[Freeciv-Dev] Re: Project goals

[saywhat@xxxxxxxxxxxx]

Per Inge Mathisen <per@xxxxxxxxxxx> wrote:
> What features define Freeciv?
  I have some thoughts on this.  I hope they are helpful.  They
are not *all* directed at you particularly.  They are intended as
a general discussion of Freeciv's future direction. 

  AIUI "good game design" involves boiling down a lot of game
elements until you're left with only the essentials.  Elements which
dilute the fun are thrown out; elements which give the player "fun"
(in the form of interesting choices) are kept in. 

  Freeciv is a game.  Therefore the principles of good game design
(not just good *computer* game design) should play a strong role in
how Freeciv is designed. 

  Single player computer games have an additional requirement -
the AI has to be a challenging opponent.  Multi-player computer
games can sometimes get by without it.  Freeciv tries to be both -
which makes the job of designing it harder. 

  I believe you have correctly identified many of the issues
relevant to making Freeciv fun to play.  I agree with some of your
conclusions but not others.  Please allow me to explain why. 

  I believe that supporting 30 players (and the large map that is
needed to support that many players) is the most problematic feature
of FreeCiv.  Why?  One equation sums it up: 

Large map + Large number of players = lots and lots of choices 

  Lots and lots of choices means that each decision is less
important.  But isn't making resource management decisions what
makes 4x games fun?  When such decisions become so numerous that
they become a chore instead of an interesting choice, then the
game has failed (as a form of entertainment). 

The abundance of choices causes other problems.  E.g.: 

1) Each city is less important. 

     Lose a city?  Then just build another.  This undermines the
 importance of cities in the game. 

2) Each unit, building, trade credit, science credit, tax credit,
 etc. becomes less important. 

     Because of the sheer volume of resources produced in a large
 Freeciv game, the ability to make good decisions (about how to
 manage resources) tends to be less important than the ability to
 make quick decisions. 

     IOW there might be an interesting game happening in front of
 you.  It's too bad that you're too busy clicking to enjoy it. :-) 

3) AI calculations are *expensive*. 

     Perhaps worst of all, the plethora of choices makes AI
 calculations much more expensive in CPU time.  This limitation
 strikes right at the heart of making the single player game
 challenging. 

  My point is that Freeciv must first be fun to play on a small
scale (i.e. small map, small # of players).  If it isn't, then
increasing the scale is not likely to help much.  Would chess be a
significantly more interesting game if it had 10 times as many
pieces per side (on a board 10 times the size)? 

  So for Freeciv, I propose limiting the default number of cities
that a player can create (via a settler) to a relatively small
number (e.g. 5) - controlled by an option, of course.  That one
change will have numerous effects: 

1) Where you place your cities becomes much more important. 

 The early game exploration phase (finding the "good" city sites
 and getting a settler there first) becomes much more important. 

2) Defending your cities becomes much more important. 

 After each player has founded 5 cities, if you lose one city to
 an opponent, he now has 6 cities while you have 4.  So he has an
 obvious production advantage. 

 You *can* build a settler and found another city (to get back to
 your maximum of 5 cities). But that still leaves you at a 6 to 5
 disadvantage (relative to the player who took your city). 

3) Alliances become essential for many players. 

 After the first few city conquests, one or more players will
 likely have a significant production advantage (due to their
 larger number of cities).  At that point the remaining players
 have almost no choice but to form alliances against the stronger
 players in order to survive. 

4) Smaller numbers of cities + smaller maps +  a smaller number of
 opponents = less micromanagement (particularly late in the game). 

 Imagine how much easier it would be to manage a few dozen cities
 instead of hundreds of cities (in the latter stages of the game). 

5) AI coding becomes less constrained. 

 Because AI calculations aren't working on such large data sets,
 more kinds of calculations are practical (within the CPU time
 available).  Perhaps that could make more complicated AI
 decisions viable? 

  If the maximum number of buildable cities *is* made an option in
Freeciv, then those players who want to play big games (like they do
now), can do so.  They will also know that, by doing so, they are
going "out of bounds".  So they should expect more micromanagement
work, a greater likelyhood that modal dialogues will popup at
inconvenient times, poorer AI performance, etc. 

  It may even turn out that some of Freeciv's features *will*
scale up well.  But if not, then at least the default settings can
provide a competitive game (for single player or multi-player
games).  That might be better than splitting Freeciv into separate
single-player and multi-player versions. 



I have some other comments and questions: 

>  I) Up to 30 players can play in multiplayer
  I'm curious about how many Freeciv games with more than 10
human players are played to completion.  I'm really asking (because
I have no idea how popular large multiplayer Freeciv games are).
Perhaps that info should help guide the future direction of Freeciv. 

  If such data is available, does it show that those who win
games with fewer players (e.g. 5 ) win proportionally the
same number of games where the number of players is a multiple of
that?   E.g. on average, if a player wins 3 out of 6 game where 5
players are involved, then does he also win 1 out of 2 30-player
games as well?  In part, I'm asking if large numbers of players
introduces "noise" into the game results data? 

  Is "30 players" where the "maximum number of players" line
should be drawn?  Game designers typically set a minimum and maximum
number of players for their games.  Presumably they set those values
based on how well the game "works" with each number of players.
Perhaps Freeciv should set its limit this way too.  Maybe "30" is
too high (for a variety of reasons). 

  And finally I have a smartass rhetorical question: 

  Is winning a 30-player game of Freeciv 6 times more satisfying
than winning a 5-player game of Freeciv?  :-) 

>  A) It is a turn-based game (TBS)
>  H) Gameplay is simultaneous
>  I) Up to 30 players can play in multiplayer
  As you noted, these issues threaten to split Freeciv's
development.  For a long time, I've wondered if Freeciv could use
techniques like those used in Dan Bunten's Global Conquest to
implement all three of the above features.  (Global Conquest was a
turn-based game that supported multiple human players with both
simultaneous turns and simultaneous movement.) 

>  G) Players start out on an empty world (except huts)
> I think the most important features are A to E while G is a bit silly and 
> rather boring, and H is inherently broken. I do not think H can be fixed 
> without losing I.
  Exploring an unknown map has been a popular game element since
the days of "Seven Cities of Gold" (and before that "Empire").  If
Freeciv's design has evolved to where it has made that part of
the game "boring" then I think that that is unfortunate.  I hope
that it can be fixed. 

  That's all.  What do you think? 

-Eddie