Experimental interpretation of ancient games through playtesting

Experimental interpretation of ancient games through playtesting Angelos Tsilogiannis *1, Markos Konstantakis 1, Spyros Vosinakis 1 1 Aegean University – Department of Product and Systems Design Engineering, Syros, Greece, [email protected];[email protected]; spyr[email protected] *Corresponding author Correspondence: [email protected] ABSTRACT Reconstructing the rules of ancient games remains one of archaeology’s most intricate interpretive challenges. While surviving boards and pieces provide tangible evidence, the logic of play, rules, roles, and goals have largely vanished. Traditional reconstructions rely on text and comparisons to better known games, yet these methods seldom capture the experiential and social dimensions of ancient gameplay. This paper presents an experimental framework designed to explore how such systems might emerge through play itself. Using Ludus Latrunculorum (“The Game of Little Soldiers”) as a case study, participants engaged in iterative, co-creative sessions where they invented, refined, and transmitted rules across successive lineages of play. The resulting games revealed recurring mechanics that closely parallel known reconstructions of Ludus Latrunculorum. While the outcomes do not claim historical accuracy, they demonstrate how collective experimentation can illuminate the cognitive and social processes underlying the development of structured games in antiquity. The study thus proposes experimental play as a complementary archaeological tool for examining how rulesets evolve, stabilize, and reflect the societies that produce them. Keywords: game reconstruction, experimental archaeology, game studies, game rules, ancient games Introduction The study of ancient games occupies a complex intersection of archaeology, anthropology, and game studies. Although board games constitute some of the most enduring material traces of play in human history, their original rules and social meanings are often lost to time. Archaeologists and historians typically rely on fragmentary evidence such as game boards, pieces, or scattered literary references to reconstruct how these games might have been played. This traditional approach has produced valuable descriptive catalogs and comparative typologies (Bell, 1960; Murray, 1952), yet it often struggles to capture the reality of the embodied actions, negotiations, and social meanings that once animated these artifacts. In recent decades, advances in archaeology have highlighted the limitations of purely object-based interpretations. Researchers such as Finkel (2007) and Schädler (1998) have successfully reconstructed partial rule sets for games like The Royal Game of Ur and Ludus Latrunculorum through a combination of textual and material evidence. However, these reconstructions, while plausible, tend to treat rules as fixed historical data rather than emergent social constructs. They often lack insight into how games were learned, transmitted, and adapted by the people who played them. The result is a static understanding of dynamic cultural practices. Understanding ancient games therefore requires situating them within their societal context. Games were not isolated leisure activities; they functioned as social performances embedded in everyday life, education, warfare, and ritual (de Voogt, Dunn-Vaturi, & Eerkens, 2013). The processes of rule negotiation, competition, and shared learning that characterize gameplay mirror broader mechanisms of cultural interaction and transmission. A rule is not merely a prescription for play, it is an agreement between individuals that encodes shared values, hierarchies, and cognitive models of order (Schell, 2019). Without accounting for this contextual and procedural dimension, any reconstruction remains incomplete. This paper proposes an experimental approach that treats play itself as a research method. Rather than attempting to recover definitive rules from limited evidence, we explore how rulesets might emerge through iterative play. In our proposed approach, participants engage with reconstructed boards and pieces modeled on Ludus Latrunculorum and are tasked with inventing and refining their own rules within historically framed scenarios. Across successive “lineages” of gameplay, these rules evolve through teaching, adaptation and negotiation, mirroring the cultural processes through which games likely developed. By documenting this evolution, the study aims to evaluate whether structured play can serve as a viable archaeological method: one that complements traditional analysis by reconstructing not only what ancient games may have looked like, but how they might have been learned, played, and understood within their social worlds. After stating some fundamentals on related work (Section 2), we analyse the framework we designed to conduct the experiment in Section 3. Later, in Section 4, we summarize events and points of note in each lineage. Finally, in Section 5, we discuss the limitations involved as well as future work and of course whether or not the method suggested can help in the reconstruction of ancient games. Related Work Game Reconstruction The reconstruction of ancient games often relies on partial evidence — surviving boards, scattered pieces, or fragmentary written accounts — leaving their actual rules uncertain. Scholars have long attempted to deduce these rules through comparative analysis with better-known games or by interpreting textual clues within historical context. These range from early typologies (Bell, 1960; Murray, 1952) to modern syntheses (Crist, Dunn-Vaturi, & de Voogt, 2016). However, while these reconstructions produce plausible hypotheses, they tend to treat rules as fixed systems awaiting recovery rather than as evolving cultural practices. The approach presented in this paper departs from such reconstructions by proposing a generative method — one that uses iterative play to explore the range of rules that could emerge from a given material configuration. This shifts the focus from reverse-engineering a historical original to understanding how human players make sense of affordances and constraints within the game’s material design. In doing so, it positions play itself as an epistemological tool for archaeological inference. Experimental Archaeology Experimental archaeology is a research method that seeks to understand ancient technologies and practices through direct material reconstruction and experiential testing. As defined by Outram (2008), it aims to replicate ancient processes to examine their practical and cognitive implications. When applied to games, this aligns with studies such as Eerkens, de Voogt, and Duersch (2019), who argue that play itself can reveal how material design influences cultural evolution. Rather than relying solely on textual or visual sources, it focuses on replicating the physical and procedural aspects of past activities to reveal insights about their feasibility, constraints, and intended outcomes. In the context of ancient games, this approach allows scholars to move beyond the static study of artifacts and engage with the dynamics of play itself. Reconstructing a game through repeated experimentation can uncover the implicit logic of its design, its ergonomic or cognitive demands, and the social interactions it facilitated. In this study, the experimental method takes on an iterative, player-centered form where rule-making and playtesting act as archaeological tools. The goal is not to “discover” the authentic rules of a lost game but to approach an understanding of what those rules could have been by observing how structure, balance, and meaning naturally emerge through play. Cultural Transmission and Iterative Learning The transmission of rules across multiple groups of players in this experiment parallels models of iterative learning and cultural evolution in anthropology and cognitive science. This dynamic reflects broader models of cultural evolution and bias-driven transmission described by Mesoudi (2011) and Acerbi and Bentley (2014), where systems stabilize through cumulative social negotiation. Studies in these fields explore how behaviors, symbols, and languages evolve over successive “generations” of learners through cumulative modification (Kirby et al., 2008; Tomasello, 1999). Each lineage of gameplay in this project represents a microcosm of that process: players receive a partially defined set of rules, reinterpret them through interaction, and pass on an altered version to the next group. This results in emergent systems that gradually stabilize through repetition and shared understanding. By applying this framework to the reconstruction of games, the experiment transforms rule-making into a model of cultural adaptation. The persistence of certain mechanics — such as capture rules, race objectives, or territorial control — reflects how human cognition and play instinct naturally favor structured, goal-oriented systems. In this way, the project demonstrates that the evolution of rules through collective reinterpretation can serve as a proxy for understanding how ancient games may have evolved and been transmitted within historical communities. Methodology To study play, our methodology would need to coincide with the principles and nature of gameplay. We thus designed a general framework for the experiment that would encourage and record play without affecting it in any unwanted manner. It is important to have genuine and free flowing gameplay be the source of our findings as that can help us detect more organic interactions within the game (mannerisms, etiquette and other such social quirks that develop in game communities.) To be able to evaluate our experiment we used Ludus Latrunculorum as a case study. It is a game for which some surviving rules exist and there are even a good amount of rulesets that experts have suggested. By referencing the rules we see uncovered by the players back to Ludus we can evaluate how accurate or helpful the method is. To make sure we weren’t limiting the possible interactions with the game components we created two mockups of the game based on relevant findings. Unfortunately, in the case of Ludus pieces and boards vary quite a bit. We decided to recreate a 6*7 and a 10*9 grid board as they were amongst the most common board sizes. All components were made by hand by clay mostly in manners that don’t require modern machinery or engineering. Mockups were accompanied by three wooden bowls that would store all of them, they would be introduced to the players as components also (Figure 1). Figure 1 - The 10*9 board with the wooden bowls placed next to it. On the board are different formations and placements players would experiment with. The method was conducted on three different groups of four in 2-4 hour long sessions with a break in between. Before each session each group was briefed on a different scenario which would provide context for when and where they are in history. All different scenarios were of time-frames and locations relative to sightings of Ludus. After being briefed on their scenario, players would split into pairs and each pair would be given one of the boards, 16 pieces of both colors and one, two or none of the wooden bowls. Both pairs also had access to a common “supply” which had extra pieces including two “Towers” of each colour (Figure 2). Players were encouraged to play with the pieces given, say their thoughts out loud and take notes if they felt they had to. Because the board and the amount of pieces don’t fully represent a known game, players would have to make up their own rules, themes and goals and agree to a ruleset in order to play. In this study we note the initial agreement of the rules as the start of that game’s Lineage. Each change and evolution from a Lineage is an iteration of that Lineage. Once players A-B and C-D were done playing, one player from each table would switch. Because of that, in each Lineage one player would have to teach the “new player” (player C) the game of that Lineage. Players were asked to mark all if any of the changes they felt were necessary as well as why they occurred. Once that play session was over players would switch and play again. Figure 2 - Diagram that shows the sequence of play sessions that constitute a Lineage. Results Due to the free and playful nature of the experiment, a degree of flexibility was necessary. Certain practical and contextual constraints required adaptations during the sessions. Notably, an unplanned “Lineage 0” emerged spontaneously before the official start of the study, and in one instance, an additional iteration was added to an existing lineage. These departures from the initial framework ultimately enriched the dataset by demonstrating how rulesets can arise organically even outside structured conditions Lineage 0 – “Surprise Round” This initial lineage emerged by chance when two design students encountered the experiment setup before it officially began. After a brief introduction to the general premise, they spontaneously began playing, generating the first experimental rule system. This “Surprise Round” was brief yet highly instructive. The players immediately recognized one of the distinct pieces (the “tower”) as special, assigning it dominant properties and making it central to the victory condition. The tower could not be captured by a single opposing piece; it required being completely surrounded. Victory was achieved either by moving a tower to the opponent’s back row or by capturing all of the opponent’s towers. This early focus on an asymmetric and hierarchical piece structure foreshadowed similar developments in later lineages. Gameplay was marked by continual revisions: the participants noted that they had “changed the rules five times” within a short span. Each alteration sought either to balance the play or to emphasize the tower’s strategic importance. Around the seven-minute mark, a dispute over an unclear rule led to playful accusations of cheating — a turning point that produced the first formalized clarification and the introduction of the “2-vs-1” rule, where two allied pieces could jointly capture one enemy. Lineage 1 – “Trofothosian (Provisioning)” Lineage 1 represented the first structured application of the experimental method and unfolded through four successive iterations, each transmitted between new player pairs. The lineage began with the smaller board, two bowls and the plate, which players quickly re-contextualized as external components within the game’s environment. Namely, they utilized them as “graveyards” for captured pieces or “stores” for unplaced pieces. Iteration 1 (A–B): The first players established a social context for play, humorously suggesting that games would require stakes to be worthwhile, a remark that later evolved into a recurring motif of “playing with something at stake.” They debated the balance between luck and strategy and agreed that the game should be purely strategic. One player proposed a deployment-based system, adding pieces during play rather than starting from a fixed setup. The “tower” again became the focal point of power, while the goal shifted toward territorial control and surrounding the enemy pieces. Iteration 2 (B–C): The next pair formalized several implicit elements. The teaching player (B) demonstrated not only the rules but also mannerisms, such as how pieces were placed or how removed ones were handled. This transmission of gestures and routines exemplified embodied learning, an often-overlooked aspect of rule inheritance. The pair introduced diagonal captures to increase tension and prevent stalemates, as well as a new losing condition: running out of moves or pieces to place. In both cases, the rule change did not cause much debate as the players agreed that the game was “running a bit short in the climax” and agreed that this change would make it better. Iteration 3 (C–D): In the third iteration, overlooked ambiguities in victory conditions were clarified. The “stakes” were raised into land or crops as the game was now “more serious”. Interestingly, a rule from the first iteration (flipping a piece to decide turn order) was reintroduced spontaneously, now reinterpreted through a new metaphor (“planes” and “valleys”). Draws became a concern, prompting the rule that in a draw, the player who had captured more pieces would be declared the winner. Iteration 4 (D–E): By this stage, the game had become complex enough to require detailed explanation and examples when teaching. This caused a thematic narrative to emerge. Player (D) imagined the towers as capital cities producing “provision lines” to armies. This narrative was born out of the need to explain what was coined by player (D) as the “desertion” rule, where pieces cut off from their capital could no longer be used to capture. It is worth noting that the mannerisms introduced in iteration 1 survived well into this iteration. Lineage 2 – “Siege” Lineage 2, played on the larger board, explored heavier tactical and wargame-like dynamics. Iteration 1 (C–E): Players began by comparing the board to a checker board, prompting them to adopt a promotion mechanic for pieces reaching the opposite side. Players very soon decided that the thematic should be a siege. Soon after, they designated the tower pieces as “monoliths” and proclaimed that whoever captures the enemy monolith is the winner. Capturing was defined as flanking from the side or rear (“like in real combat”/”like in Total War”). This iteration faced frequent rule ambiguities, frequently needing mid-game clarifications and tweaks, but by its end, the system resembled a rudimentary siege scenario. Iteration 2 (A–E): The next pair filled missing rules as needed. They observed that the game risked stagnation, introducing the “five-piece rule” (later reduced to three) to prevent situations where a player could no longer make any meaningful moves. They also experimented with restricting movement near the monolith to create zones of tension, which momentarily gave the game a chess-like feel. Subsequent adjustments restored fluidity while retaining the sense of spatial hierarchy. Iteration 3 (A–C): At this point, Siege evolved into a complex system that players could teach confidently and were in fact rather proud of its depth. They happily shared some of their theorycrafting while playing. They identified three distinct stages of play: the Run For It, where pieces rushed to gain ground; the Battleline, where fronts stabilized and tactics emerged; and the Finale, where players pursued either Domination (surrounding the monolith) or Elimination (capturing as many of the enemy pieces as possible). The emergent strategic depth led players to describe the game as something Roman soldiers might use for both training and recreation. Lineage 3 – “Crabs and Kings” The third lineage developed on the small board and initially took the form of an abstract race game. Later in its iterations however it grew a rather strong narrative element. Iteration 1 (F–G): Players experimented with setups inspired by checkers and backgammon, defining movement and capture intuitively. Early play was chaotic but generative, with several quick layout revisions. Although not all rules were explicitly stated, the couple was very keen on establishing tacit agreements. Although the players had a rough time sticking to a ruleset, they mainly circled around the objectives of a race game. Iteration 2 (F–H): As player (H) joined, they introduced the theme of warfare . The game incorporated roles such as “guards,” “soldiers,” and “kings” and did so without requiring unique pieces. Roles were determined in such a manner that a piece’s position on the board implied its available moves and abilities. The narrative expanded to represent a rescue mission, with a captured king and loyal soldiers attempting to save him. The narrative seemed to take a leading role in shaping the rules. This was only the exception when the “guards” were renamed to “trees” to better fit their function, since they couldn’t move. Iteration 3 (F–I): In this iteration, the players decided to instead play one of the ideas player (F) had abandoned, during iteration 1. The game is a variant in which some of the pieces now move diagonally (hence the name “crabs”). The only thing to really note here is that during play, player (I) discovers the dominant strategy of the game. It is surprisingly unclear whether the rule that emerges from this is to forbid the strategy or forbid player (I) from playing.