Re: GM-controlled Bad Luck points

On Mon, 23 Jul 2007 15:42:04 +0100, Simon Smith
<simon_smith_news@[EMAIL PROTECTED]
> wrote:

>In message <46a4aee7$0$4695$4c368faf@[EMAIL PROTECTED]
>
>          "gleichman" <fox1_217.nospam@[EMAIL PROTECTED]
> wrote:
>
>> 
>> "Simon Smith" <simon_smith_news@[EMAIL PROTECTED]
> wrote in message 
>> news:401024064f.zen44412@[EMAIL PROTECTED]
>> > Now here I'm with Peter. I too like variable-numbers-of-dice
mechanics. I
>> > like competent characters to be consistently competent. 'Consistently
>> > competent' is a bit of a nebulous phrase, but among other things it's

>> > going
>> > to include fumble chances for routine tasks on the order of 1/1000 or

>> > less,
>> > which gibe reasonably well with real-world failure/fumble chances for
>> > routine tasks.
>> 
>> This is a good example of what I meant by "chasing an illusionary
target 
>> that's meaningful only to him".
>> 
>> Many games designs (AoH being one example) do not invoke fumble rules
for 
>> routine tasks, but rather only for stressed or combat conditions. Even
games 
>> where that isn't directly stated are often ran in that style- people
don't 
>> roll skill rolls for walking up stairs, taking a bath, or driving to
work 
>> (although all those are causes of death in the real world) <snip>
>
>
>Here's an attempt to articulate why ultra-low (but non-zero) failure and
>fumble probabilities are sought by some game designers (incl. me and
Peter,
>it appears):
>
>
>To many strict simulationists, the view given by gleichman, while
pragmatic,
>is bothersome. When considering a game world, I like to assume that the
game
>rules apply universally. That is to say, even off-camera, the plebs are
>driving to work, driving home, crossing roads and so on and - at least
>theoretically - dice are being rolled by some nebulous GM in the sky to
>determine what happens. So if the rules as written give a 1% chance of
each
>NPC being run over every time they cross the road, that spoils the game
for
>me. I stop trusting the rules because I know they generate ludicrous
results
>for routine tasks that they ought to handle properly. (Even though those
>routine tasks are being done by NPCs off-camera, and do not impinge on
the
>game.) And if I don't even trust the rules for routine tasks, I'm
strongly
>disinclined to trust them for anything more demanding, such as the stuff
PCs
>usually get up to. Inability to sensibly handle routine operations breaks
>SoD. So, given any system that /has/ an automatic failure or fumble
>mechanic, the dice system, whatever it is, must be able to generate
sub-1%
>probabilities, or it won't be able to handle road-crossing and other such
>trivia in a sane way, and I for one am not going to be happy running in
that
>system.
>
>Such a target may only be meaningful to a small majority of us, but for
>those of us who do worry about the integrity of the rules systems we use,
>it is a dispro****tionately big deal.
>
>Having a rule that fumbles only apply in stressful situations is one
>possible fix, of course, and it's certainly one I use in some systems.
But
>there are other legitimate ways to achieve the same end. And certainly,
if a
>game system /can/ handle tiny failure chances elegantly, I see no reason
to
>prevent it from doing so. I also feel that trying to design a system that
>can do so is a worthwhile goal.

There are also simulationists (me, for example) who consider it better
to ignore very small chances of fumbles, "rounding the chance down to
zero" rather than ginning up a mechanic to handle such small chances.
Even a 1/1000th chance of a fumble is often orders of magnitude too
large, and a 1/1000th chance is about as small as any reasonable
mechanic can deliver. 

If the 'true' chance of a fumble occuring is a 10% chance of a fumble
in 10,000 trials, then a system that gives 0 fumbles (because it
rounds the chance to zero) is going to be more accurate than a system
that gives even 1 fumble, on average, in those 10,000 trials - to say
nothing of a system that gives an average of 10 fumbles per 10,000
trials (a 1/1000th chance of a fumble per trial). 

And there are a couple of other factors: First, for very common tasks
the "independant Bernoulli trial" assumption may no longer hold.
Someone who drives to a given location 200 times a year may very well
NOT have a 10x higher chance of an accident along that drive than
someone who drives to that location only 20 times a year. 

Second, the game setting may well be one where fumbles are even less
likely than in real life, at least for PC-caliber characters. Since
games systems that incor****ate fumbles already have a strong tendency
to produce too-frequent fumble results, that makes a "round fumble
chances to zero" approach even more accurate for the game setting. 

As I've mentioned before, *my* idea of a "good" fumble system is "roll
1d6. If the die breaks into two or more pieces then the character
suffers a fumble." (OK, slight exaggeration. If a character is doing
something particularly fumble-genic, then I'll allow what the
fumble-system-fans consider an "ordinary" chance of a fumble. E.g. if
a character is using an improvised or cheap-piece-of-**** weapon then
I'll allow a <1% chance of a "weapon breaks" fumble result.)

-- 
Erol K. Bayburt
ErolB1@[EMAIL PROTECTED]