New Windows Installer and Source Tarball

New Windows Installer and Source Tarball

[Eric McDonald]

Hello Xconquerors,

   I have released a new Windows installer and source tarball for Xconq 
at the usual spot:
   http://xconq_hacker.home.comcast.net

   This is newer than the 09/26 release from a few days ago. The 
rationale for releasing new packages so soon again was simply that I was 
feeling guilty about breaking the Opal game, and degrading the 
performance somewhat with all games that utilized advances.

   Addressed in this release:
(1) Opal now works again. You will get a warning about a cycle being 
detected in the advances graph. You can choose to ignore it and continue 
playing the game.
(2) All games that use advances should experience no delays on startup 
now. The cycle-checking code has been improved. Also a piece of the 
research goal choice code has been disabled because it was hogging too 
much CPU. This should not adversely impact the choice of goals too much.

   New features (for game designers):
(1) Updated documentation, including on some advances-related things 
that were never documented or were documented incorrectly.
(2) New GDL global variable: 'ai-initial-research-goals' is a weighted 
list of research goals that the AI will evaluate if it currently does 
not have a research goal. The "initial" is perhaps a bit of a misnomer 
since this list can be evaluated at any point in the game where the 
above condition is true.
(3) Examples of using 'ai-next-goal' and 'ai-initial-research-goals' in 
'advances.g'. (I did not make an example of using 
'advance-precludes-advance', since I assume Elijah will probably be 
providing it in his next release of Opal....

   I also reduced the ACP of Elephants in the Advances game in an 
attempt to make that unit type less deity-like and thereby make other 
units such as Chariots more attractive.

   Enjoy,
     Eric

P.S. Since I am not currently checking into CVS, I realize that using my 
updates may be inconvenient to some. If there is an interest, I will 
provide diffs between releases to help ease the pain.

Re: New Windows Installer and Source Tarball

[Jim Kingdon]

> I also reduced the ACP of Elephants in the Advances game in an attempt
> to make that unit type less deity-like and thereby make other units
> such as Chariots more attractive.

Something needed to be done, that's for sure.  Elephants were very
powerful and dirt-cheap to build (in terms of what advances you need
to get that far).  I don't know if ACP alone will really fix the
problem.  It might also be necessarily to play with the hit-chance
(for example, the help hints that elephants are particularly good
against cavalry.  Well, maybe they aren't so hot against other kinds
of units).

Re: New Windows Installer and Source Tarball

[Eric McDonald]

Jim Kingdon wrote:
>>I also reduced the ACP of Elephants in the Advances game in an attempt
>>to make that unit type less deity-like and thereby make other units
>>such as Chariots more attractive.
> 
> Something needed to be done, that's for sure.  Elephants were very
> powerful and dirt-cheap to build (in terms of what advances you need
> to get that far).  

Agreed. I also think that it is mentioned somewhere that Elephants are 
considered to be an archery platform. If so, then perhaps Joinery (the 
advance needed for Archers) should be required as well. Also, some sort 
of construction advance should be required, since I seem to recall 
images of elephants with little towers on their backs. Additionally, 
there is the animal training/guidance aspect; perhaps some animal 
guidance skills (like a caravan) should be required. They certainly need 
to be made more expensive. The ACP reduction was just the first screw 
being tightened....

>It might also be necessarily to play with the hit-chance
> (for example, the help hints that elephants are particularly good
> against cavalry.  Well, maybe they aren't so hot against other kinds
> of units).

That is another possibility.

Eric

P.S. Those looking for a little bit of fun with the game could try 
something like the following:
   sdlconq -g advances -e,mplayer+3 10

Re: New Windows Installer and Source Tarball

[Jim Kingdon]

> I also think that it is mentioned somewhere that Elephants are
> considered to be an archery platform. If so, then perhaps Joinery (the
> advance needed for Archers) should be required

Makes sense.  I don't remember in what order I was in the habit of
doing Joinery, but perhaps not at all (it appears to be a dead-end
advance, since Horsemen are just as easy to get as Archers, and
generally as good or better at combat).

> some sort of construction advance should be required, since I seem to
> recall images of elephants with little towers on their backs.

Joinery already depends on artisanry and carpentry.  Barring a new set
of construction advances, I don't really see anything else which fits
the bill (it needs to be 4th generation, at least the way the rp
values are set now, because 3rd or lower will be easy to get by the
time one is thinking of going for Elephant, which is 5th generation).

> Additionally, there is the animal training/guidance aspect; perhaps
> some animal guidance skills (like a caravan) should be required.

That would appear to be already covered; elephant depends on a whole
line of animal advances (horse, donkey, camel, etc).

Another thing to tweak is construction points.  Right now the game
board quickly gets overrun with units, and it is really tedious to
move them all.  There are some commented out values for construction
points in advances.g which I thought were an improvement, but Hans
didn't like the slow start to the game.  The slow start could be
speeded up by making a few early units cheaper (say, slingers and
spearmen - I'd probably leave colonizers kind of expensive as they are
the source of the exponential growth).

We shouldn't go too far overboard on elephants; since Elephant is a
dead-end advance, it does make sense for the units to be kind of
powerful.  Making the player choose between Elephant now (for
short-term gains) or setting themself up for Phalanx some turns from
now is an interesting trade-off situation.  I suppose if we really
wanted to take this to an extreme, we wouldn't seek to have Elephant
depend on existing advances like artisanry, but instead to have its
own set, which don't mix with the advances which take you towards
Phalanx.

Tweaking advances.g (was Re: New Windows Installer and Source Tarball)

[Eric McDonald]

On Tue, 28 Sep 2004, Jim Kingdon wrote:

> Makes sense.  I don't remember in what order I was in the habit of
> doing Joinery, but perhaps not at all (it appears to be a dead-end
> advance, since Horsemen are just as easy to get as Archers, and
> generally as good or better at combat).

I don't think much of Archers either. But anything that can 
reasonably increase the number of deps required to get to 
Elephants is probably a good thing, even if some of the units that 
are enabled along the way are not effective.

> Joinery already depends on artisanry and carpentry.  Barring a new set
> of construction advances, I don't really see anything else which fits
> the bill (it needs to be 4th generation, at least the way the rp
> values are set now, because 3rd or lower will be easy to get by the
> time one is thinking of going for Elephant, which is 5th generation).

Okay, thanks for looking.

> > Additionally, there is the animal training/guidance aspect; perhaps
> > some animal guidance skills (like a caravan) should be required.
> 
> That would appear to be already covered; elephant depends on a whole
> line of animal advances (horse, donkey, camel, etc).

Yeah, I realized that was probably what those advances represented 
after I wrote the above.

> Another thing to tweak is construction points.  Right now the game
> board quickly gets overrun with units, and it is really tedious to
> move them all.  

Agreed. However, implementing the ability to handle multiple unit 
selections and to more handily set formations might largely 
mitigate this tedium.

>There are some commented out values for construction
> points in advances.g which I thought were an improvement, but Hans
> didn't like the slow start to the game.  The slow start could be
> speeded up by making a few early units cheaper (say, slingers and
> spearmen - I'd probably leave colonizers kind of expensive as they are
> the source of the exponential growth).

Currently the generations are pegged to a 2^n curve, IIRC. Perhaps 
we need to make the curve more of a slanted s-shape so that the 
last few generations don't take forever to research. I say this, 
because there is a limit on city densities (due to their 
increasing size and reach into the terrain), and so one cannot 
continue to settle new cities at a pace that keeps up with the 
growth of research costs. This is especially true if two sides 
are the last ones remaining and in a deadlock situation.

> We shouldn't go too far overboard on elephants; since Elephant is a
> dead-end advance, it does make sense for the units to be kind of
> powerful.  

Agreed. My intention is not to tune the Elephants into oblivion. 
They are, and by all rights should be, a powerful unit.

>Making the player choose between Elephant now (for
> short-term gains) or setting themself up for Phalanx some turns from
> now is an interesting trade-off situation.  

I have found that, in practice, I can wipe the board with 
Elephants before I can even finish researching Phalanxes (let 
alone Legiones).

It would be nice to make Elephants (immediate gratification) 
versus Phalanxes (serious firepower, but later) more of a 
tradeoff. 

>I suppose if we really
> wanted to take this to an extreme, we wouldn't seek to have Elephant
> depend on existing advances like artisanry, but instead to have its
> own set, which don't mix with the advances which take you towards
> Phalanx.

That's a thought. If you have a specific proposal, I am all 
ears.... (Or just check it in, and if I like it, I will adopt it 
for my branch of the sources.)

Eric