Ohhh I just had a good idea for this:

LOSE the game? Option to extend the map and continue!

Think about it for a second. I, the great Evil One am vanquished by an allaince of tree-hugging, kitten-cuddling goody-goody two shoes fluffy-bunny twits.

I want revenge!

So why not extend the world, and get revenge. What could be MORE satisfying to the freshly-vanquished player than an entire game of vengeance against the infuriating cabal of computer players that killed you off?

In fact, why not have a "extend for revenge" option if you're defeated. Perhaps double the map and give the vanquished player a starting city smack bang in the new empty area.

Return with a horde and utterly DESTROY your enemies!

I think this would be extra cool because it actually makes losing a story-building event rather than an incredibly frustrating game-termination.

Thoughts?

I don't like the original idea but i could accept it. This last one... no. If you have been defeated, you have been defeated. If it's an option for only the player, it's unfair because it means that you cannot actually lose (in theory) while the AI can be defeated. If it applies also to the AI, then... well, endless game?

Also, if you only start with a city and some troops, how do you expect to defeat that whole alliance that destroyed you? If you start with some cities and a good army... Where did they came from? If all your empire was at the other point of the map (and is still there altough not controlled by you anymore), where did all that stuff come from? I'm going to suppose, that you "killed" channeler found refuge in a far away land that was already developed and that had no knowledge of channelers, receiving with open arms such a powerful being (altough your channeler should start weakened to reflect all the previous mess). In that case, i could like the idea somehow.

In any case, an idea like this i would prefer it to be "one shot". Where is the fun to win if you needed to use this option four times? (with the consequent four expansions in the map)

Well when you extended the map after winning why wouldn't some of the new AI players be returning vanquished foes?

As for resources, if you're on another island or far enough away you could have time to build up, or yes indeed you could start as having taken over an existing empire. Villains do this stuff all the time... if we continue the LoTR comparisons the Witch King and Sauron both came back with new empires after being vanquished.

As to where's the fun in winning after 3 defeats? That would have been a pretty challenging and epic game, yes? If there's no appeal in that for you - cool, I bet some people would love this and others hate it, that's why it's proposed as an option

didn't supreme commander's campaign system have an expanding map?

the system worked really well in the campaign map, and i honestly would have loved a mode like it in skirmish

the concept to me of generating a huge/gigantic map, and then playing a small section before expanding sounds fantastic fun, would also be nice as it could dynamically scale the opponents per expand, keeping the challenge flowing.

I started to think... I wonder how hard it would be to randomly generate an established civilization of a certain strength, as if it had been playing since the start of the game, expanding, fighting, researching, deploying troops, etc. Having the map expand upon victory seems like it would be an easy thing to do. But having the civs in the expanded world being generated well... It might require more time and effort than the feature is worth (and I really like this idea).

My guess is that the design challenge might be more on the "how long to make the player wait" problem than on alogrithms to generate factions that still reflect the same game rules a player must follow. If the game doesn't have to do any significant live graphics, a modern PC can do a pretty ludicrous number of calculations. But if that still means several minutes (or more) of waiting for your expanded map, you might be losing steadily more of the potential audience for the feature as your next-map interval grows (large-map players dropping out first, maybe everyone eventually?).

The only way I can think to do it is to have the AI's run, turn by turn, essentially playing the game without a human. And based on how long AI turns tend to take passed the beginning... It could take a very long time. Dunno about you but I wouldn't want to wait around for 5 AIs to play through 500 turns in a row . Even if it only took an average of 10 seconds per turn, that would still be over an hour!

I'm sure you could code some random city and resource generation fairly quickly, a big short-cut on playing it out turn by turn.

Even if this generated content wasn't spot on, surely it would only take an AI player a few turns to take the raw, generated empire and turn it into something formiddable.

Seriously though, when le frog codes his AI he's going to have tables of what buildings and units they prefer and in what order - just run those tables on a dozen cities and you're done.

Yeah, but what about calculating what long-term magic might have caused to happen? What about the different POIs on the map - would the civilizations explored some already and had a chance to get whatever rewards might be found in them? Just generating a bunch of cookie cutter cities would make for a pretty stale experience. I'm hoping that the AI is smart enough to build different things based on its needs, not just run down a set priority list. The games would start to feel the same after a while, otherwise. So to generate a random empire in a way that doesn't leave the player with a bad taste in their mouth wouldn't quite be that easy. Expanding the map to encounter a handful of civs the same size as you are, but feel as if they've just started playing the game, would be awkward. To implement this feature well they'd need to be able to generate civilizations that feel as if they've been played through until that point in time. 

To get around any wait time required after victory, here is my suggestion: when the game detects the player owns over 80% of the cities or whatever on the map, it assumes victory is coming soon, and so uses any spare cycles to start calculating out the generation of the new empire. This game is turn based remember, and so when the player is doing things on his turn, almost no calculations are being done. All the calculations for what is done that turn is calculated in between turns. If the game utilized these periods of relatively little calculations to start work generating the new civlization(s) by the time the player won the game, the game would have the data ready for the new civs to pop up pretty much instantaneously.

Also, a wait time of 10 turns or so between victory and when it allowed the player to interact with the new civilization(s) would be fairly reasonable. And by interact, I mean it would not let the player move his/her armies to sit right next to the new AI's capitol for 1 turn victory, but more of a time to build up one's forces on the border if war is to be declared, or a time to make peace treaties with the AI.

One of the main reasons for a system like this is after a 20 hour game, I am really starting to like my civilization I built up, whether it be in Elemental, Sins or any other strategy game. If given the choice between continuing on with my empire against more powerful adversaries and starting all over again from scratch, the choice to me is obvious.

I think we'll have to agree to disagree there. While I share your concern, I'm just not seeing the need to turn-by-turn calculate what the AI would do. If the player isn't there, it doesn't need to be scrupulously fair - so a few shortcuts will do.

As to the stale experience, real turn-by-turn AI would take over as soon as the map is done expanding - and unless you've got some spell which reveals the whole map you aren't going to see anything "stale" or "cookie-cutter". You wont see anything but blacked-out areas to explore, and while you're getting there the full-on AI will do whatever it does with points of interest and developing the map.

MoM generated plenty of cities at the start of the game - handing something a little more complex than that over to turn-by-turn AI would surely be enough. Really only one way to find out though... an alpha test

cheers,

Kul

I'm not seeing that there would need to be a turn-by-turn calculation of what the AI would do. It would be too impractical. I think what would need to be done is to have a civilization generator that can generate civs intelligently based on surroundings and maybe some other factors. 

I disagree. Hopefully there will be a lot of strategy involved in city development. It seems to me that making a civ generator capable of creating completely different, yet functional civilizations would be difficult. Even though the player likely won't interact heavily with the AI immediately, if every time you expand the map the game creates similar civs to fill the world, the player will still notice the similarities even long after the AI has taken over.

And, your point about the full-on AI doing whatever it does with points of interest and developing the map while you're still getting there would just put the new civs at disadvantages. It would put the AIs hundreds of turns behind the player in that aspect.

Generating individual cities would be relatively easy. Generating a civilization consisting of multiple cities that make strategic sense, plus relevant standing armies, etc, seems like it would be a whole new level of complexity.  

I still really like this idea, but if the civ generator isn't sophisticated enough to create a different experience each time then I don't think I'd find myself using it very often. On the flip side if it's done extremely well I would probably almost always start games on smaller maps and then choose to expand it on victory.

I don't think I'd ever use this feature, but it seems to have a lot of support here, so what they hey... as long as it doesn't take too much time from developing other features, more options for the player is always a good thing.

To throw in my 2 cents... I would much rather start playing on the full sized world map right away.  I don't like the idea of "unlocking" portions of the map by "winning" certain discrete stages of the game.  Say that I have a small kingdom on the coast of an ocean, surrounded by larger, more powerful enemies.  If there's another continent on the other side of the ocean, I don't want to have to defeat my local enemies before I can unlock it.  Instead, why can't I build up a fleet right away, sail across the ocean and colonize?  That way I might build up enough strength to take on my enemies back home.  Did Portugal defeat Spain before setting up colonies in South America?

Similarly, if there's an empire forming on the other side of the world, I don't want it to wait for me to be ready before it attacks.  If I have a small kingdom, fighting with other neighboring kingdoms, it would be interesting to have to put aside my old differences with my former enemies in order to create some sort of hasty alliance to fend off the invading empire. 

Basically, I want the world to develop in a somewhat organic and believable fashion as the game progresses - something that I don't think can be achieved with this expanding map.  It's certainly not going to be achieved by having the game generate 3-5 random kingdoms at roughly my strength every time that I'm ready to "level up."