Introducing the Medieval Pack II

[XMon]

A couple more thoughts...

If there are no new sprites in the WWII scenario then they may as well just keep it as far as I'm concerned.

Wes, would you consider having the militia units be one step behind the current defense unit? Militias tend to be more along the lines of armed citizens and thus would not have state of the art weaponry. If you just discover machine gunners then all your cities get infantry militia etc.

How about losing the "City Wall" graphic in the modern age? You can keep the defense bonus effect but a wall surrounding a modern city looks weird.

I know you had concerns about the tech tree in Med Mod 4 causing the game to be unstable but I must say that my last few games of MM4 were quite stable. Why? Don't know really. It had Gedrin's latest version of the Dissemination Trigger (better alternative to auto upgrade IMHO), and I removed rail launchers and a few futuristic units. No crashes at all. I do hope you include an enhanced tech tree. The default tree is a mess.

[WesW]

quote: Originally posted by wheathin on 12-11-2000 10:22 AM To keep the AI in the game (or help a straggling human), Harlan's got some good ideas (catch-up wonder, etc.) Here's more: 1. The tech-sharing from CtP1 MedMod 2. See my thread on upgrading obsolete units 3. Bigger DiffDB.txt bonuses for AI in late game 4. Increase the odds that AI starts with two or more settlers (can the AI start with three?!) 5. SLIC: every wonder increases the rate of tech diffusion from the civ that builds it? Each wonder attracts tourists and pilgrims - it enhances tech transfer from that advanced civ as the pilgrims return home. I.e., each wonder functions as a Great Library in reverse. 6. Feats of Wonder geared towards the AI? We tend to want feats of wonder that reflect our own building styles, but surely there are AI play strategies that could be rewarded. First to move a unit into and out of the same city for 20 turns in a row. Seriously though, there's gotta be something... 7. One wonder per age per player

It may be easier to adjust the penalty given the human in the diffdb than to try and slic a bunch of stuff to help the AI. Btw, the feat of wonder proposal is hilarious.
I am getting worried about having to use slic for any and every adjustment we want to make to the game. I am going to try and find other means to solve problems whenever possible, and save the slic for big things that can't be addressed any other way.

Also, I have found out how to implement Harlan's new city styles into the game. I am going to wait about adding the Industrial age for now, unless Harlan or someone else figures it out, then I would be happy to drop it in the mod.

I started a new game last night to test out the city styles, and it was immediately apparent that setting wages to a base of 12 was too high. Therefore, I have set them to a base of 8 with each notch changing the level by 2, which is midway between 12 and the original setting of 4. This should give good results, and we can tinker with building settings to further refine things.
I also reduced the base workday from 10 to 8, with each notch representing a change of 2 hours. This was what things were in Ctp1. Unit, building and other costs are going to be changed in the mod anyway, so we can adjust them to fit this new setting.

Finally, if someone would change the necessary text files so that each pop only worked an average of one tile in the first radius, I will put this in the mod, and see if helps address the issue of cities growing too fast early on in the game. I think the settings are in strategies text, but I could be wrong. If you do send me something, please specify what changes you made to which lines.

Btw, can anyone confirm or deny colorme's statement that the Min and Max ahead and behind bonuses are applied linearly throughout the game, rather than being triggered either on or off when the AI crosses the specified threadhold? This is a big deal.

Most of the time here has been spent discussing military units, which I enjoy as much as anyone, but we need to get these "domestic" issues of growth and production nailed down first, as well as getting building and tile improvements cost-to-benefit ratios evened out.

[wheathin]

Food issues:

For initial city growth, if there is not a net positive amount of food in the first 8 squares + center, the city can't grow. So, if you give certain types of terrain *negative* food values, you might be able to tailor better which terrains will support cities? I.e. give tundra a negative value high enough so that if a city has 4 tundra squares adjacent to it, it just can't grow at all, no matter what kind of land it's on. After all, any city located near that much tundra is just too cold - odds are it's nearby beaches, climate, forests, etc. are too hostile to support a multi-million pop city...


BTW: Is the mod gonna end up changing everything back to the CtP1 values? Just curious - it seems that's where we're going.

[colorme]

Wes,

In DIffDB, I read a comment line which talked about the interpolation business, and hence my suggestion.

Also, I mentioned this in some other thread, that in most CTP2 games I've played, the AIs seem to have huge ups and downs in their power graph, often over just a few (50 or so) turns. It isn't clear to me why this should be so. I usually end up with a pretty linear power graph, usually with minimal downs.

Do the AI do badly because they fight with each other a lot? But then, given that they're incapable of a massive attack, whereas their defense is pretty good, the huge ups and downs in power don't seem right.

Anyone have a clue about this?

[wheathin]

The powergraph is highly weighted to military units. It could be the AIs are building large armies and then having them destroyed in combat. It could be that the same few cities keep changing hands. Or, it could be that once an AI takes a few cities, it is willing to let a victim make peace. Later, the victim regroups and takes back its losses.

Solutions?
Generally, the AIs have to be more willing to go for the kill - if an enemy is down, keep kicking them. The little nickle-and-dime attacks are wasteful and don't get much payoff for the invested time and production.

[lago]

A lot of AI civs build massive armies, and there's no retreat option for them. So when it's 12 against 12, and the battle goes down to the last hit point, that can show up pretty clearly in the graph.

Just a guess.

[Alpha Wolf]

I'm having fairly good results with lowering the initial production values of a city. Under the current production model, a new city gets the full production of its tile, plus a city bonus, plus 1/8 of of the surrounding area. I cut the city bonus to 5f-10p-10g. This seems to greatly slow the growth of cities. I also gave snow and desert tiles 0 food. I'm also playing with the way the AI decides where to settle so that they wont start cities that cant be supported. You can also determine how closely the AI packs its cities. A major prob i saw was that AI cities would grow rapidly but then stagnant as multiple cities tried to share the same tiles.

------------------
History is written by the victor.

[Alpha Wolf]

Also, I mentioned this in some other thread, that in most CTP2 games I've played, the AIs seem to have huge ups and downs in their power graph, often over just a few (50 or so) turns. It isn't clear to me why this should be so. I usually end up with a pretty linear power graph, usually with minimal downs.

Do the AI do badly because they fight with each other a lot? But then, given that they're incapable of a massive attack, whereas their defense is pretty good, the huge ups and downs in power don't seem right.
---------------------------------------------
I've noticed this in my graph after losing major battles, or the lose of a city, especially a powerful one containing one or more wonders. BTW, I have been attacked by a stack of 8 which totally caught me by surprise.

------------------
History is written by the victor.
[This message has been edited by Alpha Wolf (edited December 11, 2000).]

[Daniel Frappier]

Hi,

Taken from DiffDB (Impossible)
----------------------------------------------------
AI_MIN_BEHIND_PERCENT 1.8
AI_MAX_BEHIND_PERCENT 0.8
AI_MIN_AHEAD_PERCENT 2.0
AI_MAX_AHEAD_PERCENT 3.0

AI_MIN_BEHIND_GOLD_ADJUSTMENT 1.0 1.0 1.0 1.0 1.0
AI_MAX_BEHIND_GOLD_ADJUSTMENT 1.5 1.5 1.3 1.3 1.1
AI_MIN_AHEAD_GOLD_ADJUSTMENT 1.0 1.0 1.0 1.0 1.0
AI_MAX_AHEAD_GOLD_ADJUSTMENT 0.9 0.9 0.9 1.0 1.0
----------------------------------------------------

wich means i believe: if the AI is rated at 3 times or more the strenght of the human then use AI_MAX_AHEAD_GOLD ==> the AI will have a 10% penalty for gold.

if the AI is in between 2 and 3 times the strenght of the human then AI_MIN_AHEAD_PERCENT is use ==> no bonus or penalty.

if the AI is in between 1.8 and 2 times the strenght of the human then nothing special.

if the AI is in between 0.8 and 1.8 times the strenght of the human then AI_MIN_BEHIND_PERCENT is use ==> no bonus or penalty.

if the AI is less then 0.8 times the strenght of the human then AI_MAX_BEHIND_PERCENT is use ==> the as a 50% bonus for gold.

It's the same principle with Science & Production. Me to i have won the first game i played (at impossible) but like i mentionned to WesW before, the AI is not as bad as it looks but with those settings it simply can't and will never compeat with a human player.

In my game I play with:
----------------------------------------------------
AI_MIN_BEHIND_PERCENT 1.5
AI_MAX_BEHIND_PERCENT 0.85
AI_MIN_AHEAD_PERCENT 1.5
AI_MAX_AHEAD_PERCENT 2.5

AI_MIN_BEHIND_GOLD_ADJUSTMENT 1.1 1.25 1.4 1.55 1.7
AI_MAX_BEHIND_GOLD_ADJUSTMENT 1.4 1.6 1.8 2.0 2.4
AI_MIN_AHEAD_GOLD_ADJUSTMENT 1.0 1.05 1.1 1.15 1.2
AI_MAX_AHEAD_GOLD_ADJUSTMENT 1.0 1.0 1.0 1.0 1.0
----------------------------------------------------

and something similar for science & prod and the AI do LOOK much better.

[Celestial_Dawn]

Thank you Daniel. Based on what I've seen of DiffDB I am in agreement with your observations.

Wes, the scope of your mod looks great! I am in agreement with Harlan however as far as the "less is more" approach - minimalist changes are better in this case if the timeline isn't going to be extended.

CtP timeline seems to have a lot more turns, so adding one or two units here and there shouldn't be problematic. However, you're talking about lots of new units ... it'd be better I think to start with a few, and then scale the timeline up progressively as you add more - and only if necessary ... don't want game to be too long.

And my humble suggestion - perhaps basic game balance should be addressed first before worrying about adding new units ...

That said, can't wait

And sigh - anyone know where I can find a tech tree? Scanned? I don't have a foldout, and it'll save me having to reconstruct the entire tree by hand ...
[This message has been edited by Celestial_Dawn (edited December 11, 2000).]

[lozina]

about cities growing too fast- i think i made modified by files just the way i like it... firstly- i reduced the percentage of plains and grasslands and included more inhospitable terrain like deserts and white tiles. then i made deserts, mountains, white tiles and sea tiles negative food value (except beaches and shallow water which gives 0-- i hated the huge megalopolises on a shore of a barren wasteland) but when u add nets you start getting food from water tiles. makes sense dosent it? by that same token- i made plains give 0 food and grasslands only 5. jungles & swamps give 10 food and forest i raised production. why do jungles give 10 food and plains 0?? well- think about it- a plains by itself is just grass. unless you're a buffalo its not useful to you. so in order to get food from plains you must have farms. so now i dont have huge cities growing in some barren plains with no farms- with original model i can get a size 30 city WITHOUT any farms in plains and i found that quite ridiculous. but now- my cities grow much much slower- but more realistically. if you work on a city with improvements and such you can get a nice size 12 in the early age- otherwise if u neglect it- you'll only get maybe a size 4. it really does work great for me. another thing i hated was how advances are researched so fast. by time i discover fighters and a few turns after i build my first fighter i get the intercepter technology. thats just silly. well good luck!

[wheathin]

Another AI-heleper idea that doesn't require SLIC: how about giving the capitol improvement a few extra bonuses, such as a city-wall benefit? This has to help the AI, because how many human players lose their capital in a game? And having or losing the capital is such a huge blow...

Perhaps also, to mimic the extra size of the capital, (as the seat of government, court, etc.) the capital could get a food or gold bonus?

[MarkG]

quote: Originally posted by Celestial_Dawn on 12-11-2000 11:15 PM And sigh - anyone know where I can find a tech tree? Scanned? I don't have a foldout, and it'll save me having to reconstruct the entire tree by hand ...

http://apolyton.net/ctp2/downloads/u...treeposter.pdf

[phenyl]

lozina: the thing with having specific country style units in world war 2 is that they are going to be very similiar in game play. maybe there should be a different sprite loaded for units with maybe a small amount of tweaking stats. depending on the civ who owns them? the idea about limiting unit production to cities that have specific advances is really good. something like that was done in the end game of ctp1, so for sure it is possible!

re:fire. that is a great idea. i suspect that it would invovled slic, but having fires (set or otherwise) as well as other natural disasters triggered randomly would mix up the game a bit. they (i am thinking earth quake) could affect terrain, cities and improvements too!!

pillboxes would basically mix tile improvements with unit abilities. ?? the only way i know of doing this is locking a unit to one location. the auto-defenders in cities that would die when moved are a good example. not really what you are suggesting as you could even hear them die when you moved em

xmon: your ideas are very good. the idea of tweaking combat strategies (flanking, etc) is good. this could also be used to introduce combined arms tactics if it were reliable.. (modern units working better together than seperately)

ship to ship combat seems like a pirate kind of thing to me esp the potential to damage/lose the ship, etc, potential gain cargo (gold), etc... pirates could be a barbarian only unit, along with other "anarchy" types (bandits, etc..)


wesw: i think it is a very good idea to pre-package the sprites ahead of time. even going so far as to pre-define empty sprites that have redundant graphics for later extension ... (free variables) if its that much easier to package at once..... this would also allow for a degree of flexibility if people were ever going to try and combine mods.

[lozina]

phenyl- on limiting unit production to cities with necessary buildings-

i was fooling around the alexander the great scenario included in the game and have found it does something similar. the scenario has specific cities allowed to make the triremes... these cities were predetermined by the creator. I've been studying the code and am trying to modify it so instead of checking for specific cities- it checks if the city has the buildings required in order to build that unit. if i can get it to work ill need help with a rational and realistic 'requirement tree' for units in the game (i.e. maybe in order to build a caravan that city needs a bazaar, catapults require mill and academy? etc..) ill let you know if i get lucky and it works!