les-mockingjay / sinssoa2sacrificeofangels

Automatically exported from code.google.com/p/sinssoa2sacrificeofangels
0 stars 0 forks source link

Cap Ship Research didn't go as planned #171

Closed GoogleCodeExporter closed 8 years ago

GoogleCodeExporter commented 8 years ago
I see that JTaylor added the revised cap ship research such that all cap
ships except the colonizer need researched, but it doesn't seem to be
working right.  I have the following theory on why, which I intend to put
to the test over the next day or two (I'm providing this untested theory
for info in case anyone is getting antsy and wants to beat me to testing it):

1.  The AI seems to randomly select a Cap Ship, then try to build it.  If
it's got no pre-requisites (like in vanilla sins), it just builds.  If
there are pre-requisites, it tries to do those by building stations,
performing research, and then building the ship.  

2.  HOWEVER, it seems that the AI wises up if it realizes that the cap ship
it wants to build requires more research stations than it can build at the
capital planet.  Unfortunately, it only realizes this AFTER it starts
building the last station that will fit in the grav well.  It then switches
to a ship it can build (i.e., the colony ship w/ no pre-requisites).  

I base this theory on the following:

A. When playing the revised version, some of the AI picked the colonizer
right away and played fine.

B. When they didn't build a cap ship right away, they would do one of two
things:
  B.1 Build a couple research stations and build a cap ship only requiring
a few stations (e.g., defiant, which requires 3 war stations).
  B.2 Build a couple research stations, then build the colonizer
(presumably wanted, say, the Sovereign or Galaxy, but couldn't build enough
stations and picked the one it could build ASAP).

C. SOGE never has this problem, and I think it's because low-level cap
ships (i.e., Star Destroyers) don't have research pre-requisites, but
high-level cap ships (i.e., the Executor Star Dreadnaught/Super-Star
Destroyer) require, like, 7-8 research stations, which clearly can't fit in
one grav well.

If I'm right, then the only solutions to allow the AI to play the mod well
are to (1) have no pre-requisite research for any cap ships like in the
earlier beta version, or (2) make all cap ships except the colonizer
require more stations to build than can fit in the capital planet's grav
well (i.e., 5 or more stations for a home planet w/ 24 logistics slots). 
If that's not acceptable, maybe the mod team could consider an alternate
download version w/o pre-requisites so that people who don't play other
people can still have a good gameplay experience with the mod.

Original issue reported on code.google.com by thomas.s...@gmail.com on 20 Jan 2010 at 1:21

GoogleCodeExporter commented 8 years ago
good point

Original comment by jtaylor...@gmail.com on 20 Jan 2010 at 6:45

GoogleCodeExporter commented 8 years ago
Okay, scratch some of that.  It looks like making all ships but one require more
stations than can fit in the home planet's grav well doesn't really accelerate 
the
AI's decision to build its first ship.  It still needs to build up and fill 
that grav
well with stations.

However, I found out how SOGE pulled it off.  You start out with a lot of money,
enough to build several stations, but the cap ships are really expensive (will 
take a
minute or two to gather the cash to build).  That means that the AI and the 
player,
both behaving sensibly, will buy a few things and then hoard cash for the first 
cap
ship (even if the AI decides to go for a supercapship, which it can't build, it
builds its stations with the readily available cash, figures things out, then 
builds
a sensible cap ship about as fast as a person).  This results in similar build 
times
for the AI and savvy players.

However, I don't get the feeling that's what people had in mind for this mod 
since
that would mean jacking up the cost of cap ships a lot.

The problem with the AI is alleviated by giving players enough cash that the AI 
can
figure things out faster (build soon rather than wait for the cash it needs to 
build
all the structures [which it must do before it figures things out]), but that's 
the
best mitigating option there is.  At any rate, the AI doesn't start out much 
behind
players (typically 60-120 seconds build behind human player), and it seems from 
my
testing that the AI catches up soon b/c it doesn't need to build any research
stations for a while after that.

Bottom line: let's keep testing it with the (newly) revised cap ship research 
and see
if the AI gets all hosed up, but I think the AI will play competitively enough
(unless your name is STARTREKKER).  The AI will certainly catch up faster with 
more
money to start out with, which is a two-edged sword: it's a huge boon for good
players but a huge problem for inexperienced players who blow it.

Original comment by thomas.s...@gmail.com on 21 Jan 2010 at 2:11

GoogleCodeExporter commented 8 years ago

Original comment by jtaylor...@gmail.com on 22 Jan 2010 at 8:11