Design Considerations Over “The Strange Forgeries Of Mr. S.C.Rheber”

The Strange Forgeries Of  Mr. S.C.Rheber is a game of Induction Logic, and for many people it might be the first time that they read about it: one player, the Game Master (later identified as GM), invents any rule/s of a game and the other players (later identified as PCs) they have to try not to loose against each other while figuring out the concealed rule/s. A typical example is the game of Penultima which is a traditional chess match where a third player (the GM) had secretly decided how certain pieces move.

Accordingly to Boardgamegeek.com, which accounts the biggest database of the industry, as of November 2022, the mechanism of Induction appears only 38 times out of 321841 labeled mechanisms of the +100000 games registered in the database. Is one of the so-called rare mechanism. In general board games’ mechanisms are not uniformly distributed: the chart below shows that approximately 83% of the existing mechanisms are only the 19.9% of the total labeled mechanisms.

The bias seems to favor of Dice Rolling (count of labeled games: 36440), Hand Management (19504), Roll / Spin and Move (16836), Variable Player Powers (14821), Set Collection (14735), Open Drafting (10274), Cooperative Game (9977), Simulation (9358), Hexagon Grid (9223).

One of the goal of The Strange Forgeries Of  Mr. S.C.Rheber is to innovate on the long tail of this ranking: I picked Induction (38) and I did so by breaking certain “staples” of the genre, namely:

1. In other Games Of Induction Logic, the GM is outside of the game, acting as a moderator: they cannot loose, but they cannot win either if not in being the enabler of an interesting logic puzzle;
2. Furthermore, it is assumed that is the GM responsibility to make the process of discovering the concealed rules as rewarding and as pleasant as possible for the other players, but nothing either provide a consistent feedback if this was actually achieved. And, more importantly, nothing blocks a GM from creating a riddle which is too complex or too ambiguous for their audience, which means that there isn’t anything within the game itself that consistently measure fairness.

With (1) and (2) in mind we decided that the The Strange Forgeries Of  Mr. S.C.Rheber would have been characterized, respectively by:

1. Everyone plays against each other, including the GM. The GM can win or loose, and, conversely, “PCs can defeat a GM“;
2. A GM can invent any rule what-so-ever, because the way the GM scores would have been inferred from the quality of the secret rule that they had produced. In other words: ambiguous or complex rules would have needed to be sub-optimal strategies for the GM to pursue and, conversely, elegant and innovative riddles would have been the best choices  to improve their score.

In terms of design integrating the last (1) and (2) is not at all trivial: what forces a player to not invent rules that will make them win? If we establish, as for instance, two different currencies (say carrots and potatoes) on the basis of which the GM and the PCs, respectively, would have score with, then we would have solved (2), because across multiple games the performance of each GM would have been comparable to the performance of other GMs on the basis of how many carrots they gained, but would have been incomparable with the performance of the PCs which are measured in potatoes leaving (1) unsolved. Hence, making potatoes and carrots a fungible currency (i.e. creating an exchange rate to turn potatoes into carrots), would have solved (1), but would have been a critical problem to (2). Which brings me to the biggest design challange we had: how can we turn potatoes into carrots and vice versa and balance out a god-like rule-forming player?

The solution came with the ideation of the game phase SCAM-OR-AUCTION which is a traditional guessing round (in the style of other Games Of Induction Logic) followed by a less traditional one, where PCs would have attempt to guess after having peeked at written documentation of the concealed rule. Playtests confirmed that is extremely difficult for a GM to write this kind of documentation such as enables PCs to correctly behave in front of some difficult task like the classification of free-hand drawings. But the GM role became very interesting and fun: some players that participated to our playtest sessions compared it to the construction of Knowledge Based System or the process of writing a patent and carefully choosing words and a set of finite pictures to describe an array of infinite embodiments of the same invention. And where there is difficulty and reward there is also mastering; hence, was exactly based on the classification of free-hand drawings that the game was optimized during development. 

This is when I start to introduce another under-represented mechanism:  Drawing (50). Games Of Drawings are typically party games and their popularity is probably undermined by a “staple” in the genre: a recurrent bias for the players which is the most skilled in visual arts. Abstract Strategy Games like Games Of Induction are, on the other side, recognized as bias-less and, due to this aura of being “fair games“,  they tend to be favored.

But many logic tasks are actually drawing puzzles, like the Standard Progressive Matrices in your classic IQ Test, so a Game Of Logic and Drawing can very well exist:

But Raven’s Matrices are not the only kind of cognitive assessment tests that make use of images. In fact The Strange Forgeries Of Mr. S.C. Rheber is based on the work of Russian computer scientist Mikhail Moiseevich Bongard (Михаил Моисеевич Бонгард, 1924–1971), the father of Bongard Problems. Each problem consists of 12 small boxes: six boxes on the left and six on the right. Each of the six boxes on the left conforms to a certain rule. Each box on the right contradicts this rule. Your task, of course, is to figure out the rule. here is an easy example, Bongard Problem number 6:

But Bongard, as the great riddle maker he was, invented incredibly clever ones, like number 32:

The answer to number 32 is incredibly simple and all the essential information are comprised in only 12 images. I stayed 2 days anchoring on the wrong pattern before seeing the “true pattern” and solving it in 2 seconds. The same happened to some friends to whom I proposed the riddle. But others manage to get the solution on the fly in 10-20 minutes of careful examination. Examples like this (or the ones in Justpuzzles blog) provokes reflections about the criteria that makes a compelling rule:

1. a “simple”, yet uncommon pattern;
2. similarity of said pattern to more common patterns, which allows the riddle maker to leverage on the anchoring bias of the majority of the observers;
3. therefore a high variance in the response time between different observers: which, in my case, went from 5 minutes to 2 days.

I use the word “simple” in (1) as the opposite term for “complex” and I define a “complex pattern” as “a permutation of common patterns in an uncommon logic tree“. As for instance, using the same common patterns introduced in Bongard Problem 6 (above), we can create the following complex rule: 

“the drawings contains only triangles or only an odd number of tetragons or whenever the drawing comprises an even number of tetragons or no tetragons, then, if the drawing comprises also an odd number of triangles is still fulfilling the rule”

We do not want to allow players to invent those kind of rules and (1), (2), (3) give us interesting tools to learn how to avoid this; in fact Complex rules like the one in the example do not fulfill (3). Unlike the anchoring bias of (2), which requires a task-independent skill to be avoided (inhibitory control as for instance), a riddle maker that makes a complex rule leverages on the cognitive load of the solver to conceal the solution of their own invented riddle. However cognitive load cannot be reduced on a task-independent level: it requires prolonged task-specific training (and training is not the fun part of a game) and, therefore, the response time for a complex riddle might be very high in average, but definitely not significantly different from one untrained individual to another untrained individual. This is because the cognitive assets that allow a player to solve a specific complex puzzle (like chess), might not be the same cognitive assets required to solve a second complex puzzle (like Hnefatafl or any other chess-like game). And in a Game Of Induction those puzzles needs to change every time because the designer do not write them in the rulebook, but is up to the players to come up with something. So, in conclusion, 

• a game which features riddles-forming rules that are always changing (like in a Game Of Induction) enables complex riddles which do not fulfill (3) and
• integrating (3) into the scoring logic is an integral part to make riddle-forming rules that are able to punish the creation of complex riddles.

So we need riddle-forming rules such as they reward a riddle maker every time they are credited of a riddle that fulfils (3), because the riddle is credited that it features (1) and (2), which, ultimately, means that the riddle is credited to be solved by at least one solver.

Of course “being credited that at least one solver will solve your riddle” doesn’t imply that “a solver had solved your riddle”: in fact the phase SCAM-OR-AUCTION was designed around this subtle difference. The design features the PCs (which are the solvers of the riddle) involved in a Prisoner Dilemma where they can decide to cooperate with each other or to betray by cooperating with the GM (which is the riddle maker). The scoring logic is, therefore, optimized such as:

• if “a PC solves the riddle”, then it will betray. Prisoners Dilemmas work best if there is uncertainty in the decisions of the other Prisoners/players, so every GM’s rule which features a (3) high variance in the response time between different PCs will “be credited such as at least one PC will solve it” even if the crediting player is not that PC. So rules which features (1) or (2) are the best options for the GM to fool the prisoners;
• The more PCs betrays the more the GM scores. So is in the interest of the GM that the existence of a player solving the riddle is only speculative for the purpose of having the highest amount of PCs betraying each other.

For more information about the final design of the game phase SCAM-OR-AUCTION the reader can refer to the official rules: here.

Now that I had tried to explain my very personal solution to create rule-forming rules in The Strange Forgeries Of Mr. S.C.Rheber I need to explain my interpretation of the reasons why players had so much fun in inventing and testing the rules of their friends during playtests. I am convinced that inventing, teaching and learning rules has deep connections with the instinct of play itself and, maybe, with being human.

One can see any game as if it is made by 2 components: Rules and Moves. As for instance “the knight moves like a 3×2 L and takes pieces in the place it lands ignoring anything on its path” is a rule and “black knight moves from b8 to a6”, “white knight moves from b1 to c3” are some of the moves allowed by that rule. Rules are general and Moves are particular cases of that rule: so Rules are for saving time. With one rule you can explain many moves.

As trivial as it sounds, the ability to infer thousands of different situations (like Moves) from a single sentence (like a Rule) is quite rare in the animal kingdom and it was observed in Homo Sapiens, as well as in a Bonobo Chimp (Pan Paniscus) by the name of Kanzi which, unlike its friends Sherman and Austin (Pan Troglodytes), was able to acquire this kind of ability during Sue Savage-Rumbaugh studies in the late seventies. This ability is known under many names: language, symbolization (check out Terrence Deacon, 1997 The Symbolic Species for one of the most beautiful dissertation in that regard), and, lastly, inductive thinking.

Learning a game and/or creating rules for it, is something hardcoded in our DNA and is much closer to language and symbolization than the common man likes to think: as for instance there are nearly as many American young adults who thinks that video games are a waste of time (26%) than young adults who believe the opposite (24%)(Pew Research Center Survey, June 10-July 12, 2015, n=2001).

All play is human and I even believe that humanity is play, but is evident that this is a sentence which hardly finds agreement at the time I write this article. 

So, I decided to create a game who exploits the most human part of gaming:

a game about inventing, teaching, and learning rules.

During a classic game one player (the game Teacher) will teach the other players each of the rules (which are Public Rules due to the fact of being taught): they will go from the general rule to the particular move by showing examples (a process called Deductive Reasoning). If a move is not allowed then the Teacher will explain in details why the move is not allowed: this behavior of the teacher is known as Explanatory Feedback. But during a game of induction the following 3 differences arises:

1. • Some rules are not explained by the game Teacher (Concealed Rules as opposed to Public Rules).
   • Players will need to infer the concealed rules going from the particular moves to the general rules, by designing experiments and interpreting results (Inductive Reasoning as opposed to Deductive Reasoning).
   • The “results” are provided by the Teacher which is not allowed to give Explanatory Feedback when a move is not allowed by a concealed rule, but only to inform the player the degree in which the move is or is not allowed (Conformational Feedback as opposed to Explanatory Feedback).

These 3 features are a great tool to train creativity in young teens, especially in the context of the mainstream schooling system which is based on explanatory feedback and doesn’t reward individuals when they find a solution by their own: as for instance Abbott’s Eleusis card game is a game of induction which became famous for being employed as an educational tool to prime children into scientific heuristics.

But then why The Strange Forgeries Of Mr. S.C.Rheber?

Other induction games focus on the activity of learning rules by Inductive Reasoning, but do not typically implement a rule-enforced system to reward/punish good/bad Teachers. Either the activity of inventing a rule is allowed to be as large as possible (like in the game Penultima) or the space of possible rules is constrained (as it happens in the lovely Zendo): we don’t mind the necessity of having a good Gamemaster or to be able to select our rules among the subset of rules printed on cards, but we thought that those games exist already and, on the other hand, there was no game that could bring a good challenge to the player that is in charge of inventing the rules. A game Teacher is the enabler of fun, because it needs to be emphatic with their friends trying to be sure that everyone has the same understanding of the game systems and mechanism. Reading rules and explaining them is a passion that leads some of them to create great games at a certain point:

every game teacher is a game designer at heart.

We created the role of the Gallerist, the riddle-maker and the GM of the game, to bring a challenge to those players: if they underestimate their players with a riddle which is too easy, they will be punished, if they try to trick them with excessive complexity, then the rules allow the Collectors, the solvers and PCs of the game, to punish them. If the Gallerist write bad rules and/or provide inconsistent examples, they will be punished by the Collectors.

So The Strange Forgeries Of  Mr. S.C.Rheber started as a game like any other game of Induction: a game about learning concealed rules, but it become a game where inventing rules and teaching them via Conformational Feedback is part of the scoring system itself:

we wanted to design a game made to design games within.