1.9.5 Feedback

5% shift down from Emperor onward
That would probably do it. Maybe have Deity get 10% shift down (I seem to remember someone posting that unit spam on Deity was lower than what would typically be encountered - I assume in comparison to vanilla civ).
 
The idea of "adaptive difficulty" is discussed for many different games. If you have an option to *decrease* the difficulty level if the player is doing badly, I think it dilutes the achievement of actually winning. The game is giving you a break when you mess up.

I think *increasing* the difficulty has a similar problem.

I agree with David. I think if you want a more challenging game, you turn up the difficulty level for your next game. Don't change the difficulty level during the game. If you get a tough start, you have a tougher game, or you lose.
 
I agree, too. Sometimes I restart a dozen of times to get a "sweet spot" or, more generally, a starting location of a certain kind. Maybe I just feel good with the AI of a certain level in all respects except of the beginning, and want to "correct" it for me this way. This behaviour would be impossible with "adaptive difficulty".

It would also be susceptible to exploiting, once the criteria for the game routine are known.

Also, it's hard to adapt to all the different playstyles. Even the designers of, say, WoW, failed very frequently to unexpected "inventions" by players. Even "only" hundreds of Players will typically be smarter than you could be.

Furthermore, I'd simply hate it to not know the current difficulty setting.


Final plus: It's a complete waste when thinking of multiplayer. To which player would you adapt?
 
Well, to say the least, I guess I can classify the reaction to that suggestion as 'negative' :lol: (did I mention the word 'option' in there). While I do think this has more potential than the general consensus here has given it(and I don't think what I generally had in mind was well understood), I do also think it's also a minor point that would only bring a small enhancement to a small percentage of players. Axed, and... moving on to more important things.

Right now I'm taking a close look at AI tech decisions, seeing what techs are over/under valued, what likely orders the techs are being researched in (taking flavor values into account for different leaders, as well as things like economic conditions and warplans/current wars, etc.). I've written up code to give me useful feedback on the current valuation of all possible tech choices by any given AI on the given turn, and output it into the help text area so I can do this more efficiently.

The main thing I'm looking for is techs that should be considered valuable that aren't (again, considering the civ in question and various conditions) and those it probably shouldn't consider valuable but yet it is. Of course the ultimate goal is decently situationally intelligent choices. If there is no immediate threat (and they aren't planning on making themselves a threat to someone else), they should be making sure they are seeking techs to grow their economic and science output, while making sure at a minimum their defensive units are up to snuff. If they are planning or in a war they should be valuing more powerful units higher. If there is a religion available to be founded that fits the civ (I'll be giving religions 'flavor' values like techs have) it should be as willing as a human to try to take a shot at beelining it. It's a real balancing act between smart decisions and unpredicability.

Fortunately the things I'm aiming for aren't anything new, it's how it's already coded (in principle). So what I'm really doing is not re-writing the code, but finding things that aren't implemented well and finding better ways to value them. An example is the improved well and mine improvements. Around the time that Planetary ecology comes within three 'path steps' of being researchable (the maximum number of techs the AI is programmed to look ahead), it was uniformly becoming one of the top valued techs to shoot for by pretty much everyone. While it is a decent choice, it's not that much better than the other techs it was competing and winning against.

The primary reason is the deep well. The human knows that while the deep well gives x additional water in a tile, it's really only 1 more water than their current shallow wells give-not that big of a deal. The AI was looking at it without this perspective, instead seeing it as really good because it gives all this extra water!(which the AI rightfully values more than other improvement yields). To correct for this overvaluation, I added a tag for improvements to indicate if it is an improved version of an other improvement essentially, and the AI now looks at the deep well in terms of what it adds on top of a shallow well improvement, just like a human. This same tag is also used for the improved mines.
 
Right now I'm taking a close look at AI tech decisions
Something to keep in mind if you're doing simulations is the very deliberate manipulation through flavor values, particularly for religion founding and the like. Ix will pick industrial techs more; Fremen will get ecological techs more, Bene Gesserit will favor religion and culture techs, Ecaz will favor economic techs and will try to found CHOAM, Corrino will try to found Imperial, etc.

I don't think we want all the AIs picking out the same tech paths, and it is a good thing to have some faction-specific flavor even if it is slightly suboptimal.

The primary reason is the deep well. The human knows that while the deep well gives x additional water in a tile, it's really only 1 more water than their current shallow wells give-not that big of a deal. The AI was looking at it without this perspective, instead seeing it as really good because it gives all this extra water!(which the AI rightfully values more than other improvement yields). To correct for this overvaluation, I added a tag for improvements to indicate if it is an improved version of an other improvement essentially, and the AI now looks at the deep well in terms of what it adds on top of a shallow well improvement, just like a human. This same tag is also used for the improved mines.
This sounds good; the over-writeable improvements are a special case.
 
I'm actually paying a lot of attention to how the flavor values impact tech choices. In fact in many cases the current numbers are too conservative to have a meaningful impact and I'm adjusting them accordingly as needed. This is a lot easier and more accurate to adjust now that I can directly see the comparative tech valuations for different leaders in real time in a game.

As I casually alluded to above, I've also given flavor values to religions that ties more directly into how likely a leader is to go for a religion founding tech if the religion is unfounded (unlike the flavors in the tech infos which encourage that tech choice even if the religion is already taken).

In these cases I'm not removing the tech flavor, as it's still flavorful for that leader to get those techs generally even without the religion, but tying their flavor values directly to the religion founding in the tech valuation is generally a better method for what's trying to be accomplished there.

I'm also having too use the iAIWeight techinfos tag in special cases where a tech is routinely over or under valued, but changing the actual tech valuation code would upset the balance of other similar techs. The only real downside of this is that if we ever rearrange the tech tree significantly again, cases where this tag is used would all need to also be reevaluated.
 
I'm actually paying a lot of attention to how the flavor values impact tech choices.
Ok, it's cool that you're going through all this stuff, all I ask is that you keep some of the broad flavors in mind from the existing design.
Some of the religion flavors were used not just for founding techs, but for techs that give benefits to those religions (eg: that allow religious units or buildings), or that have associated flavor.

It sounds like what you have in mind is much more sophisticated than the brute force approach I used before.
 
All those are staying in, and I'm only strengthening some flavor associations where they are too diluted to really impact tech path. Most instances work quite well, there have just been a few cases where I've made any major (more than a shift +/- 1) changes to tech or leader values, while sticking to all the original intents as close as can be deduced.

As a general rule, tech values before adding in flavor and the general random factor (0 - 2000) and then being adjusted for tech cost are around a standard deviation of something like 1500 - 4000 lets say (rough estimate). Flavor values multiply the (leader flavor * tech flavor * 20), so if both are 10, the flavor value is 2000. However a tech with 2 and a flavor with 2 adds 80, which isn't enough to make any meaningful statistical dent in tech choices unless two potential techs are very close in value otherwise. Of course it's not supposed to make a huge difference in some cases, while a larger in others, but anything adding below two or three hundred can basically be ignored in the big scheme of things.

(I know I technically used standard deviation wrong above, but I'm presuming you know what I mean ;). To clarify, a rough estimate of the min and max values within one standard deviation of the mean base tech value are somewhere around what I said)
 
I get the idea. Mostly I used the vanilla flavor values as a guide. I didn't really want super-strong changes except for the religions, I didn't want to interfere with the AI optimization too much.
 
While making observations on AI tech paths and such, I also looked a little into Spice/Paradise Civic values and decided to dig up an old idea:

+ Have civilizations decide at the beginning of each game whether they are going to pursue Arrakis Paradise or Arrakis Spice for that game. Give each leader a pro-Spice probability to control this:
Never adopt Paradise: Bene Gesserits, Ecaz
80/20 spice: Corrino, Harkonnen, Ix, Ordos
80/20 paradise: Leto I, Alia, Tleilaxu?, others?
Never adopt spice: Liet-Kynes, Stilgar, Leto II, Muad'Dib

I added LeaderHead tags for probability to go spice, neutral (fairly rare), or paradise. The AIs will roll a decision based on those probabilities at the beginning of a game and will only give value to that civic, unless vassalized (in which case they'll value them all equally so that any demands from their master should stick). The only real problem that existed in it before was for leaders whose 'hated civic' was not spice or paradise it was pretty much a crap shoot which they'd pick with no real value to distinguish one from the other. It also meant that even if paradise was the leaders hated civic, if they researched Way of Liet before Spice Industry, they'll go Paradise until they get Spice Industry. I don't think that's intended or optimal.

As for values for all the leaders, I tried to stick to their flavor as much as possible. Harkonnens, Ecaz, Corrino you can count on going spice, Fremen, Leto II, Alia paradise, Nobody knows what the Tleilaxu will do (though they are slightly favored towards spice). Bene Gesserit will never go Paradise, but lean more towards neutral than most. As that indicates, some have a small chance of staying neutral, which I think is good for flavor, but maybe not for AI performance. I'm thinking of adding a handicap tag to alter the odds of spice neutral AIs, where at higher difficulties it's much less likely, but at lower it's a fair possibility.

I know the general design is to encourage picking a side here, but even players will keep neutral for diplomatic/flavor reasons sometimes (or at least I do), so I'm not decided really on how much neutral odds I'd allow in the end. But without a neutral chance, most leaders become all one option or the other, as very few feel right taking either or (spice or paradise).

Duke Leto I gave a 40% of neutral (but still 60% spice) since flavor wise he'd want to keep good terms with the fremen. What I'm considering doing to account for this kind of thing is modify the odds if the leader's favorite civ is in the game based on that civs favorite/hated civics if they are Arrakis civics. So in other words, going this route, he'd only shy away from spice if the fremen are in the game (staying neutral instead).
 
As that indicates, some have a small chance of staying neutral
I think for the uninclined, I would tend more towards equal chance of going either way than much of a chance of staying neutral.
Maybe, 40/20/40 spice/neutral/paradise?

I wouldn't block Leto I from paradise.

What I'm considering doing to account for this kind of thing is modify the odds if the leader's favorite civ is in the game based on that civs favorite/hated civics if they are Arrakis civics. So in other words, going this route, he'd only shy away from spice if the fremen are in the game (staying neutral instead).
I think I'd stay away from this kind of thing. I already think we're risking making the alliances a bit too deterministic.
 
My thoughts is that for the base mod we need to keep Arrakis civic choices as close to plausible for what that character might have done, we can create an option easy enough that adds more randomness to it. Duke Leto's concern is the survival of House Atredies, and he'd only really support a policy of transforming Arrakis if he felt it'd be the best way to achieve that goal. I can see a little wiggle room to justify a rare move towards Paradise, but it's very tenuous. In such cases I'd like to err towards the the side of caution. Seeing the good Duke valiantly leading his House in the transforming of Arrakis more than a few times would leave a lot of players going WTF? I would think.

Like I said, for the sake of those wanting less predictability let's include an option that opens it up a little bit and gives more variation.
 
My thoughts is that for the base mod we need to keep Arrakis civic choices as close to plausible for what that character might have done, we can create an option easy enough that adds more randomness to it.
This kind of thing has been thoroughly debated throughout the history of the mod. The general consensus (which I support) was to hard-code only a handful of things. Part of what people enjoy about Civ is alternate histories, we don't just want to try to recreate a canon set of alliances. It can get boring if the Atreides and Fremen are allied every game.
We have a few hardcoded things; hard-coded vendetta relationships between some factions (though I would really like to see this able to be disabled with a single option), hardcoded religion access (though again I'd like to see this able to be disabled with an option) and some hardcoded AI flavor preferences. But I think we should not go beyond that, and I strongly think we shouldn't start getting into *conditional* flavor modifications (Faction A has preference X only if Faction B is present).

There were some very heated debates. Some people very strongly wanted no hard-coded restrictions at all. We ended up with a compromise that people could live with. I think we should respect that compromise and not start shifting to an increased degree of hard-coded plot.
 
Well, spice civic preferences aren't a new thing, as most civs have been 'hard-coded' to one or the other by the 'hated civic' mechanism up to this point. It's what makes the Fremen always go paradise and the Harkonnen always go spice for instance right now, and I think making sure there is a common sense set of preferences in this respect is the right path. I agree everything shouldn't get too hard-coded, but there is a certain level that's necessary to keep the flavor of the different civs, such as the Fremen not being able to found CHOAM, that serves the mod very well. But the purpose of this code change isn't even to add new restrictions, it's just to make one already placed (on the AI) work more coherently.

But I doubt we're really disagreeing much that there should be preferences to an extent. The discussion is how detirministic those preferences should be. The pattern I've followed is that any leader that has spice/paradise as it's hated civic has a zero chance of adopting the hated one, with varying shades of chances to go neutral. That mostly leaves the guys in the middle ground to decide on. That includes Goya and Scytale, both of whom I've roughly split down the middle with little odds of neutrality, Duke Leto, and Lady Margot Fenring. Both Ixian Leaders and the Ecaz leaders don't have a hated Arrakis civic if I recall correctly, but neither are good candidates to have much chance of going Paradise?

Ultimately, I'm pretty happy with the way the current setup tends to only have certain leaders that make sense to go paradise do so, while most of the Houses are going to stick with the 'spice' block. It makes sense to players and I don't want to suddenly have terraforming ixians on a regualar basis :crazyeye:. All of that I'd like to stay roughly where it is, I just want to iron out some holes in the system and make sure the AI knows where it's going with these civics and researches and prioritizes accordingly to whatever extent it can be made to. And like I said, from there an option to make it more unpredictable is quite doable. We can add options to disable vendettas to, that's even easier.
 
But I doubt we're really disagreeing much that there should be preferences to an extent. The discussion is how detirministic those preferences should be.
Right, I agree that the discussion is just about a matter of degree.
I just wanted to make clear that there are some very real downsides to going too far, and that were were some mod contributors who were very emphatic that they didn't want any hard-coding at all.

I think it is ok to have some hard-coding (as the current model does), and I think that this should probably tend to lean more towards spice than towards paradise.

So, maybe something like:
Fremen, 100% Paradise
Atreides. 100% Paradise except for Leto I, 40% Paradise 60% Neutral
Bene Tleilaxu 40% Paradise 20% Neutral 40% Spice
Ordos 40% Paradise 20% Neutral 40% Spice
Ix 20% Paradise 20% Neutral 60% Spice
Bene Gesserit 20% Paradise 20% Neutral 60% Spice
Ecaz 20% Paradise 20% Neutral 60% Spice
Corrino 100% Spice
Harkonnen 100% Spice
 
I'm not opposed to the odd chance that Ix or Bene Gesserit (at least Margot Fenring) should go Paradise, but one in five games is a bit high. For the most part I think the default option should be somewhat more predictable. We really want players to see they have such and such next to them and have a real feel for what they will do, big surprises should be quite rare and feel like a surprise.

Now remember, I'm talking about the default option here and there will be an option for less predictable spice civic choices, but for the purposes of the base option there should be three 'archetypes' that leaders fall into with variation within them:

  • Spice Leaders
    • this includes 100% spice leaders like the baron, leaders that are mostly spice, but can lean neutral, and in some cases have a 5-10% outlier chance of Paradise. Some of the more diplomatic leaders could be nearly evenly split Spice/Neutral.
  • Unpredictable Leaders
    • these are the guys that you have no idea what they might do, and should be a short list. The Tleilaxu, to a lesser extent maybe the Ordos.
  • Paradise Leaders
    • these are the Fremen obviously and Leto II and Alia, and much like Spice leaders should only have small chances to vary from the usual path

The key difference I have with what I want (for the default option) and your proposed numbers is I want players to have a real feel for who is what. If there are too many 'gray area' leaders, and I'd argue 80 spice/20 paradise is really grayer than you realize in a player's mind, it feels too chaotic to tentatively plan around. That's fine for an option when the player wants it. But the Ix just to take an example should probably be 85/10/5*. Gaius Helen of the BG hates the Paradise civic, so shouldn't ever go spice. the BG should also be less gung ho about spice than say the Ecaz and more diplomatic. I'd say Gaius should be maybe 60/40/0 or 50/50/0. Lady Margot could by contrast possibly go paradise, but rarely (the reverend mothers do depend on the stuff), lets say 50/40/10.

But as a general rule, nobody in the 'Spice Leader' category should be over 10, maybe 15% paradise as an extreme. Just to really summarize the point once more, I'd support more predictability for the base mod, with the option for less. We can also consider an option that removes all flavor from the decision and makes everyone 33/33/33 if people want a real wild ride :lol:.

Leto I, 40% Paradise 60% Neutral

Did you mean this? Everything else in Leto's personality is CHOAM heavy, which really means diplomacy and spice trade. This would be a heavy departure from that. It doesn't mean he has to be 100% spice by any means, but I was thinking maybe around 30/60/10?

* in case it's not obvious, anywhere I've got an x/x/x, it means spice/neutral/paradise
 
And just to clarify, it's actually not extraordinarily important to me wether the looser or tighter interpretation is the default or not in the mod. We can make the standard less predictable with a 'cannon' spice civic option. But my feeling is that the normal new player will be better served by a more predictable, more cannon version, while lower predictability is included as an option for more re-playability for veteran players that prefer it.
 
For the most part I think the default option should be somewhat more predictable. We really want players to see they have such and such next to them and have a real feel for what they will do
Why? It isn't that predictable what civics or religion or strategy will be taken by particular leaders in vanilla Civ4.
I think rather the opposite, I think the game is more fun if you can't always predict what the others will do, and instead need to have a decent element of basing your strategy on what choices they make, rather than being able to do this pre-emptively.

While I think there should be an anti-Paradise bias, it shouldn't be overwhelming, if it is 4:1 in terms of faction choices then the Paradise players are going to get crushed by alliances almost every game.

I think the concept of Paradise, Unpredictable, Spice leaders is fine, but I think it is a problem if you put too many leaders in the Spice category.

Did you mean this? Everything else in Leto's personality is CHOAM heavy, which really means diplomacy and spice trade. This would be a heavy departure from that. It doesn't mean he has to be 100% spice by any means, but I was thinking maybe around 30/60/10?
If you start doing this, then there are only going to be 5 paradise leaders left in the game, 1.67 / 9 factions. I think that is too low. The Paradise/spice mechanic starts being uninteresting and unappealing if almost no-one is willing to adopt it.
I think it is more important to set these by faction than by individual leader.

Anyway, it would be great to hear what some others think on these issues.

*edit*
Part of the reason for the semi-paradise favor of the Tleilaxu is the idea that they are the ones who will/might develop artificial spice, and so have rather something to gain from constraining natural spice production (to increase price on their artificial spice monopoly).
 
The reason I keep stressing the option thing is that we have separate views that, in this case, can't be compromised and create something that meets what either of of is looking for. The difference in our opinions between wether it should prefer cannon or unpredictability is also going to be shared by players, so opening up either possibility is the best route. But one of the routes has to be a cannon route, not everyone (me included) will think it makes sense to see Ixians or Ecaz going paradise on anything but the very rare occasion (as an example). It would annoy the hell outta me if that was the only option we had.

Now, I've said it before, but I'll revisit it, I'm trying to keep the same feel that we had before for the most part in terms of spice vs. paradise. In the mod at present, it's very uncommon to see anyone but the two relevant atredies and the fremen stick with paradise if they can take spice. I don't remember any games where I've seen someone else go paradise into the late game. And that works well. There's a big diplomatic danger of trying to terraform the planet, that diplomatically tricky spot is part of terraforming-as it should be. You are threatening to destroy the spice in the process afterall...

In practice, the odds based approach I've added already increases the odds of any other leaders/factions outside of that small group of always paradise'ers of taking paradise. It also makes it possible and likely that there will remain a few neutral parties, which also benefits the paradise guys diplomatically over the current setup. Your proposed odds represent a larger divergence from the status quo by contrast that I don't believe we should be taking in the base game.

I also definitely see where more unpredictable behaivor in this regard can be a fun route to go. I think it would be a fun alternative... some of the time. Some people would probably always play with it on. But I can't support it as the only option.

Part of the reason for the semi-paradise favor of the Tleilaxu

I definitely agree the Tleilaxu should have a healthy chance of going paradise.
 
My opinion :
1. The mechanic is fine as is. Any valid terraforming strategy always involves getting a leading position first, then switch to Arrakis Paradise (100% win rate). Going Arrakis Spice you can switch immediately but it seems there is enough inherent conflict (religion etc.) where spice factions still war with spice factions - I've never seen spice block vs terraforming block. In my current game, I have pleased relations with several Arrakis Spice civs even though I am Arrakis Paradise. I really don't see what needs fixing. P.S. Nobody goes Spice Neutral - that's just dumb (and poor strategy).

2. If you want to "make a better system", fine (I'll game any system you come up with anyway :p). I do see merit in making an option to play standard(current) vs alternate (hell, maybe even more fun) system. Choice, IMO, would be the best way to deal with it. Should satisfy everyone. Will try whatever you come up with - it's always fun to try out new things. :)
 
Top Bottom