What I would love to lurk..

[Krill]

Gaspar hates you for reminding him that PB16 happened.

[spacetyrantxenu]

Too bad, he'll have to get over it. We're going to need almost everyone, including retirees (that includes me I guess?) to fill up a 50 player game.

[Krill]

Would still need minimum of 2, preferably 3, subs to players that dropped out.

[Old Harry]

Do you think you could manage a turn every three days Krill?

Also be interesting to see if a server can handle it and just how long login can take before people get bored and don't bother.

[Krill]

Meh, I figure any game with 25 or more players is essentially the same thing. It's going to reach the modern era unless the game isn't totally unbalanced.

[spacetyrantxenu]

I'm sure a server could handle it, the question would be more of what the game engine can support I think. Cal's server is running PB18 no problem, along with several other games at the same time. The final constraint to game size would probably end up being the 4 GB RAM limit for a 32-bit process but even PB18 doesn't come anywhere near that much RAM use in the pitboss server instance. Even with the added overhead of running graphics the game only uses up about 1.5 GB RAM on my system, last time I remember anyway. I think a 50 player game would be manageable, and logging into PB18 didn't really start taking forever until the game saves got to be really big (they're over 2.5 MB now!). The constraint will be in getting enough players for a game that big.

[Fintourist]

Been thinking about how to make a map for 50 players, and how to even roll the civs for that many people. There would have to be duplicates due to only being 34 civs in teh game. Only method I've come up with is splitting the player base into 2 or 3 groups and then either running the random pairings with rejection pick method, or doing 3 separate snake picks.

I suppose the reason I'm posting is: lot's of people seem to hate change, or duplicates, but does that sounds unreasonable? Any problems with those pick methods?

How about this:
1. Make a list consisting of all restricted pairings
2. If there are less players than possible leaders, start removing duplicate civs randomly until there are as many civs left in the list as there are players
3. Now if you have e.g. 40 players you can roll random pairings using all the leaders, 28 unique civs and 6 duplicate civs
4. Then either the pairing is done or people get to reroll like in PB18 or some other creative/fun next step is taken before the pairings are final

I don't strongly object the methods you suggested, but this way the amount of duplicate civs is minimized and everyone gets a unique leader (unless we have more than 52 players) :P

[Krill]

Problem: if there is only 1 person whom re-rolls, how does that work?

[Cheater Hater]

Crazy idea: since we're so sure this game will go to the modern era, maybe some of the "duplicate" civs could be the original (or tweaked) versions of civs where we moved the UU/UB forward? That way there's still more variety rather than just duplicate civs. Obviously that doesn't work after the first couple (America, Russia, Germany, were there any others RtR moved forward?) but it's something.

[Fintourist]

Problem: if there is only 1 person whom re-rolls, how does that work?

He gets the same civ back, but the leader is randomly assigned from the available ones. If we have exactly 52 players (all leaders taken) then he just gets the same combo back.