-in whiskeydispensers, tastes like pizza when you lick the dashboard and has a set of great tits to rest your head against - a man's wet dream, if only the designer had thought of adding a proper engine as well.
 
by the way: as usual, it's meant as positive criticism and by no means as an attack on your work or choices. Kudos for what you've accomplished, team, especially also for breaking some boundaries like the restrictions on the amount of resources. I'm rather posting this since I'm not sure you're really aware of the impact of the balance/AI issues as they stand and work out when the game is played for real, or that lots of new mechanics and a short turn time aren't the crucial factors anymore for deciding the worth of the mod. IMHO, if you can restore balance and AI competitiveness, it's potentially the greatest mod ever. If it's left unattended, the mod is made pointless and mostly a tedious matter of clicking through the years to win. I prefer the first option :)
 
Well, I kinda agree with you as I haven't had a challenging game for ages.

I'm thinking of moving to working on a mod-mod that essentially FREEZES the mechanics and assets at the point I branch it from C2C, and then:

  • Making minor mechanic changes for balance purposes (and onl for balance purposes)
  • Balancing the asset values in isolation from their values in the root C2C mod
  • Concentrating soley on AI improvement

Periodically I'd probaly pull in some more assets from the main C2C mod, and balance them to match the changes made to the ones I started with.

It would be my intent to make all ncessary DLL changes in the main C2C DLL and NOT to have that drift apart. Any new mechanics that C2C introduces would be turned off (either by not having XML that triggers its use, or by global define filtering of the fucntionality in the (common) DLL), pending a case by case decision process on how it would affect balance.
 
Well, I kinda agree with you as I haven't had a challenging game for ages.

I'm thinking of moving to working on a mod-mod that essentially FREEZES the mechanics and assets at the point I branch it from C2C, and then:

  • Making minor mechanic changes for balance purposes (and onl for balance purposes)
  • Balancing the asset values in isolation from their values in the root C2C mod
  • Concentrating soley on AI improvement

Periodically I'd probaly pull in some more assets from the main C2C mod, and balance them to match the changes made to the ones I started with.

It would be my intent to make all ncessary DLL changes in the main C2C DLL and NOT to have that drift apart. Any new mechanics that C2C introduces would be turned off (either by not having XML that triggers its use, or by global define filtering of the fucntionality in the (common) DLL), pending a case by case decision process on how it would affect balance.

So you would essentially make a "C2C lite" with all of the awesome DLL improvements but less content and assets, so those who wanted could have a greater challenge from the AI?
 
So you would essentially make a "C2C lite" with all of the awesome DLL improvements but less content and assets, so those who wanted could have a greater challenge from the AI?


Not exactly. I'd take (at least more or less) the full set of assets from some release version (latest when I do it, if I do), and then tune THOSE assets (iteratively and based on balancing feedback) in a more restricted environment, where some mechanics are turned off (stack limits for example). From that point on I would NOT add new assets from C2C except through a controlled process that involves a balancing and AI support step.

Part of the initial setup would be deciding what mechanics to restrict the mod to, and therefore what choices to leave. Some restrictions would be mechanic-on rather than mechanic-off, but the option would cease to be exposed to the user. The whole point is to provide a much smaller set of environments to tune against.

It certainly wouldn't be for everyone, because some people would almost certainly hate some of the choices I fix in the mod-mod (e.g. - if I decide it should be REV on, and so on). My intent would be to open with some polls on what choices to fix, but certainly it will have some of my own biases in it.
 
It certainly wouldn't be for everyone, because some people would almost certainly hate some of the choices I fix in the mod-mod

Though some might not like the choices, it does seem the most logical approach. Having a choice whether to play with revolutions on or off is nice from the players perspective, but essentially means you have to tweak the balance for two entirely different games OR choose a middle ground in the balance which is suboptimal for either option. Realistically speaking you probably can't both give the player unlimited control over which mechanics to use while at the same time giving him a finetuned game.

Another approach you might opt to take is a bit like the Magna Mundi mod for Europa Universalis III: that mod finetuned for a specified setting/starting moment they strongly advised to use but left the final choice to the player. You could likewise finetune for a certain default setting you strongly advise, and still allow people to alter some options they feel really strong about. For instance - I dislike Tech Brokering enough to ruin a game for me since it's so exploitable. If I turn it off but am smart enough to keep my other options to the suggested defaults, I can still profit from the overall balancing while knowing it's slightly less than optimal. The good part for you would be: less people will feel too restricted, which means more followers and less moans you'll hear, while at the same time you can profit from having to concentrate on only a restricted, controlled set of mechanics to balance. It comes with the price of players feedback being less valuable for you if they don't use the default settings, but if you communicate that clearly it shouldn't become a critical problem.
 
Well the main factor in the game being imbalanced is the conflicting feedback. People like Zaga here say there is too much gold and its not challenging anymore, even on the most difficult setting, while people like JosEPh_II say there is not enough gold and they hate the buildings that give -:gold:.

This makes it so my hands are bound in that I fear if I go one way or the other it will piss off people. Personally I think there is a bit too much gold and we could go with harsher gold penalties. However I wonder if the issue is not the game but the players.

Hear me out on this, we have players of many skill levels and of many play styles. Looking at the Let's Play videos has shown me how different people play and then in the feedback of what people do. While I cannot just say "your playing it wrong" since there is not wrong way to play C2C, there just are better ways to do things.

Those that complain about too much gold have become masters of this. So much so that even those without that much skill can amass a huge gold reserve. Thus if we are going by feedback I would say I hear more "too much gold" than JosEPh_II's "not enough".

What do you guys think? Should I stop holding back and tip the scales to where gold is more rare and feel the wrath of JosEPh_II? Or should I just not touch things have both sides pissed off?
 
What do you guys think? Should I stop holding back and tip the scales to where gold is more rare and feel the wrath of JosEPh_II? Or should I just not touch things have both sides pissed off?
Make a poll with 5 settings (far too much gold, a bit too much gold, the right amount of gold, slightly not enough gold, by far not enough gold). Then adapt accordingly.
 
Well the main factor in the game being imbalanced is the conflicting feedback. People like Zaga here say there is too much gold and its not challenging anymore, even on the most difficult setting, while people like JosEPh_II say there is not enough gold and they hate the buildings that give -:gold:.

This makes it so my hands are bound in that I fear if I go one way or the other it will piss off people. Personally I think there is a bit too much gold and we could go with harsher gold penalties. However I wonder if the issue is not the game but the players.

Hear me out on this, we have players of many skill levels and of many play styles. Looking at the Let's Play videos has shown me how different people play and then in the feedback of what people do. While I cannot just say "your playing it wrong" since there is not wrong way to play C2C, there just are better ways to do things.

Those that complain about too much gold have become masters of this. So much so that even those without that much skill can amass a huge gold reserve. Thus if we are going by feedback I would say I hear more "too much gold" than JosEPh_II's "not enough".

What do you guys think? Should I stop holding back and tip the scales to where gold is more rare and feel the wrath of JosEPh_II? Or should I just not touch things have both sides pissed off?

I might have a solution to this.

Since I couldn't work on multimap stuff today (pending some discussions that are ongoing with Lytning who owns that mod), I decided to revisit a suggestion I made regarding the gold imbalnce in another thread.

As suggested in that other thread on the gold issue, I have now implemented (as a default off, global-define triggered option for now, so that modders and people who want to can play with it without it impacting anyone else) a change to the way negative gold from buildings is handled. If you change the global define TREAT_NEGATIVE_GOLD_AS_MAINTENANCE (in assets/xml/A_New_Dawn_GlobalDefines.xml) from 0 to 1, it will cause negative gold to be handled as extra maintenance instead of subtracting from base commerce. The effects of this are:
  1. It removes the counter-intuitive effect of the actual negative gold increasing the more gold modifiers you have in the city (banks etc.)
  2. It makes the penalty subject to maintenance modifiers instead of commerce modifiers. Since these are easily manipulated on game difficulty (for now it uses the corporation maintenance modifier of the difficulty, but I can easily add a new tag to split buildings off to having their own) it is possible to tune down the gold at higher difficulty levels very easily
  3. Because maintenance modifiers are lower near administrative centers, it makes negative impacts on core cities less (but on outlying cities higher)
  4. It just seems more natural to model it this way (after all the only reason that builings would cost you gold is really through maintenance overheads)
 
What do you guys think? Should I stop holding back and tip the scales to where gold is more rare and feel the wrath of JosEPh_II? Or should I just not touch things have both sides pissed off?

Most definately!! :lol:
Thats why I have to change the games .xml's and give the AI additional bonuses and periodically give away free techs and stuff even on Deity for it to be a challenge- and I haven't played the mod for that long.. At first I had trouble with emperor, first game I won vs monarch lost a few to emperor then got the hang of which buildings were good and how to order production and stuff and then immortal became too easy around 10 games in.
 
@Hydro/Koshling: I think that, from looking at JosEPh's and other peoples' saves that the vast majority of expenses is maintenance already. Maintainance accounts for 3/4 of expenditures in the Finance screen in all of the games I've looked at, and Unit expenses account for the rest. This suggests that IF we want to decrease Gold (and that's a big if) we should look at unit upkeep as opposed to building/city expenses.
 
@Koshling

But the reason we changed from maintenance to -:gold: was the fact that so many buildings and then corporations ment that it killed the game. Basically corporations became something you did not want in your city because of their huge maintenance.

That was because when you specified maintenance in the building tags you only had maintenance MODIFIER tags to use. This doesn't add a modifier (so won't chnage the maintenance of corporations for example), it adds a fixed base amount (that the modifiers then act upon).
 
@Hydro/Koshling: I think that, from looking at JosEPh's and other peoples' saves that the vast majority of expenses is maintenance already. Maintainance accounts for 3/4 of expenditures in the Finance screen in all of the games I've looked at, and Unit expenses account for the rest. This suggests that IF we want to decrease Gold (and that's a big if) we should look at unit upkeep as opposed to building/city expenses.

All the negative gold on buildlings didn't show up in the finances screen (it just reduces the postive you see there from income), which is why it APPEARS that maintenance was already dominating (which acually it wasn't). The new system makes things more transparent and you can see where the costs really are (which makes it easier to balance too)
 
All I'm really trying to do is to keep C2C from falling into the Trap that AND fell into by the "Too Much Gold and Not enough Maint." Mantra the the "elite" Players pushed Afforess into. It Killed the Mod in the long run. Afforess said as much to me in a PM. While ppl like killtech had great ideas they went overboard. I don't want to see that happen here.

So I raise my voice in protest when I deem it necessary. I guess I can stop.

JosEPh
 
All I'm really trying to do is to keep C2C from falling into the Trap that AND fell into by the "Too Much Gold and Not enough Maint." Mantra the the "elite" Players pushed Afforess into. It Killed the Mod in the long run. Afforess said as much to me in a PM. While ppl like killtech had great ideas they went overboard. I don't want to see that happen here.

So I raise my voice in protest when I deem it necessary. I guess I can stop.

JosEPh

Yes, but this doesn't ADD to the costs, it moves them from one category (where they caused strange artifacts like making things look worse when you had banks etc.) to another. The net effect will depend on modifiers from other things (like difficulty level), so for some games it will generate MORE gold, and for others less. It also makes it more obvious where your costs ACTUALLY lie, and also makes it easier to scale things with difficulty level.
 
@Koshling,
My response wasn't to your new method Koshling. Sorry if it seemed that way. Actually your new method may be the ticket for better balance. Only using it will tell for me anyway.

No my response was to the posts prior to yours. My name is showing up in a lot of posts lately. I must be the "Mean Mr. Mustard" of this mod's development. ;) :lol:

JosEPh
 
All I'm really trying to do is to keep C2C from falling into the Trap that AND fell into by the "Too Much Gold and Not enough Maint." Mantra the the "elite" Players pushed Afforess into. It Killed the Mod in the long run. Afforess said as much to me in a PM. While ppl like killtech had great ideas they went overboard. I don't want to see that happen here.

So I raise my voice in protest when I deem it necessary. I guess I can stop.

JosEPh

Well that's one reason I have not gone either way at the moment. Your long time experience with RoM/AND brings wisdom to the table. There must be a better solution to keep both sides happy. I just don't know what. Which is why we are having this discussion in the first place.

In short your opinion carries more weight than a bunch of new people like Zaga who I have not heard from before.
 
In short your opinion carries more weight than a bunch of new people like Zaga who I have not heard from before.

Exactly, JosEPh has in the past pointed out some big balance issues such as the excess Inflation on Epic which I fixed a while ago. I trust his assesment on this issue, especially after looking at his saves.

@Koshling: Oh, I did not know the finance screen worked that way, I guess I learn something every day. How will this new system work with the Corporations, that was as Hydro said why things moved away from maintainence las fall.
 
In short your opinion carries more weight than a bunch of new people like Zaga who I have not heard from before.

I can imagine me not having posted a lot means that you give more weight to other people's opinions. But I'm aware there's a big discussion about the gold with strong opinions on both sides, so I tried to give some specific and objective points where the problem was felt instead of just yelling "there's too much gold" or "it's too easy". If you can't tell, I gave some thought about what I'd write and put effort into adding something useful to the discussion. If you aren't impressed by my amount of posts, at least consider the argument itself :)

But don't focus specifically on that gold-issue. The main point I tried to make was about general balance and AI. The gold is merely one aspect of that.
 
Top Bottom