Mika R.
Finland
Unspecified
flag msg tools
badge
Avatar
mbmbmbmbmb
Scott Nicholson has proposed a neat categorization for board game experiences called SNAKS. It describes five archetypes of game experiences: Strategy, Narrative, Arcade, Knowledge and Social. The categories seem logical and well founded to desribe the personalities or traits that the games typically contain. You can see a lecture of the archetypes here:

http://www.youtube.com/watch?v=ak01Jaa1qNs

And here's the original thread for the archetypes:
http://www.boardgamegeek.com/thread/415440/page/1

Now that the archetypes have been proposed, next question is whether these archetypes can be identified from the concrete characteristics that are part of the game design. I.e. is it possible to define an attribute space that would separate games into clusters that represent the stereotypical experiences and position games around these clusters.

The discussion at Scott's original thread was derailed (sorry about that) so I decided finally to move it into its own thread. The status is that we (Tall_Walt, webnard, fellonmyhead and me) have had a lot of discussion about principal game characteristics that would be orthogonal, descriptive and condensed enough to describe the multi-faceted 'personality' of the game.


So far we have been able to identify five main features that describe the board game. Some of them may be debatable, but this is the best we have for now:


======LATEST VERSION===================================


PERCEPTUAL ELABORATENESS:
very low |--------| nearing a book or other aesthetic work (painting, miniature world etc.)

RULES ELABORATENESS:
very low |--------| complex, nearing a simulator

COMPONENT GRANULARITY (SYMBOLIC ELABORATENESS):
very low |--------| very high (has various component types (symbols) that need to be considered in the game)

PLAYER INTERACTION:
very low |--------| very high (direct confrontation)

PLAYER CONTROL:
random |--------| no randomness, open information




Perceptual elaborateness describes game's attention to descriptive details, i.e. elaborate text descriptions, detailed miniaturization, rich playing board, costumes, masks etc. The richness of aesthetics.

Rules elaborateness describes the amount of rules and statements that need to be followed during the game. The more there are individual rules, the more complex the game is to run.

Component granularity describes the amount of symbolic elements (components) that are needed to represent the game state. Players or game mechanism interact with the components according to component rules. Two card games having same number of cards can have different component granularity if one game uses more symbols. A large-scale war game with many types of units and unit attributes has high component granularity. Rock-Paper-Scissors has low granularity.

Player control describes how clearly players can perceive the factors affecting to their available choices and predict the outcome of their actions. Chess has very high player control. Roll and Score Dice Derby Game has very low.

Player interaction describes how much player actions have influence to other players situation in the game. Games with high player interaction are games with direct player elimination element or requires tightly collaborating actions to achieve common goals.

=======================================================


I am sure that some of the features could be easily divided into more detailed sets, but the point here is try to identify absolutely reasonably minimum set of attributes that would give a well-rounded characterization of games (mainly board games).

Ken Shogren (Audacon) has suggested an interesting and even more concise set of principal game characteristics in his later post. He introduces only four features: mechanics control/mechanics interaction, player control/player interaction. The discussion is still ongoing about this model, you can read the alternative model and participate to the discussion.

Have we missed some fundamental attributes that cannot be embedded into ones described here? What do you think of these axis, are they descriptive enough to separate the archetypal experiences? Thoughts? Musings?

A post from Ken Shogren suggested that the attributes introduced here have been used mainly in the same form in his Game Classification System back in the day. So we give homage to the prior art. For more info, check out this thread and find Game Type Indicator section:
http://www.boardgamegeek.com/thread/51705


====HISTORY: The initial list of characteristics (1st version, Jul 2 2009) ======

Game composition describes game's physical appearance
World description describes game's relationship with theme
Player control describes game's relationship with the player
Social interaction describes game's relationship to inter-player communication


6 
 Thumb up
2.01
 tip
 Hide
  • [+] Dice rolls
Ian Wakeham
United Kingdom
Chester
Cheshire
flag msg tools
badge
Avatar
mbmbmbmbmb
Completely off the top of my head (and unrelated to your question) - as I'm at work and don't have time to read back over the original thread at the moment but - radar charts.

I saw that you were looking at some way of displaying the attributes of a boardgame. Currently you have four features which could conveniently fit on the x and y axes of a graph - or the sliders of a graphic equalizer. That's fine if you don't identify any futher attributes, but using a radar (or spider web) chart you could accommodate further features.

The difficulty with the attributes, as you suggest, is that there are shades of grey; such features will be subjective. But using polls or ratings for each feature (similar to game weight) you could produce radar charts for each game in the BGG database - well, at least the top 100 or 500. For example - on a scale of 0 to 9 (or 1 to 10) do you find Chess to be abstract or dripping with theme? Games with similar charts should theoretically have similar play experiences.

Maybe from these charts you could identify how these similarities fit into Scott's idea of the SNAKS archetypes? I guess the question at the moment is determining whether fewer or more than four attributes will assist in this.
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Mika R.
Finland
Unspecified
flag msg tools
badge
Avatar
mbmbmbmbmb
batdog wrote:
Completely off the top of my head (and unrelated to your question) - as I'm at work and don't have time to read back over the original thread at the moment but - radar charts.

I saw that you were looking at some way of displaying the attributes of a boardgame. Currently you have four features which could conveniently fit on the x and y axes of a graph - or the sliders of a graphic equalizer. That's fine if you don't identify any futher attributes, but using a radar (or spider web) chart you could accommodate further features.

The difficulty with the attributes, as you suggest, is that there are shades of grey; such features will be subjective. But using polls or ratings for each feature (similar to game weight) you could produce radar charts for each game in the BGG database - well, at least the top 100 or 500. For example - on a scale of 0 to 9 (or 1 to 10) do you find Chess to be abstract or dripping with theme? Games with similar charts should theoretically have similar play experiences.

Maybe from these charts you could identify how these similarities fit into Scott's idea of the SNAKS archetypes? I guess the question at the moment is determining whether fewer or more than four attributes will assist in this.


You are right, radar chart could plot multiple features nicely. With proper alignment the center would contain abstract, mechanical, full control games with minimal player interaction. The game "universe" would explode towards the borders of the chart with more elaborate/specific game designs.

Thanks for the tip!
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Maarten D. de Jong
Netherlands
Zaandam
flag msg tools
badge
Avatar
Forgive me for asking the following question, but what is the characterisation supposed to accomplish?
3 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Mika R.
Finland
Unspecified
flag msg tools
badge
Avatar
mbmbmbmbmb
cymric wrote:
Forgive me for asking the following question, but what is the characterisation supposed to accomplish?


Optimistically an attribute space that would separate games into clusters that represent the stereotypical experiences and position games with respect to these clusters.
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Maarten D. de Jong
Netherlands
Zaandam
flag msg tools
badge
Avatar
Very well. Do you already have objective criteria to rank games according to various attributes associated with each cluster? Or, phrased more accurately, criteria which differ in a meaningful and statistically significant way in sofar this is possible with what I presume to be ordinal data? Without such criteria, the whole thing seems a bit pointless to me. Fun, no doubt, interesting, perhaps, but ultimately a case of majority rule with too much variation to make the resulting classification worth anything beyond what people would already instinctively pick for a given title.

For what it's worth, a long time ago one of the BGG Admins did some clustering analysis on BGG's users, where the number of clusters was an adjustable parameter in the calculations. That was an analysis of users and not of games, but the principles are of course very similar.
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Mika R.
Finland
Unspecified
flag msg tools
badge
Avatar
mbmbmbmbmb
cymric wrote:
Very well. Do you already have objective criteria to rank games according to various attributes associated with each cluster? Or, phrased more accurately, criteria which differ in a meaningful and statistically significant way in sofar this is possible with what I presume to be ordinal data? Without such criteria, the whole thing seems a bit pointless to me. Fun, no doubt, interesting, perhaps, but ultimately a case of majority rule with too much variation to make the resulting classification worth anything beyond what people would already instinctively pick for a given title.


So far we have the extents of attributes, not the levels in between. However, if there are extremes, there must also be ordinal steps to reach them.

What is the level of granularity and its subjectivity are questions which are hopefully addressed in this thread. For example, we may attempt to find quantifiable measures for social interaction elements in games and calculate them, but another way to approach it would be to introduce pairwise ranking through questionnaires and thus end up with final ordinality between games. Certainly subjectivity affects to the local ordering but by averaging over group of gamers a normative ordinality could emerge.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Ken Shogren
United States
Rochester
Michigan
flag msg tools
badge
Avatar
mbmbmb
I like the thoughts behind your idea here and it's something I have worked on in the past. Here's an ancient article I wrote on the subject using a 4 axis system with a comparison to Myers-Briggs type-casting. I've updated some of my ideas on this since I wrote this years ago, but maybe it has something you'd find interesting...

Game Classification System
1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
A Derk appears from the mists...
United States
Portland
Oregon
flag msg tools
badge
Avatar
oooooOOoooo. Big words.

I find it helps to position the board in the middle of the table.

goo
15 
 Thumb up
0.05
 tip
 Hide
  • [+] Dice rolls
Mika R.
Finland
Unspecified
flag msg tools
badge
Avatar
mbmbmbmbmb
Audacon wrote:
I like the thoughts behind your idea here and it's something I have worked on in the past. Here's an ancient article I wrote on the subject using a 4 axis system with a comparison to Myers-Briggs type-casting. I've updated some of my ideas on this since I wrote this years ago, but maybe it has something you'd find interesting...

Game Classification System


"Abstract -- Concept -- Themed
Player -- Contact -- Group
Skill -- Control-- Chance
Deep -- Complexity -- Light"

surprise

Two independent thought processes ending up almost to the same conclusion. What would the likelihood of that if results were arbitrary?! I say we are onto something here.

The main difference is in the complexity attribute. I feel that the information is integrated into the Game Composition and Player Control features and is an emergent characteristic out of the two principal ones, similarly to Theme. So my thesis is that the perceived theme/story and complexity (as in 'game depth') are emergent characteristics, not principal.

Did you have time to experiment at all with your system after you had developed it? According to your OP, you pretty much left it at that. I agree that these kind of systems would eventually lead to a great recommendation system for games, if applied in practice, but the effort to accomplish it might be huge.

Perhaps principal component characterization could be another suggestion for collective geekmodding effort, similar to image labeling?

 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Mika R.
Finland
Unspecified
flag msg tools
badge
Avatar
mbmbmbmbmb
derk wrote:
oooooOOoooo. Big words.

I find it helps to position the board in the middle of the table.

goo


What if you have multiple different sized boards? Do you consider them as a single board and use center of gravity?

1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Ken Shogren
United States
Rochester
Michigan
flag msg tools
badge
Avatar
mbmbmb
eikka wrote:
Two independent thought processes ending up almost to the same conclusion. What would the likelihood of that if results were arbitrary?! I say we are onto something here.


I think so.

eikka wrote:

The main difference is in the complexity attribute. I feel that the information is integrated into the Game Composition and Player Control features and is an emergent characteristic out of the two principal ones, similarly to Theme. So my thesis is that the perceived theme/story and complexity (as in 'game depth') are emergent characteristics, not principal.


The concept (abstract / theme axis) and the complexity (deep / light axis) axis were problematic. You've avoid the complexity one and your comment was what I basically determined the hard way - complexity is more emergent than principle. You still have the abstract - theme axis, so see below....


eikka wrote:

Did you have time to experiment at all with your system after you had developed it? According to your OP, you pretty much left it at that. I agree that these kind of systems would eventually lead to a great recommendation system for games, if applied in practice, but the effort to accomplish it might be huge.


I spent a good portion of time working on classifying my personal collection. Some of my ratings still contain the GTI code I assigned. A lot of my collection hasn't been updated on BGG because of the ongoing work in this area. So in other words, I basically stopped because the Concept and Complexity axis really emerged as troubling. It was/is very unclear just how to assign values consistently (even for a single reviewer/classifier). Contact (Player vs Group) and Control (Luck vs Skill) were much easier. So I began to look heavily at what concept (abstract vs themed) meant as well as complexity - at least how I saw it.

What occurred to me was that what I was really after was a better understanding of mechanics/mechanisms in games. So I shifted to some analysis of the existing database items. I started with BGG's defined mechanics/categories and with the actual game data in the database. I wrote a number of K-nearest neighbor auto-clustering routines on the data and found that some clusters were fantastic. But the majority of the time I found that the data in BGG for these game attributes was basically junk. Inconsistent and often outright wrong and most frequently just missing. The BGG mechanics taxonomy is ill defined at best. So what I have been working on for 2 years (on and off - this is a hobby after all!) is a better way of descriptively classifying mechanisms in games. There is still tons of work that I need to do.


eikka wrote:

Perhaps principal component characterization could be another suggestion for collective geekmodding effort, similar to image labeling?


I think geekmodding is a way of avoiding the hard work of defining an adequate taxonomy. Let the public consensus determine it. In reality a good clustering routine will take care of that with good descriptive attribute vectors, but that requires a good taxonomy for ascribing attributes. So you end up full circle with a need to do the hard work anyway. What is likely needed is a connected attribute set in which branches emerge from a basic concept and increase in detail towards the ends but can still connect to other basic concepts as well. For example, a mechanism of auction is a good base item, but each type of specific auction (English, Dutch, silent, etc) would need to flow out from the root. Anyhow, I'm not sold on the idea that games can or should be classified within a hierarchical structure (which is where many classifiers end up pushing things.) Personally, I'd like to see something more heterarchical in nature.

Anyhow, I've babbled enough for now on my thoughts.

So...A question for you. How do you envision the World Description attribute to work? Are the ends of the scale truly exclusive of each other? In board games, this may be true(r), but in RPGs, it may not. I think whatever system is developed ultimately ought to work for both - not to mention video games and even other types of play. What I'm getting at is that I think you'll find yourself in a similar situation looking at your Mechanical end of scale and thinking - 'this is really important stuff here, but I don't know exactly what to do with it." I'm very curious to see what you do with it though!


 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Mika R.
Finland
Unspecified
flag msg tools
badge
Avatar
mbmbmbmbmb
Audacon wrote:

think geekmodding is a way of avoiding the hard work of defining an adequate taxonomy. Let the public consensus determine it. In reality a good clustering routine will take care of that with good descriptive attribute vectors, but that requires a good taxonomy for ascribing attributes. So you end up full circle with a need to do the hard work anyway.


Yes, creating a taxonomy for the attribute values might turn out too complex process.. I wonder how a pairwise ranking would work. Users could rank game pairs in their collection by selecting which game is more abstract, which is more interactive etc.
Then, out of these pairwise rankings, ordinality would emerge. This way degrees of things would not be needed to be defined.


Audacon wrote:

So...A question for you. How do you envision the World Description attribute to work? Are the ends of the scale truly exclusive of each other? In board games, this may be true(r), but in RPGs, it may not.
I think whatever system is developed ultimately ought to work for both - not to mention video games and even other types of play.


World Description was not meant to be exclusive but rather (Textual weight - Mechanical weight), i.e. a combination of the two. They could be separated for complete independence but I think people only perceive the bias towards one, so if text and mechanics are used in balance, the value should be in the middle. Even as the absolute values are lost in combination, abstract-detailed attribute would reveal the real weights for the components: For example if we know that a game uses balanced amounts of text and mechanics, we can see where the game lies on abstract-detailed scale to understand how much they are being used.


Audacon wrote:

What I'm getting at is that I think you'll find yourself in a similar situation looking at your Mechanical end of scale and thinking - 'this is really important stuff here, but I don't know exactly what to do with it." I'm very curious to see what you do with it though!


Yep, I am not that far in understanding what I am doing
Maybe only pushing ideas forward for future use, if nothing more.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Mika R.
Finland
Unspecified
flag msg tools
badge
Avatar
mbmbmbmbmb
Webnard wrote:
eikka wrote:

Axis should illustrate different characteristics:

Game composition describes game's physical appearance
World description describes game's relationship with theme
Player control describes game's relationship with the player
Social interaction describes game's relationship to inter-player communication


First, thanks for making an independent thread.

Second, what are the labels describing the opposite poles of the axes named above?


You're welcome!

Here are the poles:

Abstract--Game composition--Themed by details
Mechanical--World description--Textual/Linguistic
Full control of the environment--Player control--Random/uncontrolled environment
Independent--Social interaction--Interactive
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Ian Wakeham
United Kingdom
Chester
Cheshire
flag msg tools
badge
Avatar
mbmbmbmbmb
Geek users are already in the habit of rating each game in some form or another: their overall rating, the weight of the game, the language dependence of the game, etc, etc. So, it's not too much of a stretch to get users to rate each of the four game attributes or features for each game. Whether this is a scale of 1 to 10, -4 to +4 or some sort of descriptive scale behind these numbers would need to be established.

In suggesting the use of a radar chart (or some other similar graphic representation) to display the data I was envisaging that each game entry (or a least those that had sufficient ratings in each feature - 30?) would have its own constantly evolving radar chart displayed on its page. A quck glimpse at a chart for one game might show a similarity with another and hence a general idea of whether this would be a game I would enjoy or not. It may also give interesting correlations between designers, publishers, etc. Mac Gerdts may have a signature chart and it would be simple to identify other designers or games with the same "signature".

I just wonder whether four features would be enough for this sort of graphic representation of a game though?
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Mika R.
Finland
Unspecified
flag msg tools
badge
Avatar
mbmbmbmbmb
Webnard wrote:


I. Abstract--Game composition--Themed by details
-4...-3...-2...-1....0...+1...+2...+3...+4

II. Mechanical--World description--Textual/Linguistic
-4...-3...-2...-1....0...+1...+2...+3...+4

III. Player Full--Environmental control--Random/uncontrolled
-4...-3...-2...-1....0...+1...+2...+3...+4

IV. Independent--Social interaction--Interactive
-4...-3...-2...-1....0...+1...+2...+3...+4


What games go where? Where does Chess fall in each axis?
Dominion? Arkham Horror? Agricola? Advanced Squad Leader? Combat Commander: Europe?

I think we're at a place now to try this thing out. How would you fill in these stats?


| I | II | III | IV |
------------+-----+-----+-----+-----+
CHESS | -4 | -4 | -4 | ? |
------------+-----+-----+-----+-----+
DOMINION | | | | |
------------+-----+-----+-----+-----+
ARKHAM HORRO| | | | |
------------+-----+-----+-----+-----+
AGRICOLA | | | | |
------------+-----+-----+-----+-----+
ASL | | | | |
------------+-----+-----+-----+-----+
COMBAT COMMA| | | | |
------------+-----+-----+-----+-----+


Later, once the analytic is improved and more stable, it will be interesting to compare the games of individual designers. No one will be surprised if the wonderful rondel games of Mac Gerdts have similar numbers. I wonder whether other designers will be as similar? And whose designes will be most different?

We're closer, I think, but not quite at that point.


Yes, it is difficult to objectively name chess an objective number for social interaction, it is quite interactive (all pawns will be killed, complete player elimination etc.), but I can't tell you objectively whehter it is +4, +3 or +2 for example, I'd say it is probably around +3.

Regarding Agricola, I don't have much experience with it yet although I have the game waiting for plays, but here goes some values (out of the hat):

I: +2 or +3
II: -1 maybe.. may change after I get to feel of the play myself
III:-2
IV: -1 or -2

These may be completely different than your assessment with the game if you have played it a lot. In my earlier post I said that making an objective taxonomy for the values in the axis may be impossible because personal assesment is related to the person's other game experiences.

There are few points I have noticed after thinking the valuation of the games you provided:

1) This classification might allow separation of archetypes and to position games around these archetypes, but for creating complete detailed clusterization of specific design styles? Maybe not. This is because the attributes abstract the personality behind the design choices, for example a rondel mechanic may not be perceived much different than simultaneous role selection mechanic..

2) The the more difficult point. The absolute values and taxonomy for these values is rather difficult to establish quantitatively. So in earlier post I have talked about ranking that would position games between other games and thus create ordinality without scales. If the positioning of the games would be made by large group of users, it would produce the scale of ordinality as emergent property..

So, how would that go then?

========

THE PROCESS FOR GAME CHARACTERIZATION

BGG'ers would go through their game collection for each axis and they would position their games by creating ranking for that axis (no values, just mutual ordinality for each game against each other). So, they only identify rank order of their games for that axis. Example: I would create a top-down list of my games based on whether they are very abstract or detailed, then I would create a top-down list of games that have the most social interaction etc. Perhaps even a mechanic that allows games to be at the exact same rank could be provided.. Anyway, this would give us the way to assess the scales without the need for number taxonomy.

After suffciently large number of people have done it, the positioning of the games can be turned into global rank order list for the games for each axis. Mutual ranks would gain weights based on how many people have voted certain games above other games so the community created ordinality 'votes' would eventually create a certain scale.

Finally, all the statisticians could use the data with their clustering algorithms trying to identify classes etc. Someone could even build a recommendation system that retrieves the k nearest neighbours from the feature space etc. A range of applications could be developed out of this.

===================

This is the process I have imagined we could use for the model realistically.

1 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Mika R.
Finland
Unspecified
flag msg tools
badge
Avatar
mbmbmbmbmb
Above post edited to create a better page layout. BGG didn't seem to like this style of text in it....
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Mika R.
Finland
Unspecified
flag msg tools
badge
Avatar
mbmbmbmbmb
Webnard wrote:
And it appears a useful classification cannot be reached with opposing poles to each axis. So maybe we can throw away that idea, too. It is good to know what works and what doesn't. And I would be satisfied if this is as far as we could get at this point, but I believe we can still move forward.


Hmmm, I do understand that you see the opposing poles less clear way of describing the attributes than using isolated countable, concrete elements in the game design, but I don't understand how did you manage to include textual with your Mechanics granularity then? You wrote:

Quote:
Mechanics Granularity: A lower number would be more purely mechanical. A higher number addresses linguistic necessity of a game, whether reading flavor text or negotiating between players. Includes rules density.


Now, if we take adventure book style games (such as Tales of the Arabian Nights for example), they are low in mechanics but high in text, so in your scale they would have values close to 9. When contrasted to your example valuation of games:

Quote:
Mechanics Granularity
1...2...3...4...5...6...7...8...9
CS! AH ASL Go/Chess


we would then have Go/Chess/ASL + Adventure books all located near the maximum. Was that your original purpose? Maybe I didn't understand something...

In the World Description feature (my opposite-poles, measuring the balanced use of text and mechanics) we would have Go/Chess on the mechanical side, AH around the middle and Adventure book games on the linguistic side of the spectrum. I feel this would be useful separation for these very different games, don't you?

Webnard wrote:

Building on your idea of ordering the games in relation to one another, I'm going to propose slightly different categories. This is a first proposal, but I return to the concept of "granularity" that you introduced much earlier. I like the concept because it is intuitive and carries no stained baggage from over-use.


Yes, I would prefer use of quantifiable granularity for the axis (being
engineer, I do like things to become tangible and measurable in some way) and I think your definition fits very well. You have found good elaboration for the features that were originally just vague hunches on how things could be characterized!

Webnard wrote:

Component Granularity: the number of types of components, with the number and types of uses.


Yes, doesn't conflict with Game Composition, low count of component types and their uses would equal to Abstract, no?

Webnard wrote:

Mechanics Granularity: Similar to World Description. A lower number would be more purely mechanical. A higher number addresses linguistic necessity of a game, whether reading flavor text or negotiating between players. Includes rules density. Here is where heavy/light, thick/thin are felt. Combined with higher Component Granularities, higher Mechanic Granularities allows story to emerge.


What do you mean by rules density? Lots of small rules means high mechanics granularity, right? Also, lots of flavour text but very few rules leads to high mechanics granularity? I don't quite agree on this as I wrote above using Adventure book games as the example.

Webnard wrote:

Reliability of Player Decision Making: Similar to Environmental Control. In push-your-luck games, reliability is zero. One hopes for the most favorable die roll or card, but there is no certainty. However, in Chess, I am absolutely certain that my decision to advance my pawn one space will result in just that--and nothing else. It may be a mistake to do so, with tragic consequences ten turns later, but in game terms, I am certain that my decision to move my pawn will result in my pawn moving where I want it.


Slightly disagree with this also, in an imaginary ultimate Diplomacy game (with negotiations first and simultaneous action selection next, no random number generators) you have always 100% certainty that your action will lead to certain outcome on the board but due to other people's negotiations with other people, the outcome of your move is dependent on the changes to the game environment that happen due to other player choices. This is chaos and it is similar to random effect in that you cannot ever predict the insane amount of negotiation combos the players will make with every player. But it is still causal in that you can trust that your token will cause a fixed effect on the board. Did you mean to leave it in the Reliability attribute or leave it out on purpose?

An example: I'd say that 4 Player Chess is less reliable (i.e. having less environmental control) game than Chess. But both of them have absolute certainty on what happens after your move. I'd like to keep this characteristic in the attribute as well as it affects on the game experience a lot.

Webnard wrote:

Player Interaction: If I move my pawn wisely, I potentially change my opponent's plan, thereby interacting with his game play. During play, I hope I correctly guess his next decisions, and my goal is to limit those. Of course, it is more likely that I have moved without thinking and my opponent's plans then do not change. Still, the potential is there. If I were a better Chess player, I would change my opponent's plans more frequently. This does not address game chatter, which in this scheme belongs to mechanics granularity because rules govern game chatter.


This seems to be in line with Interactive/Independent poles. Game chatter is not often covered in the rules, but when they do (Diplomacy and Battlestar Galactica: The Board Game), it's definitely part of World description/Mechanics granularity in the linguistic side of things as rules for negotiations are part of "flavour text" for the theme (i.e. diplomatic alliances in Diplomacy and planning collaborative actions in Galactica).

Webnard wrote:

It is because AH has a higher 1 & 2 that story is allowed to emerge. And yes, these are assessments, but I don't know how personal they are. You might place AH higher or lower in the Component Granularity scale.


Number 2 is something I don't understand due to reasons stated above. Mechanics granularity increases with (high text/low mechanics) as well as (low text/high mechanics)? Adventure book games and ASL get the same values?

Webnard wrote:

However, in terms of order, most gamers would agree, I think, that Go belongs where it is in relation to other games, i.e., its order along the scale.


Yes, this is what I proposed above. Go is at an end of a certain axis, Some obscure party game would be at another end of some other axis, and people should be able to relate and order the games they played along these axis so that they wouldn't have to assign numbers to these attributes at all.

Webnard wrote:

Using these classes, I learn I prefer games rated 3.5.7.9, games such as Rommel in the Desert, Imperial, Panzer Leader and the early scenarios of Up Front. It also explains why I am attracted to ASL (9.8.7.9), but haven't enjoyed playing it: the component and mechanic granularities are too high in relation to other games I do like (9.8.7.9).


Again, in your system you wouldn't like any adventure book styled games either because they are also "high in mechanics granularity"?

---

I'll be leaving for a trip tomorrow and won't be able to reply in depth for a week or two, perhaps I'll be able to squeeze in a comment or two though...

But lets keep the ideas flowing. I'd like to thank you for the discussion we have had so far! You elaborated very good descriptions for the four attributes and our thoughts seem to align quite well on this although there are some more things that need to be worked out.

But seems like we are going to the right direction, so lets keep our brains storming! Thanks again, this is interesting stuff.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Mika R.
Finland
Unspecified
flag msg tools
badge
Avatar
mbmbmbmbmb
Webnard wrote:

Does the classification explain everything? No. But I think we're on to something useful by ordering the games, rather than rating them. You tell me, does statistical significance play a part in ordinal scales? I don't know.


Yes it does, order emerging from individual rankings is voting. It could be a good statistical tool for this purpose if we want to restrain from using value scales that may be difficult to assess objectively.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Mika R.
Finland
Unspecified
flag msg tools
badge
Avatar
mbmbmbmbmb
Webnard wrote:
eikka wrote:
...Mechanics granularity increases with (high text/low mechanics) as well as (low text/high mechanics)? Adventure book games and ASL get the same values?
When you return, it would be helpful to understand your resistance to this. You may be on to something or it may be something else.

What are your reasons for not liking both D&D and ASL being 9s in Mechanics Granularity?


I had time for a quick response..

D&D are high in mechanics granularity, definitely. But they are also high in language use. So D&D would be high/high, which leads close to the center in my system but close to maximum in yours.

But I meant adventure book games (see http://en.wikipedia.org/wiki/Choose_Your_Own_Adventure), which are different than D&D. They have low mechanics (and low Game Composition) because you just select a resolution to a story paragraph and move to another page to read consequences. It almost like reading a book. So it is low in mechanics and high in language use. This means it would position close to textual pole in my system but high in mechanics in yours..

So World Description attribute would separate D&D and Choose your adventure style games, but Mechanics Granularity would not separate them. Tales of the Arabian Nights is close to the Choose your adventure games, with added simple move mechanic and a very simple mechanic to develop a character. So the gameplay experience differs greatly from D&D where you may utilize lots of different mechanics. This is not possible with Choose your adventure books.

In summary:
World description can be seen as composite feature from two features so it is a substraction of mechanical granularity from textual granularity. When both are high, result is close to origin. When other component is higher than another, WD starts approaching either pole.

In your proposal, text or mechanic can be high and the value is the same. So games that have high/low text/mechanic and vice versa are not separable with the feature.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Mika R.
Finland
Unspecified
flag msg tools
badge
Avatar
mbmbmbmbmb
Hi again, came back for couple of days..

Webnard wrote:
eikka wrote:
...In your proposal, text or mechanic can be high and the value is the same. So games that have high/low text/mechanic and vice versa are not separable with the feature.
Yes, that is what I am proposing. In practice, I see no difference.

True enough, in some games the reading is done in the rule book, in another, the reading is done while playing the game. Both create the game world through text and both must be read to play the game.


Hmmm.. I still see it differently. Rulebook + flavor text in the game components add to the game experience yes, but I have put them in the world description axis (detailed/abstract). Why? Because descriptive and detailed rulebook, as well as rich flavor text in the game components, is the 'aboutness' of the game. I.e. description of the context that the game takes place. And I feel the 'aboutness' must be separate from the pure mechanics, as it is a form of descriptive "metadata" about the game content.

Webnard wrote:

I am also equating reading the rules with mechanics. The reason is that it must be done. Rules are part of the game and part of the game experience. It can't be avoided.


In my view mechanics are not the same to the textual parts of the game. Mechanics are the 'how' the world works whereas the textual part is 'what' is in that world. They talk different language. You could probably describe a single mechanism textually by elaborating all aspects and consequences of different actions in various situations and making players to choose between the alternative textual descriptions but it would have translated that mechanic into text choice game and often limited the number of choices people can make to affect the world state.

The power of mechanics lies in the fact that, at best, a solid mechanic may generate infinite combination of story elements that would be complementary elaboration (to the textual) of the game's context. A linguistic description describes only a single event or object and there still needs to be some sort of mechanism that would add the interactivity to the textual game elements.

I think that, for example, text adventure games provide much more different experience than mechanically elaborate ASL, for example. And therefore feel the need to separate them.


Webnard wrote:

So, it isn't quite clear to me if you are working with my proposed definition of Mechanics Granularity to reconcile it to yourself or if you are struggling to come up with another way, or both or some third way. I guess we'll wait and see. Enjoy your time off!


I am still processing the way we perceive text/mechanics relationship differently and try to come up with proper reasoning for deciding which would be better representation in the OP. We have had a good dialogue about it so far and it has made me think whether it would be better to split "textual granularity" and "mechanics granularity" into separate axis. But that would probably add redundant information regarding the Game Composition axis, which already describes the "level of elaboration" in the entire game.

There seems to be an interesting documentary coming out related to early text adventure games, i.e. http://en.wikipedia.org/wiki/Interactive_fiction. Here's a trailer for it:


Single player text adventures embedded puzzles into narrative text and provided a simple mechanic where you would have to solve the problems by giving commands in right order using proper set of collected equipment to get past the puzzle. Level of linguistic narrative is high, but after you solve the game, there's not much mechanical reason to get back to it. Those games would emphasize linguistics over mechanics to provide a game experience around exploration. They are more mechanics oriented than make your own adventure books but much less than ASL. I think the distinction between the three is important here since the game experience you get out of them requires different mental processing.

To lighten up at the end, here's something I came across while browsing info about text adventures, funny stuff in a cozy, nerdy kind of way.

 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Mika R.
Finland
Unspecified
flag msg tools
badge
Avatar
mbmbmbmbmb
Webnard wrote:
eikka wrote:
Hi again, came back for couple of days..
Welcome back. I, too am just getting back in town. Unfortunately, I can't join the conversation, but I have a few comments.

I am willing to be persuaded and a quick reading of your most recent post seems persuasive. I hope to make reasonable counter-arguments. However, if I concur or at least fail to find counter-arguments, then something must change in the way I think about this: i.e., the principal characteristics that position boardgames in relation to one another. That's a bit different than the thread's title, but I hope it's more accurate.

I won't be able to read closely your post for a few days. I also want to watch the YouTube contributions as I believe they must add something to the conversation.

I do appreciate your time and patience, and especially your willingness to wait on making final conclusions on this matter. I think, though, that final conclusions are near.

That said, how would you change the earlier assessments of ASL, D&D and Go, etc., to fit what you propose with the textual parts? Another number? Instead of a 4-digit classification, are you proposing a 5-digit one?


So I have been thinking about splitting this:


WORLD DESCRIPTION
Textual/Linguistic
---
|
|
|
---
Mechanical


Into this:

USE OF TEXT: none |-----| nearing a book
USE OF RULES: none |-----| complex, nearing a simulator



EDIT:

Furthermore, if we add this

USE OF GAME COMPONENTS: very low (rock/paper/scissors) |-----| very high (filled with bits)

Then we could get rid of this as well:

GAME COMPOSITION
AND APPEARANCE

Themed by details
---
|
|
|
---
Abstract



So this is how the principal attributes would look like after splitting apart:

USE OF TEXT: very low |-----| nearing a book
USE OF RULES: very low |-----| complex, nearing a simulator
USE OF COMPONENTS: very low |-----| very high (filled with various kinds of bits)
PLAYER INTERACTION: very low |-----| very high (direct confrontation)
CONTROL OVER ENVIRONMENT: random |-----| no randomness, open information


 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Mika R.
Finland
Unspecified
flag msg tools
badge
Avatar
mbmbmbmbmb
Webnard wrote:

I believe we agree on Component Granularity (the number of types of components, with the number and types of uses) and Player Interaction (the potential for one player's decision to change another player's subsequent decision). So, polls could be constructed using those two scales, asking other BBG members to rank games in relation to each other. I might start with asking BGG members which games have the least compenent granularity and which have the most? Then you have the extremes to which all other games relate.


Yes, finding game groups for the extreme ends of the scales would be a good start. It may be difficult to find a consensus, but it could be tried. Go ahead if you want try the poll.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Jeff Khoury
United States
White Oak
Pennsylvania
flag msg tools
Avatar
mbmbmbmbmb
While I allow thst there is plenty of possibility for game theory and categorization, I would strongly argue against any kind of implementation on the geek. I feel that there are more than enough, charts, bars, pictures, ratings, widgets, sprockets, weights, averages, graphs, analysis, links, lists, posts, comments, plays, modules, collections, tags, suggestions, recommendations, forums, doodads, and kerjiggers to satisfy most every practical user.

Imagine how cluttered and chaotic BGG must seem to new/casual users. A good friend of mine is as big into boardgames as I am, but he won't use the geek because there's just too much going on. At some point, useful, laymen information gets buried under statistical analysis . . .

And pictures of cats in game boxes.
 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
Ken Shogren
United States
Rochester
Michigan
flag msg tools
badge
Avatar
mbmbmb

I like your split of World Description, but you might want a slightly different take on the 2 parts. Might I suggest:

Use of Text -> Use of Language : It would take into consideration how much the players must use language skills (including spoken and written as well as metaphor, analogy, allegory, etc) So chess might have a very low use of language, a game like scrabble or a negotiation game might be moderate while a Role-Playing Game might be very high. The moderate games would differ by the degree of player interaction. In scrabble, player interaction is limited to board position, but in a negotiation game it is much higher.

Use of Rules -> Interaction of Mechanics : It would look at how the system of rules is structured. Does it use rules that are intricately interrelated or broader and potentially independent of each other. It is the game system's equivalent to the player interaction scale. So many kids games might tend toward lower interaction levels while most euros tend toward the middle and complex simulation style games (e.g. wargames) might tend to the high side.


Lastly, I agree with the points on components. I think in describing the functioning of a game, it is important to separate the aesthetics from the game system (i.e. the form from the function). Aesthetics is necessarily subjective and as such will be hard to reach consensus on. A game with numerous bits is no more or less complex than a game with few bits and intricate interrelated rules. I think what you are after here is more about how does the theme of the game translate into the rules. A game where the theme is 'pasted on' would not show much relationship between the mechanics and the theme. Where as a simulation game would tend to have tight relationships.

I've been working with this idea and am leaning towards looking at the game's theme details to mechanics ratio as a possible criteria scale.


Keep up the good discussions. Sorry I haven't been posting much, but I have been reading the posts when I can.

 
 Thumb up
 tip
 Hide
  • [+] Dice rolls
1 , 2 , 3  Next »   | 
Front Page | Welcome | Contact | Privacy Policy | Terms of Service | Advertise | Support BGG | Feeds RSS
Geekdo, BoardGameGeek, the Geekdo logo, and the BoardGameGeek logo are trademarks of BoardGameGeek, LLC.