Abolish the distinction between Arcanus and Myrran races

[Nelphine]

From what I recall, map generation is incredibly complex. Its also very hard to define what is bad and good, as it changes from game to game and setting to setting. You might be on the smallest island, but if it can fit 12 cities, there's no real problem. You might be on the biggest continent, with a great capital - and only desert for 20 tiles in all directions past your capital area.

I've seen a game where the AI capital was not on a shore, he was chaos/life (never got chaos channels), and the continent had enough room for 3 more cities - but all 3 other spots were blocked by neutral cities on a different continent.

[Seravy]

From what I recall, map generation is incredibly complex. Its also very hard to define what is bad and good, as it changes from game to game and setting to setting. You might be on the smallest island, but if it can fit 12 cities, there's no real problem. You might be on the biggest continent, with a great capital - and only desert for 20 tiles in all directions past your capital area.

I've seen a game where the AI capital was not on a shore, he was chaos/life (never got chaos channels), and the continent had enough room for 3 more cities - but all 3 other spots were blocked by neutral cities on a different continent.

Yes, that's one of the problems, it's hard to define what's good or bad.
The island might be bad for resources but it's easy to defend and less likely to get attacked by hostile but not at war wizards. The large continent has lots of resources, but you're open for attack from all 4 sides and make lots of early contacts, risking an early war with 2-3 wizards at once.
A capital with a max pop of 8 in the mountains sounds horrible...unless you plan to spam 5-6 settlers from it in which case the production bonus is helping you tons to speed things up and you won't care about the size of your capital as you're relying on the other 6 cities instead.

And ultimately, it makes minimal difference to have that adamantium in your capital, or 6 tiles away from it where you send your first settler.

But let's say we put the AI on the largest landmass, in theory. How? Do we put all of them on the same one? That's even worse than a small continent. Go in the order of largest, second largest, third largest? Then no one ever shares a starting continent, that's boring and also the first wizard is at an advantage - they always get the largest continent. But there is more, what if the continents are sized 150, 150, 22, 20, 17 ,8 ,6 ,3? Then the third wizard is still on an island.

[GermanJoey]

Hmmm, an idea. If it is possible to detect the size of the largest landmasses on a plane before placing the players, then an algorithm like this might lead to a better distribution:

(example for Arcanus on a standard size map, numbers might have to be different for different sizes ofc)
a.) if the single largest is bigger than some very big size, say 300 tiles(?), place all players on that plane on it.
b.) else if the single largest is bigger than some medium-big size, say 225 tiles (?), place 3 players on it, then the 4th player on the next largest continent.
c.) else if the top two continents together are larger than the very big size, 300 tiles, place 2 players on each of them.
d.) else use the current distribution algorithm

The idea is that since we're only talking about placing 4 players, then there's really only 3 lopsided special-cases to handle. Maybe this could help?

[Seravy]

a.) if the single largest is bigger than some very big size, say 300 tiles(?), place all players on that plane on it.

This should never happen under normal conditions. The only way it can is if two nearby continents end up connected in the final phase where the game replaces invalid shore and land configurations with land. (no they did not make art for all 256 possible shores) However that phase is way after capitals are placed.

b.) else if the single largest is bigger than some medium-big size, say 225 tiles (?), place 3 players on it, then the 4th player on the next largest continent.

Same. Continents are limited at 200 tiles each unless the above mechanism or a poorly placed node connects them.

...also, you are under the impression that all games are played on Huge land.
On Tiny land, 150 tiles is all we get, for all continents together. There, the "Big" continent will be like, 45 tiles.

Either way this is a ton of work to achieve an imperfect solution to something that's not a real problem - yes starts are uneven but the number of wars each player gets involved in plays a several times larger role than their starting location, and ultimately the game won't be easier if one AI gets a much better position than others and takes over the world - if it is, it's a signal that the AI sucks at playing that specific realm/race/whatever and needs to be improved there. So this only really is an issue for the start of the human player...and they quit and reroll if they aren't satisfied with their starting position (No, I'm not encouraging that, but it's the truth).

With the recent improvements, AI that gets stuck entirely shouldn't happen in more than 0.1% of the cases, especially not on Myrror.

[namad]

Just imagine what it used to be like a year ago, when you picked myrran you always or almost always faced a single opponent. The pick was REALLY OP then!
I'm glad everyone is arguing about if it is OP or not now, that means it's in a pretty good place.

[Nelphine]

I know! Seravy has inspired miracles!

[GermanJoey]

...also, you are under the impression that all games are played on Huge land.
On Tiny land, 150 tiles is all we get, for all continents together. There, the "Big" continent will be like, 45 tiles.

Seravy. Please stop picking out just one or two sentences out of my posts while ignoring the rest. You do this to me nearly every time I post something that you don't like and it is very frustrating.

I said in the post that these numbers would be adjustable for different landmass sizes, and I added the question mark around the numbers because I do not know what the average size of a continent is in Caster of Magic. My experience with maps is mostly in Civ4, where our maps can be several thousand tiles in size, so I just used numbers numbers that I felt sounded about right in CoM for a fair-sized landmass map (which is all I ever play, by the way - the only setting I ever change is to set Rich resources instead of Fair). I certainly have seen landmasses in CoM that are fairly pangaea-ish, for what it's worth. If my suggestion isn't something that you think could be implemented very easily, or there's other things that seem more important to you, or you're just not interested in this aspect of the game - well, so it goes. It was just a suggestion. But please don't just dismiss me out of hand because I used bad numbers as example parameters.

As for this being a big problem, I'd say it's an extremely core problem, as demonstrated by the fact that Catwalk, Nelphine, and myself have all pointed it out within the last page! Your counter-example of wars doesn't make sense to me. I never seem to have this problem. Not getting into a lot of stupid early wars is actually fairly easy - you simply don't attack the AI if you don't want to war with them, and then give them tributes so that they won't attack you. They'll war with each other and leave you alone. When you're ready, swoop in and decapitate their capital; rinse and repeat. On the other hand, starting a game as a Myrran wizard and having a huge continent to yourself while the other two opponents are on tiny 2-cities is like stepping down the difficulty by 2 levels, and you won't even realize it until you're deep into the game.

[Catwalk]

I do agree that this is a real problem, although I have no way of knowing how much work it would be to fix it.

As a compromise, is it possible to balance starts by modifying terrain specials as the last step?

[Seravy]

I do agree that this is a real problem, although I have no way of knowing how much work it would be to fix it.

As a compromise, is it possible to balance starts by modifying terrain specials as the last step?

I guess we can (if space allows) add a step that clears ores and/or adds new ones but...wouldn't it be unfun if every game started with the same amount and types of ores?
Same problem here, hard to define what should be done but seriously, why do you think ores are that significant?
Let's say you get an amazing start with 2 gold and a silver for +20 gold a turn. That's still no better than hiring a single Noble hero. Not even changing Heroism affects that, as it's not depending on levels. Power from crystals is the same deal, except there the amount is scaled by levels on the heroes.
That leaves Iron, Coal and Mithril, Adamantium, Orihalcon, Wild Game. I don't think Iron and Coal are that significant. Not sure about Wild Game, it does give you an additional worker...but considering Sawmill is worth 4 workers it's doesn't sound all that much either.
Mithril and Adamantium on the other hand clearly makes a difference, but does so regardless of having it on the capital or somewhere else nearby instead. (except when going the military first route instead of settler first)


tl;dr version :
I have no idea what should be done as I don't even perceive the problem. So I need an actual algorithm suggested to have something to think about. Keep in mind that fun and diversity are equally important to game balance so anything drastic like "always start with exactly 1 gold ore" is out. Getting terrain specials randomly is fun.

[Catwalk]

I'm not suggesting that every game starts with the same amount and type of ore, far from it. I'm not sure where you read that in my suggestion.

I'm suggesting making an algorithm that estimates the value of the starting position, and then randomly adds some terrain specials to balance a weak starting position. I disagree strongly that starting with +20 gold is not a big deal, your economy will snowball and let you expand and develop way faster. Getting almost any hero early on is also a huge random benefit, which is why I'm playing without them at the moment.

As for an actual algorithm, I'm sure we can come up with one. It doesn't need to be perfect, it just needs to be better than not having one.

Factors:
1) Estimated value of capital terrain
2) Estimated value of starting continent
3) If the starting continent is shared, divide the value by the number of wizards on it

If the total value calculated above is below a certain level, add one or more ores to balance it out. Rough idea for valuation of ores:
Silver/iron/game/orihalcon: 1 point
Gold/coal/quork/mithril: 2 points
Gems/adamantium/crysx: 3 points