Interesting Observation about Civ 3 - Input wanted
Adam Vestri
I have crossposted to a.g.c2 as well.
Situation - A simple one. I'm a Greek regular caravel in coastal
water. Two other regular caravels are two squares away in coastal
water, one Iroquois, one American. I am at war with America, peace with
the Iroquois.
Immediately I think to attack the American caravel. It's not really a
strategic need, but what the hell, might as well stick it to them while
I have the chance. Caravels only have an attack and defense of 1, and
since both are regulars, that leaves me with a 50/50 chance of winning
the battle. A coin flip is good enough for me, so I attack the
American.
The arrows fly and and I lose all my HPs and sink. The American caravel
takes no damage. The odds of getting swept in an even battle isn't
really that implausible (2 ^ 3, or 1/8 chance), but I'm curious that
perhaps I missed some naval battle modifier.
I quickly scan the manual and only find that I missed that coastal water
(or any water for that matter) has a 10 point (10%) defense bonus.
American only had The Pyramids as a wonder (I have none, damn), and is
about the same in the way of technology, so the only difference should
be the terrain mod, which is minor. So I have a 45/55 chance of winning
instead of a coin flip. In a mathematically perfect world, in 20
battles I should win 9 and lose 11.
Curious, I go back to a save game right before the battle, then
reinstate the attack on the American caravel. Boom, swept again.
Hmmm. I go back again and try again, same results. Now the odds are
getting remote (1/512). So I try ten more times and get the same
results every time, I get swept and bye bye caravel. Now I know I've
missed something. Even if the 1.1 defensive value is rounded up to 2,
the odds of these results are infinitesimal.
I peruse the manual and can't find anything else to explain this
phenomenon. Yes the AI has to compensate for it's inherit lack of
intelligence and intuition in other ways, but I have faith that it
doesn't take advantage of battle calculations.
Then I come up with an idea. Why don't I attack the Iroquois caravel
and see what happens? Yes this means war, but I intend to go back to my
post-battle save anyway. I pull up next to it, click the confirmation
icon. Completely different story. The Iroquois takes damage, then I
do. Then he does again, then I do again. It's one to one. Luck
strikes and the Iroquois caravel is sunk on the final 'round'. This
sounds more like what should have happened in a 50/50 battle, so I'm
immediately relieved.
To figure out if this is just a fluke, I go back to the previous save
and reinitiated the battle with the Iroquois. The exact same results
happen (him damage, me damage, him damage, me damage, Iroquois sunk.)
Now I'm really curious, so I try the Iroquois battle ten more times.
Each time, the exact same battle progression happens and I win with one
hp left. I even try going back occasionally and attacking the American
caravel, as usual, I am swept every time.
The odds of this just being a chance even are impossible. So this
leaves me with three hypotheses that I can think of: the game records
and remembers battle victory and defeats regardless of save, battles
are some how predestined, or the AI is actually skewing the battles to
it's advantage. More research is definitely warranted, but I don't
have the time to check it out today. I'll mess around with it
tomorrow.
Has anyone else had similar experiences? Any insights?
Sky Render
<snip>
> Has anyone else had similar experiences? Any insights?
What you just described is a non-random battle damage system. Basically, the
game determines what unit will win in the next unit-to-unit battle before it
even begins. It's possible that the nationality difference of the units is a
part of the calculation, and given your results, I'm inclined to think so.
--
_-`-_-`-_-`-_-`-_-`-_-`-_-`-_-`-_-`-_-`-_-`-_-`-_
/ Sky Render (remove the corrupt universe to mail) \
{ Aka. Ouroboros, aka. Doom Guru, aka. Bob the Evil }
{ }
{ Hater of Shock the Monkey ads and their ilk, }
{ reviver of dead conversations, conquerer of }
{ worlds, and an all-around nice guy. }
{ }
{ AIM: SkyRendrX ICQ #:2632972 }
{ Home Page: doomguru.8m.com - Doom Guru's Domain }
{ }
{ "Tragedy is when I cut my finger. Comedy is when }
{ I walk into an open sewer and die." - Mel Brooks }
{ }
{ THIS SPACE FOR RENT }
{ (Well, not really.) }
\_-`-_-`-_-`-_-`-_-`-_-`-_-`-_-`-_-`-_-`-_-`-_-`-_/
-_- Beware, the Eyes are after you... -_-
Ź_Ź Ź_Ź Ź_Ź -We're agents of the government. Surrender. O_O -Uh, oh...
JTung
"Adam Vestri" <ave...@pacbell.net> wrote in message
-snip-
> since both are regulars, that leaves me with a 50/50 chance of winning
> Curious, I go back to a save game right before the battle, then
What you have just experienced was a way to prevent people from winning
every single encounter by loading the game if they don't win first time. I
suspect the game does this by saving the next random numbers for events each
time the game is saved.
You can demonstrate this by having two (or more) units each standing by a
separate goodie hut. Save the game. Unit 1 might release some barbarians,
while Unit 2 may find some gold. Load the game, and reverse the order of
movement. Unit 2 will find the barbarians while Unit 1 gets the cash.
This was first implemented (at least in any Sid game I know) in
Colonization. There are several ways to beat this, in which I shall not go
into detail, but basically you have to make the game run a bunch of
'variable outcome events' before your lost battle.
DennyK
the exact same outcomes in battles, even through several turns, when
reverting to saved games and repeating my actions over and over again.
(Learned this while messing with the crashing issue I have... :-/ ). If
I do some stuff, save the game, take five or six turns or so doing more
stuff, and then fight a battle, the results will be exactly the same,
blow-for-blow, every time I reload the game and do the exact same stuff
since the save. Even multiple battles fought from turn to turn turn out
exactly the same each time if my steps are repeated. So, whatever the
cause of this phenomenon, it can be noticed across several turns as
well, not just the next turn after a save, provided the same actions are
taken each time before the battle. Go figure... ;)
DennyK
loco
have to go back a round or two to get an alternative result.......
Kevin D. Foster
[snip]
>
> Has anyone else had similar experiences? Any insights?
This has been mentioned before with spies as well. Doing some more
testing, I've found that you can change the results by completely
exiting the game and starting it back up again.
This is almost certainly the result of a simple anti-cheat method,
where when you load a game, the program remembers the last seed it used
to generate a random number. So after the game is loaded, and you try
the same event, the same random numbers are generated, giving the same
results. It's a reasonable, but simple way to try and stop save-load
cheaters.
PS - For the person who mentioned Colonization, IIRC, it's anti-cheat
method was more strict. When you reloaded a game in Colonization, if
the first random action you took was in combat, you lost.
--
-=[ Keeper ]=- ICQ# 8105495
kee...@lycosmail.com kdfo...@home.com
http://members.home.com/keepershaven/
Marko
random seeds for the random number generator and i think they are set one
turn at a time. Did you try going back two turns and see if the battle
results are the same?
Greetz, Marko
zeCod
and if you lose
and reload, you might win the next time..
Loren Pechtel
wrote:
>The odds of this just being a chance even are impossible. So this
>leaves me with three hypotheses that I can think of: the game records
>and remembers battle victory and defeats regardless of save, battles
>are some how predestined, or the AI is actually skewing the battles to
>it's advantage. More research is definitely warranted, but I don't
>have the time to check it out today. I'll mess around with it
>tomorrow.
>
>Has anyone else had similar experiences? Any insights?
Your diagnosis is half right. What's no doubt going on is that
the random seed is being saved with the game. If you play things the
same way you get the same result. However, by altering events you can
get a different outcome.
jizenka
> Has anyone else had similar experiences? Any insights?
Kinda. I've been attacked on several occasions by a massive barbarian
onslaught - on one occasion involving 23 horsemen at once on one city (built
on grassland) defended by one elite and one veretern spearman (roughly at
the beginning of AD) - both fortified.
Albeit on the "chieftain" level to get used to the changes since Civ II (and
checking for bugs - so far ... none); 23 horsement defeated by two spearmen
without loss - just a few hit points damaged! The only thing I can think
that was is the additional defence bonus for being attacked cross-river.
Unless there *is* a prob with the battle calculator.
Damn! Now if only I'd bet money on the spearmen!!!!!
Ian
Kevin 'Keeper' Foster
There is a defensive bonus when you are attacked across a river.
There is also a minimum 10% terrain defensive bonus.
the_narrow_way
individual unit vs. individual unit that is stored. If you use another unit
of the same make as the first one (ex. Infantry_A gets crushed after
attacking a fortified cavalry, so you reload and bring a different unit,
Infantry_B, to attack, it will lose the fight in the identical manner).
Loren Pechtel <lorenp...@hotmail.com> wrote in message
news:3c06c32a...@nntp.vegasnet.net...
The Enigmatic One
>
>
>jizenka wrote:
>> Albeit on the "chieftain" level to get used to the changes since Civ II (and
>> checking for bugs - so far ... none); 23 horsement defeated by two spearmen
>> without loss - just a few hit points damaged! The only thing I can think
>> that was is the additional defence bonus for being attacked cross-river.
>> Unless there *is* a prob with the battle calculator.
>
>
> There is a defensive bonus when you are attacked across a river.
>There is also a minimum 10% terrain defensive bonus.
There is also a huge bonus against barbarians.
-Tim
mitc...@excite.com
other game. However, I've also noticed that if you've got someone
next to a barbarian camp (without a barbarian sitting on it) and the
first time you go into it you've disturbed someone, just sit and wait
a couple of moves, if you're not attacked, and suddenly they'll give
you something, even if it's only a map.
On Wed, 28 Nov 2001 22:07:50 -0800, Adam Vestri <ave...@pacbell.net>
wrote:
>I know that many people still might not have access to the Civ 3 NG, so
CleV
wrote:
>The odds of this just being a chance even are impossible. So this
>leaves me with three hypotheses that I can think of: the game records
>and remembers battle victory and defeats regardless of save, battles
>are some how predestined, or the AI is actually skewing the battles to
>it's advantage. More research is definitely warranted, but I don't
>have the time to check it out today. I'll mess around with it
>tomorrow.
>Has anyone else had similar experiences? Any insights?
I've done a few game saves-reloads and every outcome is always the
same. leads me to the conclusion there are deep calculations going on
rather than randomisation.
Loren Pechtel
>I've done a few game saves-reloads and every outcome is always the
>same. leads me to the conclusion there are deep calculations going on
>rather than randomisation.
All this requires is saving the seed for the random number
generator.
Note that computers generally do *NOT* generate truly random
numbers. There's no way to do so without some sort of external input.
(Note that programs that really need a random number do it this way.
Create a PGP key pair and it asks you to type. The timing on your
keystrokes is a truly random input.)
Instead, they use a mathematical formula that yields a sequence
that is random enough for most purposes. These normally function by
transformation of some seed number and extracting part of it to get
the next random result to return. If you use the same seed you get
the same sequence of random numbers. Thus a simple way to impair
save/reload is to save that seed as part of the save file.
Chris Robinson
"CleV" <CL...@balJUNKcab.ch> wrote in message
news:3c08e68f...@news.balcab.ch...
I it's intentional to stop save scumming.
Have you tried going back a number of turns or using units in a diiferent
order?
Ken Bass
determined by a random number generator. If re-loading a game restores the
'seed' to the random number generator algorithm, the results of the 'dice
roll' will always be the same. Perhaps CIV2 doesnt restore the seed.
Chris Williams
"JTung" <jtung@DELETE_UPPERCASEoptusnet.com.au> wrote in message
news:9u4l9g$8ss$1...@perki.connect.com.au...
This doesn't explain why he beat the Iroquois, but not the Americans....
Chris
Paul Nilsson
seems like one for every civ at least, mabye multiple generators per
civ for different purposes e.g. battle/goodie hut/espionage.
Of course "random" number generators aren't really random they are
psuedo-random and use seeds to get the next random number, it simply
saves the seeds in the save file along with everything else. The
upshot seems to be that you will observe identical behaviour every
time you reload. Perhaps this is to stop people cheating by saving and
reloading when they lose.
cheers, paul
Powerslave
You see a goody hut, save the game and go in. Ooops, barbarians. So I
re-load the game (this is still CivI/II) go in and I get a tech or gold (I
always re-load after barbarians and mercenary (if I have to pay for it)). I
tried this in CivIII and after five tries (I'm not as patient to try it ten
freakin' times), still get barbarians. I think the AI does what you said
(remembers regardless of saves, etc) OR it determies the outcome at some
slightly earlier point (okay, never mind, makes no sense, but then, neither
does what you described, WEIRD)news:3C05D0B6...@pacbell.net...
Louie Landale
reloading will repeat the results. As for the huts, just move into them
next turn.
- Louie
"Powerslave" <som...@microsoft.com> wrote in message
news:13wR7.239251$HA6.43...@typhoon.southeast.rr.com...
Trevor Hill
I have heard that the random number generator for Civ 3 is saved when you
save your games, so that when you reload, the same pseudorandom numbers come
up. Try doing something else that requires random numbers (attack someone
else first) and THEN attack the Caravel, and you will get different results.
--Trevor