Realism and Likelyhood (Was: HERO...)
Eduard W. Lohmann
Some comments.
Given the enormous number of events(is it even finite or countable?)
it is _very_ likely that during a characters preformance unlikely
things will happen.
If only likely things happen over a long period of time, this becomes
unrealistic(as in different from the real world). So please Do not mix
realism and likelyhood(probability) too freely.
In fact in Real Life(tm), things sometimes happen that seem to me
very unrealistic. I think many of us will have had the experience
of un 'unrealisticly bad die night' (Those that use dice, that is).
This is a prime example of a series of unlikely events.
------------------------< Eduard W. Lohmann >-----------------------------
Home : http://gene.wins.uva.nl/~lohmann/
Dragons-Inn : http://gene.wins.uva.nl/~lohmann/Inn.html
Email : loh...@wins.uva.nl
Rodney Payne
"Eduard W. Lohmann" <loh...@wins.uva.nl> writes:
>Given the enormous number of events(is it even finite or countable?)
>it is _very_ likely that during a characters preformance unlikely
>things will happen.
Of course. However, over a large number of trials, they will occur at a
rate which chance dictates.
So, while a run of ten heads on a single, unbiased coin is a likely
event when n is large, 99 heads over 100 trials very likely means a
biased coin.
>If only likely things happen over a long period of time, this becomes
>unrealistic(as in different from the real world). So please Do not mix
>realism and likelyhood(probability) too freely.
I don't think anybody is. To simulate `realism', there must also be
strings of unlikely events.
However, I don't think a bare statistical analysis is really what's at
issue. Everyone is pretty much agreed that we are talking about what
`feels real'. This will vary from group to group, and in some groups
strings of probabilities won't matter at all. What system is used to
generate outcomes is an issue for the group contract.
I think one of the reasons why dice are so popular for this purpose is
that they assure the group that the probability distribution will match
that which they have decided is appropriate. Even though they may not be
conscious of the production of a believable distribution, they know that
it's there.
--
Rodney Payne | The artist should organise his life. Here
| is a precise record of the time taken by
spur...@yoyo.cc.monash.edu.au | my daily chores: I get up at 7.18,
rgp...@cfs01.cc.monash.edu.au | inspiration 10.23 to 11.47.... Erik Satie
Eduard W. Lohmann
On 21 Feb 1997, Rodney Payne wrote:
> "Eduard W. Lohmann" <loh...@wins.uva.nl> writes:
>
> >Given the enormous number of events(is it even finite or countable?)
> >it is _very_ likely that during a characters preformance unlikely
> >things will happen.
>
> Of course. However, over a large number of trials, they will occur at a
> rate which chance dictates.
Yes, and oven a small number of trials they probably will not.
> So, while a run of ten heads on a single, unbiased coin is a likely
> event when n is large, 99 heads over 100 trials very likely means a
> biased coin.
But is was the 10^10 time one dit the 100 trials it would probably
happen.
> >If only likely things happen over a long period of time, this becomes
> >unrealistic(as in different from the real world). So please Do not mix
> >realism and likelyhood(probability) too freely.
>
> I don't think anybody is. To simulate `realism', there must also be
> strings of unlikely events.
>
> However, I don't think a bare statistical analysis is really what's at
> issue. Everyone is pretty much agreed that we are talking about what
> `feels real'.
(not everyone, if I understand the 'realism debate.)
> This will vary from group to group, and in some groups
> strings of probabilities won't matter at all. What system is used to
> generate outcomes is an issue for the group contract.
Yes, but the choise of words used in the thread and the talk of
distebutions suggests too strong a connection with probability.
An I have a non-determenistic view of life.
> I think one of the reasons why dice are so popular for this purpose is
> that they assure the group that the probability distribution will match
> that which they have decided is appropriate. Even though they may not be
> conscious of the production of a believable distribution, they know that
> it's there.
>
> --
> Rodney Payne | The artist should organise his life. Here
> | is a precise record of the time taken by
> spur...@yoyo.cc.monash.edu.au | my daily chores: I get up at 7.18,
> rgp...@cfs01.cc.monash.edu.au | inspiration 10.23 to 11.47.... Erik Satie
>
This was just a sugestion about choise of words when talking about
'realistic' simulations(games).
James Ellis
Eduard W. Lohmann wrote:
>
> In fact in Real Life(tm), things sometimes happen that seem to me
> very unrealistic. I think many of us will have had the experience
> of un 'unrealisticly bad die night' (Those that use dice, that is).
> This is a prime example of a series of unlikely events.
Actually, I think I'm in the midst of an 'unrealistically bad die
life' :)
Biff
Frank Pitt
>I don't think anybody is. To simulate `realism', there must also be
>strings of unlikely events.
>
> However, I don't think a bare statistical analysis is really what's at
>issue. Everyone is pretty much agreed that we are talking about what
>`feels real'. This will vary from group to group, and in some groups
>strings of probabilities won't matter at all. What system is used to
>generate outcomes is an issue for the group contract.
Hey, just had a great of example of this sort of thing happen to me
yesterday. We found an obviously magical artifact, a white, unmarked cube,
that one of the players claimed was an artifact from the
cult of the god of Luck.
So at least two of us rolled this die. One of the characters
noticed that as the die stopped rolling both the die and the
roller glowed briefly. But nothing else was noticed.
Then, some three game hours later, we were attacked by a flying
pterodactyl-like creature being ridden by a powerful mage.
We all fired at the creature, missing horribly, and barely
survived the first pass of the creature.
Then on the second pass, one of the characters, Jared, rolled
an open-ended roll and his arrow went straight to the brain of the
flying creature bringing it crashing down. The Mage began floating
toward the ground having bailed out, and my character Rizzo,
proceeded to do exactly the same thing to the mage.
Two extremely unlikley rolls, in quick succession.
But in character we all _knew_ why it had happened, we had been
blessed by the Luck god via the roll of the artifact !
Frankie
David Berkman
Rodney Payne (spur...@yoyo.cc.monash.edu.au) wrote:
: Of course. However, over a large number of trials, they will occur at a
: rate which chance dictates.
I owuld saythat the outcome of each trial will be due to reasonable
factors within the situation, ie. skill, choices, environment, etc. We
demand that these things have their proper effect in our games as well.
So, over a large number of trials, we are simply seeing the cummulative
effect of realistic outcomes given, for the most part, conscious choices
on the part of the participants. The frequency distribution which results
handles itself, and whatever it is, is realistic, because the individual
outcomes were realistic, and that's how it went.
[snip od stuff I don;t disagree with]
: I think one of the reasons why dice are so popular for this purpose is
: that they assure the group that the probability distribution will match
: that which they have decided is appropriate. Even though they may not be
: conscious of the production of a believable distribution, they know that
: it's there.
Which seems odd. In life, the distribution is a 'chance' result of
whatever occured, for whatever reasons were present. Forgetting the
distribution, and simply making sure that the reasons for each outcome
are sufficiently believable in an RPG ought to then guarantee a
believable distribution, without even trying for explicitly. I mean,
that's the way life works.
Instead, many mechanics fix the distribution, and make the individual
outcomes of actions a 'chance' result of this distribution.
That seems somehow backwards and wrong to me.
David
Joshua Macy
David Berkman wrote:
>
...snip...
>
> Which seems odd. In life, the distribution is a 'chance' result of
> whatever occured, for whatever reasons were present. Forgetting the
> distribution, and simply making sure that the reasons for each outcome
> are sufficiently believable in an RPG ought to then guarantee a
> believable distribution, without even trying for explicitly. I mean,
> that's the way life works.
>
> Instead, many mechanics fix the distribution, and make the individual
> outcomes of actions a 'chance' result of this distribution.
>
> That seems somehow backwards and wrong to me.
>
It makes perfect sense to me, because what we see in real life is the
distribution. For instance, suppose that I notice that my favorite
parking-place is empty when I want to park there at 8 AM about half of
the time. Now, I know that whether it is empty or not is a result of
decisions made by bunches of people who had an opportunity to park
there, and either didn't or did and left again. Presumably none of them
flipped a coin to decide what to do, but if I were to simulate (say for
game purposes) whether that parking place is available, flipping a coin
or rolling a die produces a "realistic feeling" distribution; even if I
hide the fact that I'm making a die roll, the player whose character
wants to park there each day will induce from the evidence the same
conclusion that I drew in real life--that it's available about half the
time. Creating the same believable distribution through "making sure
that the reasons for each outcome are sufficiently believable" hardly
seems "simple" to me at all--I would have to simulate vastly more of the
world than I have normally have access to even in the world. That is to
say, I'd have to decide (for believable reasons) how many people had the
opportunity to park there, and for each one (for believable reasons)
whether they decided to park, and if they did whether (for believable
reasons) they left again before 8 AM, reiterating until it's 8 AM and
the spot is either empty or not. And after doing this each simulated
day, I might still fall short of providing the believable distribution I
was aiming for through some unnoticed bias in the way I generated and
evaluated believable reasons.
Mark Apolinski
Joshua Macy wrote:
>
> It makes perfect sense to me, because what we see in real life is the
> distribution. For instance, suppose that I notice that my favorite
> parking-place is empty when I want to park there at 8 AM about half of
> the time. Now, I know that whether it is empty or not is a result of
> decisions made by bunches of people who had an opportunity to park
> there, and either didn't or did and left again. Presumably none of them
> flipped a coin to decide what to do, but if I were to simulate (say for
> game purposes) whether that parking place is available, flipping a coin
> or rolling a die produces a "realistic feeling" distribution; even if I
> hide the fact that I'm making a die roll, the player whose character
> wants to park there each day will induce from the evidence the same
> conclusion that I drew in real life--that it's available about half the
> time.
Yes, exactly. David's point is that reality comes before the statistics,
which you seem to grasp in your example above. You know that the result
is *really* a result of hard determinism. And yes, flipping a coin or
rolling a die produces a "realistic feeling" distribution. So far, we
are in agreement.
David and I apparently put more of an emphasis on how things "really"
are than how they "appear" with regards to randomness.
> Creating the same believable distribution through "making sure
> that the reasons for each outcome are sufficiently believable" hardly
> seems "simple" to me at all--I would have to simulate vastly more of the
> world than I have normally have access to even in the world. That is to
> say, I'd have to decide (for believable reasons) how many people had the
> opportunity to park there, and for each one (for believable reasons)
> whether they decided to park, and if they did whether (for believable
> reasons) they left again before 8 AM, reiterating until it's 8 AM and
> the spot is either empty or not.
Are you serious?! No offense, but this seems to be a very common
misconception of how diceless games work. You do *not* have to simulate
vastly more of the world. You do *not* have to decide how many people
had an opportunity. You do *not* have to figure out their actual
decisions. You do *not* have to figure out whether they left again. You
do *not* have to reiterate. You do *not* have to have an entire universe
simulation running inside your head (although this helps :).
All you have to do is *decide* whether the spot is open or not when the
PC gets there. And asking seriously, is there some reason why some
people have to make diceless techniques seem harder than they really
are? I'm not trying to offend, I'm just dumbfounded.
Mark
Joshua Macy
Mark Apolinski wrote:
>
> Joshua Macy wrote:
> > Creating the same believable distribution through "making sure
> > that the reasons for each outcome are sufficiently believable" hardly
> > seems "simple" to me at all--I would have to simulate vastly more of the
> > world than I have normally have access to even in the world. That is to
> > say, I'd have to decide (for believable reasons) how many people had the
> > opportunity to park there, and for each one (for believable reasons)
> > whether they decided to park, and if they did whether (for believable
> > reasons) they left again before 8 AM, reiterating until it's 8 AM and
> > the spot is either empty or not.
>
> Are you serious?! No offense, but this seems to be a very common
> misconception of how diceless games work. You do *not* have to simulate
> vastly more of the world. You do *not* have to decide how many people
> had an opportunity. You do *not* have to figure out their actual
> decisions. You do *not* have to figure out whether they left again. You
> do *not* have to reiterate. You do *not* have to have an entire universe
> simulation running inside your head (although this helps :).
>
> All you have to do is *decide* whether the spot is open or not when the
> PC gets there. And asking seriously, is there some reason why some
> people have to make diceless techniques seem harder than they really
> are? I'm not trying to offend, I'm just dumbfounded.
>
No, I'm not serious, but I'm taking what David said seriously.
(There's a difference.) I'm presuming from what David said (about
creating a "believable distribution") that he's talking from the
simulationist point of view. If you're not simulationist, then it
doesn't matter whether the distribution is believable: million-to-one
shots really *can* come up nine times out of ten.
If you intend to create a believable distribution, there are two basic
ways to go about it. The diced way is to pick a distribution that's
believable, then use the dice to pick an outcome along the
distribution. Presto: instant results that will in aggregate conform to
the distribution. The diceless way, according to David, is to "simply
decide for believable reasons" and trust that in the aggregate the
believable reasons will result in a believable distribution. The simple
fact is that if you just *decide* whether the spot is open when the PC
gets there, in the aggregate your decisions will *not* conform to a
believable distribution, and there's no reason that they _should_ UNLESS
YOU SIMULATE EVERY ASPECT OF THE SITUATION. The problem is that the
distribution of parking availability isn't the result of a single reason
(for which I admit it's easy to come up with a believable reason), but
the repeated application of dozens, hundreds, or thousands of reasons.
"Nobody parked there because yesterday was street cleaning" is a
believable reason today. It's a believable reason tomorrow. It's a
believable reason the next day. Just about any _given_ day, if the GM
told me that the parking place was free because yesterday was street
cleaning, I'd accept it....but not EVERY day. The aggregate of the
individually believable reasons (in this case reason singular) does NOT
result in a believable distribution. You can attempt to patch it by
creating more believable reasons, but the central dilemma remains the
same: even though the distribution in life is the result of nothing more
than the accumulation of (ipso facto believable) reasons, there's no
reason to suppose that the GM's accumulation of believable reasons will
approximate that of life (and every reason to suppose that it won't,
human propensity for bias being what it is) until and unless the number,
diversity, and distribution of the reasons approximates that of life.
You can't expect to simulate an epiphenomenon at all accurately without
simulating all the parts that give rise to it--unless you directly
simulate the epiphenomenon (which is what the diced GMs do).
Parking may seem like a silly example, since most likely no player is
actually going to perform the activity enough times to notice the
distribution or be bothered by it, but there _are_ activities that
players do perform regularly enough that they notice the distribution,
and if the distribution is not realistic it will bother them.
Rodney Payne
bcks...@crl.com (David Berkman) writes:
>I owuld saythat the outcome of each trial will be due to reasonable
>factors within the situation, ie. skill, choices, environment, etc. We
>demand that these things have their proper effect in our games as well.
>So, over a large number of trials, we are simply seeing the cummulative
>effect of realistic outcomes given, for the most part, conscious choices
>on the part of the participants. The frequency distribution which results
>handles itself, and whatever it is, is realistic, because the individual
>outcomes were realistic, and that's how it went.
Consider the old example of leaping a 15' chasm. I really don't know what
factors go into success here, other than leaping ability, and perhaps
fear or vertigo. Let's say the character is defined as having no fear of
heights, and no balance problems.
Since I have no idea what external factors go into his success or
failure, all I'm left with is the statistical distribution. In my case,
conforming to a distribution of success/failure substitutes for the
external factors in terms of believability. I will believe the game if a
series of such jumps result in a believable rate of success.
Better examples might be non-fatal repeated events, like as a _Call of
Cthulhu_ character's Spot Hidden skill. I don't really know what goes
into a successful search attempt, I'm not interested in having the player
tell me where he or she is searching (too dull, and it's often not
applicable), so I only have the statistical distribution. This can be handled
any number of ways, but I prefer dice.
>Which seems odd. In life, the distribution is a 'chance' result of
>whatever occured, for whatever reasons were present. Forgetting the
>distribution, and simply making sure that the reasons for each outcome
>are sufficiently believable in an RPG ought to then guarantee a
>believable distribution, without even trying for explicitly. I mean,
>that's the way life works.
>Instead, many mechanics fix the distribution, and make the individual
>outcomes of actions a 'chance' result of this distribution.
>That seems somehow backwards and wrong to me.
Perhaps the `stopping power' statistics bandied about here will shed some
light here. Let's assume they are accurate. If the players in interested
in `real' reality, the figure of 65% incapacitation for .45 shots to the
torso will be useful. Recall that such a player doesn't care whether the
distribution feels real, he or she wants to know that it _is_ real.
Now, I suspect that you'll argue that not all torso shots are alike,
and probably they aren't. However, what we are looking at here is a
player who wants an accurate abstraction, not real seeming details. In
the case of real seeming details, I suspect that the final distribution
is irrelevant. What will be important often, however, is that the character
perform at the level at which the player has defined her or him, whether
this be in absolute or relative terms. See my recent post where I discuss
defining a character's ability in terms of mechanics as an aid to
consistency--a Dejanews Power Search under `abstract', `consistent',
`itemised', and my name, with the Current field selected should do.
Jeff Stehman
David Berkman (bcks...@crl.com) wrote:
: The frequency distribution which results
: outcomes were realistic, and that's how it went.
To me the overall is not made realistic simply because the individual
outcomes were realistic. If in a game you say fair coin toss results in
heads, that feels realistic. If you say a second toss also results in
heads, that feels realistic, and overall the series of tosses feels
realistic. If you repeat that a thousand times, the 1002nd time 'heads'
feels realistic, as there's a 50/50 chance. Overall, however, it
doesn't feel realistic.
A single heroic deed (ala Rambo, not the 'I take the bullet for him'
kind) in a movie doesn't necessarily make the movie feel unrealistic.
Lots of heroic deeds however, even if they are individually realistic,
and the movie feels heroic, not realistic.
Whether this view is held by the majority or not, I don't know, but
from past threads I know I'm not alone in holding it.
--
Jeff Stehman Senior Systems Administrator
ste...@southwind.net SouthWind Internet Access, Inc.
voice: (316)263-7963 Wichita, KS
URL for Wichita Area Chamber of Commerce: http://www.southwind.net/ict/
David Berkman
Rodney Payne (spur...@yoyo.cc.monash.edu.au) wrote:
: Consider the old example of leaping a 15' chasm. I really don't know what
: factors go into success here, other than leaping ability, and perhaps
: fear or vertigo. Let's say the character is defined as having no fear of
: heights, and no balance problems.
: Since I have no idea what external factors go into his success or
: failure, all I'm left with is the statistical distribution.
No, you haven;t got that either. I guess you could use average long jump
distances, but leaping a chasm is different from that as well. Even if
you are fearless, gripping rock is defferent from going feet forward into
sand. There is no statistical distribution. You can make one up that
pleases you. But that's all.
: In my case,
: conforming to a distribution of success/failure substitutes for the
: external factors in terms of believability. I will believe the game if a
: series of such jumps result in a believable rate of success.
I have been roleplaying for 20 years now, and, despite how often leaping
chasms comes up, which is fairly frequently as far as any one class of
dramatic actions goes, if you took all the examples form all such leaps,
across games systems, and characters, you wouldn't have enough data to
base a judgement upon. Well, that's not quite true. I have never had a
charcater miss and fall to his/her death, outside of the D&D spiked pit
trap. Nor have I ever seen this happen in any game I've ever watched. I
guess there's just something about 'roll, clatter, you die' that keeps
people from it. But, even discounting that (which is a lot to discount),
you still wouldn;t know. Actual action is just too varied and complex to
stuff into such neat and observable statistical packages.
How many chasm leaps can you remember from your own games, and what were
the results.
: Better examples might be non-fatal repeated events, like as a _Call of
: Cthulhu_ character's Spot Hidden skill.
That has to be the silliest 'skill'. I hate perception skills, and I don't
use them. how well you percieve depends on what you're percieving, and
that is dependent on how skilled you are in the area.
Beyond that, what's a 'realsitic' distribution for perception skills? I
don't know. I don;t think anybody else does either.
Basically, if I have to make up a distribution, I'll just cut to the
chase and make up an answer. If no one knows what details are important,
that means I can simply guestimate with impunity. If we are simply left
with 'it feels right', which I think we are in almost all cases, then
it's fine.
The dice aren't going to tell me anything I don't already know.
If you like using them, fine, but they aren;t for me.
: Perhaps the `stopping power' statistics bandied about here will shed some
: light here. Let's assume they are accurate. If the players in interested
: in `real' reality, the figure of 65% incapacitation for .45 shots to the
: torso will be useful. Recall that such a player doesn't care whether the
: distribution feels real, he or she wants to know that it _is_ real.
: Now, I suspect that you'll argue that not all torso shots are alike,
: and probably they aren't...
Well, combine that with FBI hit statistics, and the statistics for
surviving gun wounds, and you have... mud. You can guestimate an answer
and be as close as any RPG system is going to get you.
: However, what we are looking at here is a
: player who wants an accurate abstraction...
When you find one, tell me.
: ...not real seeming details.
Those are all I've ver seen.
: In
: the case of real seeming details, I suspect that the final distribution
: is irrelevant.
I suspect that in life the final distribution is irrelevant.
: What will be important often, however, is that the character
: perform at the level at which the player has defined her or him, whether
: this be in absolute or relative terms.
Sure, I agree with that, completely.
: See my recent post where I discuss
: defining a character's ability in terms of mechanics as an aid to
: consistency
As an aid to consistency, sure. Probably useful. I can do consistency I'm
happy enough with, without the dice, and I like a lot of the feel I get
by dropping them much more, and that's even more important to me.
David
Rodney Payne
bcks...@crl.com (David Berkman) writes:
>No, you haven;t got that either. I guess you could use average long jump
>distances, but leaping a chasm is different from that as well. Even if
>you are fearless, gripping rock is defferent from going feet forward into
>sand. There is no statistical distribution. You can make one up that
>pleases you. But that's all.
The one can be inferred from the other.
The key here, though, is `make one up that pleases you'. In other
words, one that `you' find believable in consensus with your group. Let's
move this away from a discrete success or failure roll. John Kim has said
he thinks he can jump about 7', maybe a few more if he doesn't worry
about landing on his feet--seems fair to me. I'd be happy to say he could
jump, on average, 10' (three metres--the world record is about 8.5), with
a variability of about three feet either side (one metre, fairly consistent
with the variability of long jumpers). If he didn't mind an uncomfortable
landing.
A d10 here could be used to model a (rough) normal distribution around
the mean of 10'. For a 10' chasm, a leap of 10' or more gets you across.
Further, I decide that a leap of (n-1)' (9') will give you a chance to grab
the cliff edge.
Now I have an approximate model chasm leaping which is believable (to
me), objective, and consistent.
>I have been roleplaying for 20 years now, and, despite how often leaping
>chasms comes up, which is fairly frequently as far as any one class of
>dramatic actions goes, if you took all the examples form all such leaps,
>across games systems, and characters, you wouldn't have enough data to
>base a judgement upon. Well, that's not quite true. I have never had a
>charcater miss and fall to his/her death, outside of the D&D spiked pit
>trap. Nor have I ever seen this happen in any game I've ever watched. I
>guess there's just something about 'roll, clatter, you die' that keeps
>people from it.
Well, yes, but this is hardly an issue of randomisers. Anyone can
overrule a die roll.
>But, even discounting that (which is a lot to discount),
>you still wouldn;t know. Actual action is just too varied and complex to
>stuff into such neat and observable statistical packages.
This seems self-evident, but it's really not. Many real world phenomena
can be, and are, modelled statistically. This technique is especially
useful when you don't know all the factors contributing to an outcome. A
roulette wheel is an extremely complicated physical phenomena--as is
witnessed by its popularity with casino owners--yet its output can be
modelled statistically over a range of trials.
>How many chasm leaps can you remember from your own games, and what were
>the results.
Oh heavens, I have trouble remembering last week.
I don't think the `across games' contention is relevant. For the most
part, I'm only concerned about believable distributions in single
games--and not in every game: I play a lot of comedic stuff.
>: Better examples might be non-fatal repeated events, like as a _Call of
>: Cthulhu_ character's Spot Hidden skill.
>That has to be the silliest 'skill'. I hate perception skills, and I don't
>use them. how well you percieve depends on what you're percieving, and
>that is dependent on how skilled you are in the area.
True. I'm terrible at spotting visual pattern anomalies (well, not
terrible, but my spouse puts me to shame), and very good at spotting
people in a crowd. But then, I doubt many GMs, diceless or otherwise,
make that fine a distinction.
As to Spot Hidden and the like, I think they are very useful skills.
Nothing bores me more than a player making a detailed account of how she
or he is looking for something. Not to mention when the character isn't
looking.
>Beyond that, what's a 'realsitic' distribution for perception skills? I
>don't know. I don;t think anybody else does either.
This relates to what I have been contending on other threads--that
mechanics, whether they be of the `good, very good' or the `60%, 80%'
variety, allow the player to delineate how good her or his character is,
_relative to the other characters, or the game world_.
A believable (I have not used the term `realistic', and if I have, I
withdraw it now) distribution is one which matches this concept of the
character. That is, a character with a 80% skill is a third again as
perceptive as one with 60%, and will perform accordingly.
What I am curious about is how you maintain these sorts of distinctions
when the characters' skills are described textually. In my systemless
games I make use of textual descriptions, but I don't claim to always
maintain these differences clearly or consistently, and for a player for
whom they are important my systemless style may be unsuitable. I am
interested if and how you, or other systemless GMs, do this.
>Basically, if I have to make up a distribution, I'll just cut to the
>chase and make up an answer. If no one knows what details are important,
>that means I can simply guestimate with impunity. If we are simply left
>with 'it feels right', which I think we are in almost all cases, then
>it's fine.
I doubt that, say, David Summers would be happy if you `guestimate[d]
with impunity'.
>If you like using them, fine, but they aren;t for me.
Are you concerned that some of us haven't realised this? ;)
>: Now, I suspect that you'll argue that not all torso shots are alike,
>: and probably they aren't...
>Well, combine that with FBI hit statistics, and the statistics for
>surviving gun wounds, and you have... mud. You can guestimate an answer
>and be as close as any RPG system is going to get you.
I can't contest that any more than you can claim it, as both of us have
professed here to knowing almost nothing about guns.
>: However, what we are looking at here is a
>: player who wants an accurate abstraction...
>When you find one, tell me.
David Summers. Has someone been using your account behind your back for
the last two months?
>I suspect that in life the final distribution is irrelevant.
Certainly. A gunshot victim isn't thinking distributions. But a gunshot
victim is not an external observer, new age religions aside.
>: What will be important often, however, is that the character
>: perform at the level at which the player has defined her or him, whether
>: this be in absolute or relative terms.
>Sure, I agree with that, completely.
Ah, good. As I said above, I'm interested in how you go about this--in
some detail, if you don't mind taking the time.
Jim Henley
Rodney Payne wrote:
> What I am curious about is how you maintain these sorts of distinctions
> when the characters' skills are described textually. In my systemless
> games I make use of textual descriptions, but I don't claim to always
> maintain these differences clearly or consistently, and for a player for
> whom they are important my systemless style may be unsuitable. I am
> interested if and how you, or other systemless GMs, do this.
I wouldn't call Theatrix "systemless." (One presumes David uses Theatrix
in his diceless games, since he designed it.) And in addition to textual
descriptions, Skills and Attributes _do_ have numbers attached to them,
of the "On a scale of 1 to 10 . . . " -- really 0-10 -- variety. One of
the jobs of a referee (ahem, Director, to use the official term with its
perhaps faint whiff of pretentiousness) is to set the Scaling of traits
and relevant Skills and Abilities according to the Setting of the game.
Generally, "typical human" attributes clock in at 3. In a sword and
sorcery game you might decide that 6 was Conan-level Strength, 7
equivalent to Greater Ajax (Aias for you Greek purists), 9 for Giants
and 10 for Dragons. The scale would be nonlinear in any case, but it
does involve setting some approximate numbers for what can be carried
and so on.
OTOH, a contemporary Spy/Eye game would compress the scale considerably.
3 would still be an average Strength (coordination, what--have-you)
while 9 might be an Olympic level weight-lifter and 10 reserved for
Mongo or someone who's been enhanced in the lab. A Marvel-style game
might set 10 as Hulk/Thor-level Strength with, say, Galactus off the
scale -- the point of Galactus is not to pummel him into
unconsciousness. Captain America would clock in around 5 and Spiderman
at 7. A DC game, though, would set Superman-level Strength at 10 and a
Hulk-level character might only clock in at 8. Again, the GM would
define in detail whether a 10-Strength character could lift a
battleship, a tank or yugo depending on the genre. In my fantasy Setting
the scale says a Physique of 9 can knock down trees while a Physique of
10 can shatter them. The point is, it is -- IMHO -- a _system_.
Skill levels are defined in terms of how capable the character is at
doing tasks of scaled levels of difficulty from Easy to Impossible. Some
people will complain that the 0-10 scale sets an upper limit on skill
achievement but this doesn't bother me -- e.g. once you can score 10,000
points using the rail nurse in straight billiards the concept of
"improvement" is probably meaningless. You've reached the point where
your ability to score 20,000 will depend more on environment, your own
stamina and chance than your shooting skills in themselves.
Best,
Not David, But What the Heck
----------------------------------------------------------------------
"I took my pistol and a hundred dollar bill
I had everything I needed to get me killed" -- Steve Earle
----------------------------------------------------------------------
Rupert Boleyn
In rec.games.frp.advocacy, bcks...@crl.com (David Berkman) did say:
Snip
>I have been roleplaying for 20 years now, and, despite how often leaping
>chasms comes up, which is fairly frequently as far as any one class of
>dramatic actions goes, if you took all the examples form all such leaps,
>across games systems, and characters, you wouldn't have enough data to
>base a judgement upon. Well, that's not quite true. I have never had a
>charcater miss and fall to his/her death, outside of the D&D spiked pit
>trap. Nor have I ever seen this happen in any game I've ever watched. I
>guess there's just something about 'roll, clatter, you die' that keeps
>people from it. But, even discounting that (which is a lot to discount),
>you still wouldn;t know. Actual action is just too varied and complex to
>stuff into such neat and observable statistical packages.
>How many chasm leaps can you remember from your own games, and what were
>the results.
OK then, In my mere 15 or so years gaming it hasn't come up hellishly often, but
when it has people have fallen, or not according to whatever set of
rules/mechanics (I'm including GM decision in that, for this case), and suffered
from the fall in accordance with whatever model of falling was being used. Some
lived, some died, and a lot ended up with broken arms & legs, especially if you
include catastrophically failed climbs. Now I've got a pretty small sample of
actual leaping, as it's never come up much, but still. It sounds to me that your
experience includes a lot of GM's who aren't consistent with the rules they're
using, presumably because they go in for cinematic scenes, rather than a hard,
simulationist feel (or maybe me, and all the GM's I've had over over the years
are just unreasonable hard-arses).
>David
My NZ$0.02
David Berkman
Rupert Boleyn (gtru...@iconz.co.nz) wrote:
: OK then, In my mere 15 or so years gaming it hasn't come up hellishly often, but
: when it has people have fallen, or not according to whatever set of
: rules/mechanics (I'm including GM decision in that, for this case), and suffered
: from the fall in accordance with whatever model of falling was being used.
O.K. What happened? What is the ratio of success and faliure? What were
the consequences? Why do you beleive they were realistic? Why don;t you
believe they were realsitic?
I think you take any particular type of action, or all actions as a whole
even, and quickly get into the same quagmire. Action and interaction in
anything believably realistic is simply too complex to get a statistical
handle on. And we don't judge realsim that way (in my humble opinion). We
judge realism from the facts and conditions at the moment, and as long as
the results conform there, and don't otherwise form some blaring pattern,
that's enough.
David
Rodney Payne
Jim Henley <jhe...@cais.com> writes:
>I wouldn't call Theatrix "systemless." (One presumes David uses Theatrix
>in his diceless games, since he designed it.)
David has stated in the past that _his_ games are actually systemless,
and that part of the purpose of Theatrix is to move players away from a
mechanised style by gradually doing away with the game's frameworks. As
such, I assume he has, in practice, done away with the numerical
descriptors. Nonetheless, I agree with you that Theatrix isn't systemless.
So, the original question remains. How do GMs who rely solely on
textual descriptions maintain consistency with regard to the relative
abilities of the characters within their games?
(Thanks for the outline of the system, BTW.)