Multiplayer Roguelikes and PvP Combat

Archived from groups: rec.games.roguelike.development (More info?)

On Thu, 30 Jun 2005 11:17:00 -0400, Shedletsky wrote:

>I've never played a multiplayer roguelike (because I feel that the idea of a
>multiplayer turnbased roguelike is fundamentally flawed without some major
>concessions to time management), and I would never want to write one.
>However, I was thinking more about roguelike AI and my game. Specifically, I
>was thinking that in my game, ideally, an AI-controlled adventurer of the
>PC's level would be a fair match for the PC and that this would be a good
>measure of how "intelligent" my AI was. But then I got to thinking that
>there could be such a thing as an AI player that was "too good". I mean it
>would be no fun if your evil doppleganger successfully killed you 98% of the
>time, right? It seems to me that in most roguelikes that there are very few
>dungeon denizens who are anywhere close in terms of being as powerful as the
>PC (this is by design). The PC is very powerful because he can carry a huge
>inventory of deadly items, have hundreds of different potions, ect ect. So
>does it work from a gameplay perspective to have two powerful PCs clash? (or
>in my case a PC and an unfriendly copy of the PC controlled by the
>computer). It seems like in multiplayer roguelikes that allow PvP that there
>would be only a very small handful of high level characters because all the
>rest slaughtered each other.
>
>It occurs to me that any nonunique monster in a roguelike with more than a
>~2% chance of killing the PC in any encounter is probably too powerful to
>include in a game.

So, you want to coddle the players? "Gee, I'm sorry, monsters shouldn't
be able to kill you unless they're very lucky" ? B-O-R-I-N-G. What's the
point of playing if there isn't *some* challenge? If the monsters are
too easy then it's just a mindless hack-n-slash.

IMO, sure, don't make the monsters impossibly hard. OTOH, the game that
got me started - DND, not really a roguelike - has balrogs (toughest
creatures in the game except for high-level dragons) on level 1 of the
dungeon. If you're not smart enough to run away, you get something like
this:
You have encountered a level 1 Balrog.
Do you wish to (C)ast, (F)ight, or (E)vade? F
You missed.

You have encountered a level 1 Balrog.
It did 20 hit points to you.
You died.

--
auric underscore underscore at hotmail dot com
*****
Weekends were made for programming.

Archived from groups: rec.games.roguelike.development (More info?)

"Shedletsky" <mylastname@stanford.edu> writes:
> It occurs to me that any nonunique monster in a roguelike with more than a
> ~2% chance of killing the PC in any encounter is probably too powerful to
> include in a game.

I'm of the opinion that (outside the early game) a roguelike [1] should
only kill a PC for a few reasons.
1) Inexperience
Monsters that haven't been seen before doing something very nasty, for
example, or thinking "I'll just eat this gelatinous cube corpse".

2) Overconfidence
"I can take on the Temple of Great Evil without going back to my stash
for the amulet of resist evil, can't I?"

3) Lack of imagination
Forgetting that you *have* a wand of cold when taking on a red
dragon and meleeing it instead.

ADoM, I think, is very much like this. After a certain (fairly early)
point in that game, my characters either win or die due to a
combination of 2 and 3. Usually they die and when looking through the
equipment it's *almost always* obvious what I should have done.

So, I think due to the equipment issues, the chance of any monster
killing a *prepared and sensible* PC _will_ be 0%. The chance of any
monster killing an unprepared or silly PC can be *much* higher. Even
100% in some cases, though that can lead to 'hunt-the-resistance'
games.

[1] Excluding small games where the loss of a character in the late
game isn't so big a deal because it only took 30 minutes to get that
far.

--
Chris

Archived from groups: rec.games.roguelike.development (More info?)

Auric__ wrote:
> On Thu, 30 Jun 2005 11:17:00 -0400, Shedletsky wrote:
>
>

>>It occurs to me that any nonunique monster in a roguelike with more than a
>>~2% chance of killing the PC in any encounter is probably too powerful to
>>include in a game.
>
>
> So, you want to coddle the players? "Gee, I'm sorry, monsters shouldn't
> be able to kill you unless they're very lucky" ? B-O-R-I-N-G. What's the
> point of playing if there isn't *some* challenge? If the monsters are
> too easy then it's just a mindless hack-n-slash.

In your average roguelike, the kill list easily reaches several thousand
monsters. Say 2000. If only 10% of these monsters has a 2% chance of
killing the player if the player uses the optimum tactics to try and
defeat them, this reduces the player's chance of winning to 0.98^200 =
2.8 * 10^-18. Now that's pretty small. And that's with only 200
monsters, each of which has a 2% chance to kill the player.

Archived from groups: rec.games.roguelike.development (More info?)

> In your average roguelike, the kill list easily reaches several thousand
> monsters. Say 2000. If only 10% of these monsters has a 2% chance of
> killing the player if the player uses the optimum tactics to try and
> defeat them, this reduces the player's chance of winning to 0.98^200 = 2.8
> * 10^-18. Now that's pretty small. And that's with only 200 monsters, each
> of which has a 2% chance to kill the player.

Pfft. I was saying that was the upper limit, not the average success rate of
monsters vs. the PC.

Archived from groups: rec.games.roguelike.development (More info?)

Shedletsky wrote:
>>In your average roguelike, the kill list easily reaches several thousand
>>monsters. Say 2000. If only 10% of these monsters has a 2% chance of
>>killing the player if the player uses the optimum tactics to try and
>>defeat them, this reduces the player's chance of winning to 0.98^200 = 2.8
>>* 10^-18. Now that's pretty small. And that's with only 200 monsters, each
>>of which has a 2% chance to kill the player.
>
>
> Pfft. I was saying that was the upper limit, not the average success rate of
> monsters vs. the PC.
>
>
As I said, if 10% of the monsters had this success rate, the odds
against success would be abysmally small. And that's not even counting
the 0,001 (0,1%) chance of success the other 1800 monsters might have.
0,999^1800 = 0,165, which is a 16,5% chance of success from trying to
kill those monsters alone.

Permadeath games are very hard to balance, which is why roguelikes have
a wealth of tactical options and inventory items that may help you
succeed in any match-up. The trick is to make the challenge for the
player what to do at any one point, rather than hoping to survive a
straight-up fight.

Archived from groups: rec.games.roguelike.development (More info?)

The problem with a purely statistical analysis of a particicular game
is that the odds vary incredibly depending on player sill, peparaion,
and most importantly, previous luck. how screwed you are in a
particular circumstance depends on alot of things, one of the most
important things being previus luck. if you make it to D:20 and
encounter 3 monstars on the way, your pretty much screwed for xp. or
maybe you have no range attacks yet and in crawl you encounter a
centaur or orc wizard in the early game. if the level is wide open,
your screwed. The way that probability dpends on all previous luck sets
up an essentially chaotic system, so any calculation of the pc's odds
has an error margin that's ridiculous. Not that I understand statistics
or anything.

My personal theory is that the game should provide a way out of every
circumstance. somehow, there should be some ridiculously obscure way
out of every situation, and it's different for each situation, so find
it or die. usually die.

As for the ridiculously low chances of suceeing, I think that the game
should work to prolong each game. Somehow dtermine whether the player
is doing excellantly or horribly, and adjust a "luck" factor
accordingly. perhaps this would give some interesting results, perhaps
this would lamify the game. it tries to keep you alive, but tries even
harder to keep you from winning. Along those lines, maybe I could even
get it to detect your skill. I've got a lot of work to tdo before I get
to that stage though

Archived from groups: rec.games.roguelike.development (More info?)

NIm wrote:

> As for the ridiculously low chances of suceeing, I think that the game
> should work to prolong each game. Somehow dtermine whether the player
> is doing excellantly or horribly, and adjust a "luck" factor
> accordingly. perhaps this would give some interesting results, perhaps
> this would lamify the game. it tries to keep you alive, but tries even
> harder to keep you from winning. Along those lines, maybe I could even
> get it to detect your skill. I've got a lot of work to tdo before I get
> to that stage though


Heh. I don't think the game should work to prolong each game;
let 'em die, hopefully learn something, and start new ones.

But yes, there's something to be said for a "luck" factor.
In at least a couple of the games out there, "luck" is an
attribute that can do a few nice things for you. One of those
things is that an attack that would normally kill you might
just leave you with a single-digit number of hitpoints
instead, regardless of the amount of damage.

And then there's a metagame of doing things that are known
(or suspected) to raise your luck.

Bear