Procedurally generated games are the best

[laughingman]

The games I play the most are the ones that give me a unique experience each time I play them. I started playing Rogue and its variants in the 1980s, and still play and love them today. I got AI war and its expansions a couple of weeks ago mainly because it used procedural generation.

I started playing difficulty 7 games with random easier AI. As you can imagine, I got pounded into a grease spot, repeatedly, in many different creative and painful ways, before I started to get a sense of the strategy required to survive even a couple of hours. Each failure taught me something important. I'm still playing difficulty 7, and I'm getting smarter and more cunning each time.

According to Steam, I've played about 40 hours total. I'm still seeing things that surprise me, and still being educated on the finer points of war. I can see myself saying the same thing after 400 hours.

Thank you for developing such a rich, interesting, throughly challenging game.

[x4000]

Thanks so much for your support and kind words!  We happen to agree on the procedural-generation.

[keith.lamothe]

Yea, it's great stuff   I think you'll like AVWW when we've got it to a playable point, as it also uses procedural methods very heavily.

There's the procedural methods, and there's the methods for creating hand-crafted content (for instance, adding a new ship type, making its art, picking and implementing its mechanics, balancing its stats, etc), but the real trick seems to be in the interface between them: determining, implementing, and refining how the procedural methods take the hand-crafted elements and use them to make changes to the gamestate (including the initial "change" of setting it all up).

In any event, learning how to use these tools is very much an ongoing process for us; it's lots of fun

[chemical_art]

I love random elements in games, it helps prevent staleness. I keep playing Master of Orion years later in part due to its randomness. In MoO each race would only get a part of the total technologies and as a result in many games a key piece of technology would be removed altering gameplay. For example, sometimes you would get combat teleports but never develop anti-teleport technology. As a result everyone used teleports because the only defense was to teleport yourself!

AI war is good for being random as well. In addition to the map themselves changing each AI feels pretty distinct further enhancing variety.

[keith.lamothe]

Yea, MoO got real mileage out of that system.  Master of Magic was made tons more fun by the fact that which spells you were able to research in a given color was significantly randomized but the more books you took in a color the higher number of spells in each rarity tier you would get.  If you _really_ wanted Change Terrain and Transmute (among other awesome spells) you could just take 10 or 11 green books, or if your whole strategy hinged on getting Crusade you could do the same with White (though you couldn't take 10 books and pick Warlord, iirc) to make sure, or you could gamble on not getting one or more of them by taking fewer books but that would let you take something else, etc.

More recent games have tried randomly-incomplete-tech-tree approaches but it tends to not be well received.  I'm not sure if it's problems with the implementations or with the players

[blastpop]

More recent games have tried randomly-incomplete-tech-tree approaches but it tends to not be well received.  I'm not sure if it's problems with the implementations or with the players

I think it is the players. I am a board game publisher- and there is now a large element of gamers who do not like too much randomness (must be something societal). Maybe an occasional die roll to keep things from being totally deterministic. They feel that randomness and luck take away from skill. I respectfully disagree that viewpoint. Yes, you will eventually become better at something that has little variation. My take is you are more adaptable and can think on your feet if the game plays differently each game. In addition a game feels fresher longer if things are not linearly scripted campaign. AI war with my limited play thus far looks nearly infinitely re-playable with so many variations and combinations.

[x4000]

I think that the reason it works for AI War and not some other games is a subtle bit of perception on the part of players.  With AI War, we have a core of stuff that is ALWAYS there for all games, for all players.  That creates a sense of consistency and predictability.  But, then there's this huge amount of other stuff that is rarely there, except some of it is there randomly in every game.  This to players feels like a bonus, rather than like something is being taken away.  And, truthfully, if the players had all ships in every game, it would be overwhelming to the point of paralysis.

With the mechanic of randomly taking something away from players, you meet resistance immediately.  "Why do you take away something I want!?" etc.  There has to be a certain core feature set identified, and then other stuff which is treated as a bonus.  If players think of that extra stuff as being core, then you've already lost them I think.

That said, I think board games are a little bit different.  In terms of die rolls or cards, it's a bit of a hard sell.  But with something like Carcasonne, that gets really well recieved because you get random options but have many ways to skillfully apply it.  With a game like DungeonQuest, so much is random and so many of the random things are penalties, that there's clearly very little skill involved in that game.  It's more about the experience and shared misery and humor, at least based on my play sessions with it.

I think the answer is always different for every game, but overall players are looking for random options that increase variety between playthroughs -- not which stand in place of strategic choices.  That's why there's no random component to the actual combat of AI War, for instance.

[keith.lamothe]

Some games rely not only on symmetrical conditions for each player, but finely symmetrical conditions.  In Starcraft I if there was a 50% random chance you would start with an extra crystal deposit next to your starting position that would be a huge, huge detriment to the game for a huge part of its audience.  In AI War if there was exactly the same random chance then most people would never even notice.

Making randomization more acceptable involves making that symmetry less important, and thus means making "winning" less important so that "I lost because of a die roll, not because I made the wrong choice" hurts less.  To some extent that means moving away from "Game as Contest" towards "Game as Story" or "Game as Journey".

There are other factors to the reaction to randomization thing, of course.

Interestingly, in MoO for some reason I never felt totally nerfed by a missing tech (probably because I never dug into info about the game to _know_ I had been nerfed), but in another much more recent 4X you could run into a situation where a bad die roll meant you had absolutely zero access to point defense technology and the enemies could just shred you with missiles pretty much regardless of what you do.  That kind of situation can be improved by making the "range" of possible random results less broad in terms of how much challenge is added (i.e. having at least basic PD tech always be available, or have some alternate missile counter always be available if PD is not, etc).

[x4000]

Well, that's all also true.  But there's also "Game As A Contest Against The Machine," rather than against an opponent.  For instance, in Chess it is sometimes an exercise to set up random board positions for the endgame and see if you can figure out a way to win for each side. 

That's just an exercise, not actually a way to play, but it's a good example of pitting yourself against random scenarios and seeing how you are able to rise to the challenge.  And that works because you're not playing against another player.  I think that most RTS single-player campaigns fall under this guise, too.  Starcraft included.  It's when you get into competitive multiplayer that symmetry becomes all-consuming.

[keith.lamothe]

Right, it works there because you've not merely reduced the importance of that symmetry, you've removed even the pretense of it

AI War is fun even if you lose, AI War is fun even if you don't finish, so razor-edge balance is simply not necessary.  The only thing that needs balancing is to make sure that there aren't any legitimate choices that are always bad ones (and, in general, only very few that are always good ones).

[x4000]

Yep, this is true.  And some variant of that applies to Carcasonne, even when played competitively.

[Nalgas]

I am a board game publisher- and there is now a large element of gamers who do not like too much randomness (must be something societal). Maybe an occasional die roll to keep things from being totally deterministic. They feel that randomness and luck take away from skill.

If the random elements are balanced and introduce variety into how the game plays and force you to actively think about each situation instead of playing from an existing known "best" strategy, I'm all in favor of them.  They can add a lot of depth to a game and keep it interesting for longer.  I suspect the most resistant people to that sort of thing in games are the ones who play them more casually and just want to be able to pick them up and play them every now and then without having so much stuff to keep track of in terms of varying mechanics or rules or whatever.  On the other hand, those are usually the kind of people who still think Monopoly is a good board game.  Philistines.

[x4000]

I actually enjoy Monopoly now and then, but it's definitely not a favorite.

[Nalgas]

I tolerate it occasionally if people are willing to stick strictly to the rules.  It's completely unplayable with all the house rules people use like all the fines being paid to a stack that goes to whoever lands on Free Parking, because once you remove the money sinks from the game, it never ends.  Jebus, is it ever annoying to try to play with infinite/free money house rules (which are distressingly common), because you know right from the beginning that everyone's just going to get bored and give up after a couple hours when no one's showing any signs of losing any time soon.

[x4000]

Oh, yeah -- I never play with house rules of that sort.  We have people bleeding cash after a while, and tend to stop the game with whoever is ahead when the first person goes out.  Otherwise it indeed does never end.