News:

Forum changes: Editing of posts has been turned off until further notice.

Main Menu

Beeg Horseshoe Theory Revisited

Started by Mike Holmes, May 29, 2003, 03:13:46 PM

Previous topic - Next topic

Ron Edwards

Hello,

I suggest that people might be interested in reading my breakdown of Gamist play in my new essay (I was expecting today, probably in a day or so) before we get too definite about "challenge," and being too sure about G+N play.

Mendel, in particular, you have mentioned "decisions" a few times, but I think you might be thinking at a scale well below what I consider a valid instance for discussing GNS.

Best,
Ron

Mike Holmes

Sorry for the line by line, but I think it's appropriate here.

QuoteSo that would mean you disagree with my assessment that this model is not useful for classifying Gam-Nar decisions. Are you saying that this hybrid is reflected by putting the dot towards the center of the axis?
No. One decision, two dots. Each at the same "elevation (Fidelity Axis), but further left or right based on how far they are from "neutral".

This was not my original intent, but it's a conclusion that has to occur to support the idea that decisions can be prioritize both. Originally I was thinking that you'd put one dot in the middle, but that's no good as it doesn't really tell you anything acurate. To the extent that this might seem overly complicated, this argues for the three D model.  


Quote from: Wormwoodpoints are individual decisions
Yes.

Quotecurves are player decision collections
Thinking more like scatter diagrams. I imagine that they'd look like shotgun blasts.

Quotecompact regions are locations in which a given design supports play.
Whoa. Who said anything about design? This is, like GNS, just a model that descibes decisions ijn play. It relates to design only in that a design will tend to produce certain patterns (and we want to avoid combinations that are far disparate, and have no mechanism to reconcile these far separate decisions).

QuoteEven if this is the case, that still implies that no decision can be both theme and challenge, just an eventual pattern of decisions. I'm more certain that that is false, as I recognize decisions of that type in my own play.
That's confusing. Decisions are plotted once. Play can be descibed in terms of averaging, but only onlong single axes (or, as I like to think of it, two at a time).

QuoteIf, rather, you are suggesting that a single decision may appear in multiple locations, then it seems that this is a poor representation of your ideas.
I wasn't origianally, but now I see that I'd have to as above. Again, this all argues for the 3D model.

QuoteSimply put, I feel that the model over corrects for the importance of Sim.
That was a definite danger. Can I assume you're then pusing for the 3D model?

QuoteI suspect that you are actually thinking in terms of bary-centric coordinates in a triangular system.
Nope, but that might work, ironically. OTOH, I think most people would be more comfortable in 3D. In case it's not clear what that would look like, individial decisions would be like stars in space, with patterns being like clusters of stars.

Mike
Member of Indie Netgaming
-Get your indie game fix online.

Wormwood

Mike,

I guess the long and the short of it, is I'm in favor of the 3D version.

When I mean designs support a compact space in the general space, is that a given design turns out to favor region of play, which can best be defined by some number of bounded portions of the space.

I often end up thinking of the standard GNS in bary-centric coordinates, but a full 3D view is signficantly easier.

I also can't shake off the feeling that what is actually being modelled is a projective space on something else, as is especially evidenced by multiple points being directly associated. On the other hand, I'm not sure I want to open than can of worms.

  -Mendel S.

M. J. Young

Quote from: Addressing me, Mike HolmesDoes the three axis model work for you then?
...and my initial reaction was that it didn't really, because it suggests that I've got multiple priorities involved in individual decisions;
Quote from: but then heGroups of decisions say something about overall play, but I contend that it's not that the player has a single priority; it's that the player has several priorities each being expressed in play to differing extents with each decision.
...and suddenly I feel like I'm doing vector addition, which I haven't done since Electronics back in high school and don't have a clue how to proceed (how do you multiply something by the square root of negative one anyway?)--but at the same time, I think I agree.

With a distinction.

I think there are two disparate activities; but let's look at an example and go from there.

We've got a decision to make, and we've got a choice that we can value as gamist 10, narrativist 6, simulationist 4. That is, I'm assuming here that a a gamist would view this as a very good way to reach his objectives, a narrativist as a moderate way, and a simulationist as a poor but not terrible way--but I'm quantifying these, because we have a three-vector graph.

Now, in a situation like this, it could be that we've got a gamist, narrativist, and simulationist playing together, and they could agree on the decision if there were no decisions whose value exceeded G10 or N6 or S4; that is, if there's no better choice for anyone, all would agree to this, despite the fact that it represents narrativist interests only moderately and simulationist ones poorly.

Note that this means we would put the decision itself at point (10, 6, 4) on our graph; but at the same time, none of the players cares about that position in total--only that on the one vector on which they are concerned, this was the highest value option.

We would call such play coherent, and if we could design a game in which all decisions always had one option which was best in all three modes, we would call it coherent hybrid design under one of those definitions (the Viet Nam War model).

This is distinct from the sort of player who is looking for something different, something which values all three modes. In this case, let us suppose that he wants decisions with a minimum sim value of 5, a minimum nar value of 7, and a minimum gam value of 8. He would look at this (10, 6, 4) option and compare it to his (8, 7, 5) values, and reject it, because although the gamist value is high enough to meet his desires and then some, both the sim value and the nar value fall short. Thus, despite the fact that he appears to favor gamism (a minimum value of 8), he might go for an option which provided (8, 10, 12), because it was the one that met all his minimums--and suddenly he appears to be prioritizing a simulationist concern.

What intrigues me about this model and makes me seriously consider it is that it might well explain much of my approach to play. Why am I gamist in one situation, simulationist in another, and narrativist in a third? Is it because I'm looking for choices in which each meet a certain minimum, or fall within a particular range, or even do not exceed a certain maximum? I'm not sure; but the three-vector model does raise the possibility, put as you have done.

Thanks for the discussion; this has been very profitable.

And I still think you're on to something with the expected level of fidelity as a separate issue.

--M. J. Young

Jason Lee

Heh, so in response to Mike's 'does anyone want to defend the model?' I was trying to collect my thoughts on the Is this really Nar? thread for a new assault on M.J.'s post with Sim/Nar Congruence as my battle axe.

However... It looks like I can make a much shorter post.

The three vectored model doesn't break anything for me, so if it clicks for other people I'm for it.  I can say Hi-Sim|Hi-Nar|Lo-Gam and that means functionally the same thing as Hi-Fi|Theme in my mind.

The three vector model maintains the reference to Congruent play as a single priority - which is the important part for me.  It also adds Nar/Gam Congruence to the mix - which I'm personally not sure anyone can actually do at a single decision point, but I do think it would be possible over an instance of play.
- Cruciel

Mike Holmes

I think that other than my objection to the fact that it's going to be hard to visualize, and with the caveat that each of the three axes (not just the sim one) has it's own peculiarities in relation to the others, I can accept the 3D model.

Basically, yes, MJ, all that stuff that you said is what I'm imagining. I'd alter it slightly to say that I think that players often are looking at two axes at once, and that one happens to be Sim more than the others, but that this is only a trend that I observe and has no weight a priori. That is, a group will tend to set minimum bars for each, but the Sim one is almost always considered. That is, for even the most Gamist or Thematic action, the sim vector for the decision must be at a level 1, which is the simple internal consistency level. That is, a player cannot say, "I kill the dragon by taking a nuke out of my pocket and forcing him to swallow it." when we all know that nukes don't exist/are too big for pockets/he didn't purchase one earlier. That's magnitude 1 on the sim scale to say that the action taken considers causality on the most basic level. No matter how much players want to allow Gamism, for example, they'll not go below this level.

That said, as I've said before, a postmodern RPG might allow for sim zero, so even this is just an observation about common play.

Stated otherwise, the most basic precepts of the game, the genre expectations, are the minimum sim requirement in most games. The only question is going higher on the sim scale is how many additional points of contact are involved in making the decision more sim: where does the text say you can buy nukes, how commonly available are they, how much do they cost, how much does a character with his profession make, how large are they, how large are pockets in terms of carry cap, etc. Consider all the above and the Sim level skyrockets. Consider only, "I went to get a pocket nuke in a Flash Gordon game", and you have only sim 1.

That's the sort of analysis that one can do with the model. Again, the way I see it, groups establish minimums. That is, players are never dissapointed with "too much" anything. They are only dissapointed when too much of one thing results in too little of another that they have a minimum requirement for. Thus, the Gamist, as an example a player with a minimum Gamism axis requirement of 7, has no problem with simmy or narrativist qualities, but tends to see problems in terms how adversely they affect the minimums. That is, if a narrativist decision because of lack of congruence also has a Gamism vector below 7, then he'll see the decision as problematic. He may or may not be able to identify the cause; but he sure as heck knows what the problem is. This example agrees with your analysis above MJ. But consider the case where a particular decision falls below two axes' minimums.

For the geometry heads, these three minimums on the axes form the "eigth space" of acceptable play for a player. The edges of which will, of course be fuzzy because players are humans and have to rely on imperfect perception, and are imperfectly consistent. In fact, a particular player's "box" will fluctuate in dimensions no doubt.

Again, power is just the authority granted by the game (defined as player using text or agreed to convention) to make decisions with a high vector along a particular axis. Thus, again, Narrativism does not clash with Sim, it clashes with whatever happens to cause the lack of narrativist power. Which may be a designer sim priority, but that's irrelevant.

As to Walt's note (one that I accidentally encouraged) that decisions tend to be problematic when hybrid, and hence less occuring in high combinations, what I'd say is that, if one plots actual decisions, there are fewer where there are multiple vecrtors with high magnitude. So plots of actual play will tend to look something like a leaning umbrellas in this model. That meaning, that density will tend to be found nearer the point at which the three minimums cross. But actually they'll not be densest there, but just a bit above. This is because players will tend to have a safety buffer that makes their decisions more certainly above the minimum. But, yes, as you get above the "umbrella", there are less and less points that occur in most games because there are fewer and fewer actual decisions that can be made that have all three vectors high. Or rather many decision points do not have these plots as representative of a possible choice.

This does not affect the space in which the decisions can be made. As long as there's a theoretical decision that's maximally all three in congruence, then the shape is a space or a box (if you assume some maximal value for a decision).

So, any problems with the 3D vector model of GNS?

Mike
Member of Indie Netgaming
-Get your indie game fix online.

ethan_greer

I'm on board with the 3D Vector model.  I dig it.

And, I stand by my previous post - but that's become a separate issue.

Cassidy

Quote from: M. J. YoungWe've got a decision to make, and we've got a choice that we can value as gamist 10, narrativist 6, simulationist 4. That is, I'm assuming here that a a gamist would view this as a very good way to reach his objectives, a narrativist as a moderate way, and a simulationist as a poor but not terrible way--but I'm quantifying these, because we have a three-vector graph.

For each decision though wouldn't you need to plot a point for each player based on how they personally perceive the decision as meeting their own expectations and GNS goals?

The guy making the decision in your example may quantify the decision as being G10,N6,S4 but what if another player with Narrativist priorities sees the decision a little differently, i.e. G13,N2,S5?

I would feel inclined to plot points based on how the player making the decision feels their decision meets the agreed upon GNS goals and expectations of the group as a whole.

C. Edwards

Hey Cassidy,

Here's how I see it.

Each individual is going to have their own particular layout on the graph for each decision. The person making the decision is going to have a layout, and each of the other people observing that decision will have a layout based on how they perceive that decision. You're also going to have what each player considers to be the baseline of the groups GNS priorities plotted onto the graph. The baseline will most likely vary between each player to one degree or another. So, even a player who thinks that his decisions fall within exceptable parameters may be breaking the boundary of another player's baseline assumptions.

-Chris

Mike Holmes

Quote from: CassidyFor each decision though wouldn't you need to plot a point for each player based on how they personally perceive the decision as meeting their own expectations and GNS goals?
Quite so. The minimums may seem consensual, but they may not be. The acceptable play region is defined personally. Coherence is achieved by the system providing a functional commonality of acceptable play regions, and players making decisions in them. The latter part is an essential part of the Social Contract and is assumed.

QuoteThe guy making the decision in your example may quantify the decision as being G10,N6,S4 but what if another player with Narrativist priorities sees the decision a little differently, i.e. G13,N2,S5?
Nothing's perfect. There will be occasions like this. Rarely is a player called on a single decision, however. When they are, often they can describe their intentions, and make it clear. Or the "true" nature of the decision will become clear later in other context.

But as it stands I'm aware of nothing that makes a decision exlicit in terms of what it is. That said, some things are more obvious than others. When a player chooses to attack the Baron because his SA indicates that he hates the Baron, that's pretty obvously narrativist. So I think that part of coherent design is sending strong signals in terms of the context of the decisions themselves.

But yes, there will be perceptual differences, and yes, sometimes they cause problems. I have this particular actual example that destroyed an entire game. A player was playing with his girlfriend, and did something that favored his character, and disfavored another player's. The player saw this a purely Social (and Gamist by the new definition), and it seemed to him to be completely without Sim merit. Sim Zero. To him there was no reason that the character would do what he did. To me and the other players, it was a valid choice in that we thought it was not out of character.

QuoteI would feel inclined to plot points based on how the player making the decision feels their decision meets the agreed upon GNS goals and expectations of the group as a whole.
Actually the group doesn't matter much other than they inform the player on what they think are acceptable minimums, and may therefore affect his. But that's Social Contract level. In effect, a decision is made, it's plotted on each player's space, and then it's seen for whether it succeeds in meeting the minimums.

So what analysis of a design does is it says that play in response to certain rules will likely result in decisions that have higher or lower magnitudes in each of the vectors. To the extent that the message is clear, and the rules do not conflict, the game can be said to promote coherent play. That is it's less likely (though not at all ever impossible) that decisions will occur that will dip below the minimum implied by the rules.

That's important as well, it's not just the other players that inform minimums, it's the rules themselves that do. If we're playing Adventure! you assume that most play will be have substantial Sim, except when it comes to using Dramtic Editing. At that point the Sim element drops very low, often to 1, and even to zero if you spend enough points. The rule actually says that with enough points you can do things that don't really make sense, thus providing us with an example of an explicitly sub 1 Sim decision. Maybe call it .5 because it doesn't seem to imply that you can take leave of all sense of causality.

Anyhow, a player reading the rule knows he's supposed to go with high suspension of disbelief (become tolerant of low Sim play), when this rule is used. And as such, it's a coherent rule (though one could debate the overall Sim coherency in Adventure!).

Mike
Member of Indie Netgaming
-Get your indie game fix online.

Wormwood

Mike,

I very much like the 3-D vector space. But a few suggestions, I'd make about the structure of the space.

Ultimately decisions are dynamic, and it may be a good idea to distinguish the current state of the game, and the currently present decision. The way I see it, is that the context that is Explored is modified by play decisions. If the context is appropriate for player goals, then it remains unmodified, or lightly modified. If it's not, the players start pulling the context around. I see these contexts as regions in the vector space, just like player preferences.

I imagine this game-play region jerkilly being pulled different directions, each player trying to keep part of it over their region of choice. This means we have a natural origin, the accepting decision, which just continues the Exploration as it had been. Calling this (0,0,0) we can open the entire 3-space to study, considering decisions such as (3,0,-2), a trade off between gamist and simulationist objectives. If the current state of the Exploration is high sim, and low gamist, a wide variety of players might make this decision. In the opposite case, the decision may be much less likely.

The thing I like about this perspective, is you can visually imagine the regions moving in a fairly stable, brownian way, for a highly coherent game, and in a jerky tug-of-war in an incoherent one. It also explains quite a bit of the sweet spotting of game play, as well as the evolution of player preferences during play. Heck, it even explains why people playing strongly in one mode, will often make decisions in other modes, simply because their core mode isn't threatened.

I'm still thinking about what this model could mean for hybrid design, in the very least the dynamical perspective would be vastly useful.

I hope that is food for thought,

   -Mendel S.

Walt Freitag

I think a three-space model is overkill for most purposes. The Beeg Triangle, with corners representing 100%G, 100%S, and 100%N respectively, gives you most of what you need. For instance:

- From a given position, you can represent drift that increases one of G, N, or S while leaving the ratio between the other two the same, as movement directly toward the appropriate vertex.

- From a given position, you can represent drift that sacrifices one priority to enhance another, leaving the prevalence of the third unchanged, as movement perpendicular to the line connecting the present position with the vertex representing the unaffected priority.

As I mentioned earlier in the thread, the triangle is a projection of the three-space onto the plane G + N + S = 1.0, bounded within the region in which G >= 0, N >= 0, and S >= 0.

It's generally conceded that most current functional play (leaving aside what might be theoretically possible) that is analyzable in GNS terms is positioned near the edges of the triangle -- that is, represents coherent play (is near one of the corners) or a hybrid with a clear priority and a strong supporting priority (along the edges, especially the N-S and G-S edges, excepting the centers of the edeges.) The centers of the edges represent the hypothetical balanced two-mode hybrids. Ron has opined that such play tends to be unstable, tending to drift closer to one of the adjacent vertices. (Also, drift that jumps across those center-edge points appears to be not uncommon).

What the 2-D projection lacks that the 3-space has is the idea that play can have "more of everything" (G + N + S > 1.0) or "less of everything" (G + N + S < 1.0). These are rather speculative ideas anyway, and I don't see any reason they can't be discussed while still using the triangle for most conventional GNS "mapping." Even reliably functional play in the center of the triangle (G = S = N = 1/3) is dubious under present theory.

- Walt
Wandering in the diasporosphere

M. J. Young

Quote from: In response to me, CassidyFor each decision though wouldn't you need to plot a point for each player based on how they personally perceive the decision as meeting their own expectations and GNS goals?

The guy making the decision in your example may quantify the decision as being G10,N6,S4 but what if another player with Narrativist priorities sees the decision a little differently, i.e. G13,N2,S5?

I would feel inclined to plot points based on how the player making the decision feels their decision meets the agreed upon GNS goals and expectations of the group as a whole.
I think this confuses the model with the illustration.

The problem is that in illustrating the model, I scaled the vectors so that I could talk about them coherently. The vectors don't really have an objective scale; they're scaled subjectively. All my statement that a given decision has a G10 value means is that it's more gamist than a G9 value and less so than a G11. If you call that a G13, but you still call the others a G10 and a G15, we're still agreed on the order of the choices along the scale, and merely debating how to express the units. That order  probably is not something that comes into question too often.

We always imagine three-dimensional graphs as being marked in cubes to represent by eye the position of any point; Mike is correct that three-dimensional arrays are difficult for some people to envision, but even so we've all done x/y graphs, and we always use graph paper to do these. Imagine, though, that you've got the graph and you've got the points plotted, but you don't have the increments marked. You can still say that decision M is more gamist than decision N, even if you scale them differently.

Similarly, there is no way to compare the scale of the three axes against each other. You say that what I call a G10 you might call a G13, and what I call an N6 you see as an N2; but what does that mean? It only means that you divide gamist steps into smaller increments and narrativist steps into larger ones. We never said that we were going to play at level N4, because that would be meaningless. Rather, we came to some sort of implicit agreement that there would be a minimum level of narrativism and all played at or above that level. That you call it "2" and I call it "6" means nothing.

What matters in your example is that the gamist player thought decision M was high enough to meet the minimum N value as he understood it, and the narrativist thought it was not. That doesn't really mean that the N value of the decision itself was disagreed; it only means that the minimum value for N (which we have already recognized is social, implicit, and in some sense amorphous) has not been agreed satisfactorily.

The one challenge presented so far to this would be
Quote from: what MikeA player was playing with his girlfriend, and did something that favored his character, and disfavored another player's. The player saw this a purely Social (and Gamist by the new definition), and it seemed to him to be completely without Sim merit. Sim Zero. To him there was no reason that the character would do what he did. To me and the other players, it was a valid choice in that we thought it was not out of character.
If it were really true that the offended player thought this was S0 while others thought it was higher, that would be a problem. However, I can't imagine that any action that was truly S0 would not be obviously so to everyone involved--taking the pocket nuke out and using it on the dragon, perhaps. There's no suggestion that the character broke the bounds of the possible with his decision. The question seems to be whether this was a credible action for the character in this context, that is sufficiently credible to meet the minimum S for the game. The offended character is, of course, biased; it may be that at this moment, not liking to have been out-gamed, rather than objecting that the G value was too high he objects that the S value was too low. He may just have raised his bar of minimum S values for play; that's more than I know. I'm inclined to think that there is an inherent stigma against claiming that anyone else's play is too high on any axis--you can't say that someone's play is too competitive, to thematic, too realistic, without sounding judgmental, so instead you say that it isn't enough of something else. Realistic was the thing that this player used, no doubt because it's easy to recognize that unrealistic play is bad (whereas non-competetive or non-thematic play would be seen as personal preference--a prejudice, not a fact).

So I think this notion of different people "rating" the elements differently is only true in the most superfluous sense--that of comparing elements on one vector with those on another. You can't have a scale that allows such comparisons. What matters is only what things are higher or lower on each scale, and where each player sets his minimums and maximums relative to those things.

What would disprove this would be a solid example in which players considered two options to be grossly reversed in value on the same scale. If someone were to say that having Luke turn off his targeting computer in attacking the Death Star is a 12N where having him defeat it with the computer on would be a 2N, and someone else in the same situation would say that having him turn off the computer was actually a 2N and a 12N would be achieved by leaving it on (or similarly for G and S variables, in the same moment of play in the same game), that would mean that the judgment of the values of such choices was too subjective for the vectors themselves to be absolute (across the group).

The notion that one player would consider one choice slightly higher than another and another player would reverse them is not in my opinion sufficient evidence, as such nuances of difference would only indicate the difficulty of measurement.

--M. J. Young

Ron Edwards

Hi there,

GNS is currently not represented by any graphics at all, no, not even a triangle. This has two consequences ...

1. I am not at all sure what Mike's construction is an alternative to.

2. Nor am I 100% sure just what that construction is any more. As first presented, it was a bit of a change, at least in explanatory terms. As currently stated, I guess I'm seeing plain old [Exploration[GNS]] as I've been stating all along, just with "Sim" being extended up/back into the Explore box.

Best,
Ron

Valamir

I lost any connection I saw in this thread back on page two or thereabouts.

I'm not even certain what's being discussed at this point.

What started as an idea to pry Sim loose from the trinity and make it an overlay on top of G or N has morphed into something I don't even recognize.

At this point, IMO, this thread needs to close down and perhaps in a few days when the discussion has been digested by the participants a new one that cuts to the chase with some better clarity can begin..