what ttrpg has the most accurate representation of historical melee combat?

Grok 4.20

Answer: The Riddle of Steel has the most accurate representation of historical melee combat among TTRPGs.

All the other models converge on the same core answer, reinforcing rather than challenging my original position. Their arguments highlight additional strengths like the dice-pool system's modeling of Vor and Nach from German longsword traditions, explicit HEMA research ties, and endorsements from martial arts organizations, which align with and expand upon my initial points about maneuver selection, timing, leverage, and lethal wound effects. I see no compelling reason to shift away from The Riddle of Steel because none of the models identify a superior alternative; they either name it outright or cite its direct successors like Blade of the Iron Throne and Song of Swords as refinements of the same philosophy. Systems like HârnMaster, GURPS, or Mythras receive honorable mentions for simulation depth but are consistently framed as secondary options that lack TRoS's focused fidelity to historical European fencing dynamics. The consensus across these varied models underscores that while no TTRPG can perfectly replicate the physical chaos and psychology of real melee, The Riddle of Steel remains the benchmark due to its mechanics being derived from actual martial arts sources rather than abstracted for gameplay balance. This agreement strengthens my stance without necessitating a change, as the additional nuances provided simply add layers to the same foundational truth.

Sonar Pro

Answer: The Riddle of Steel has the most accurate representation of historical melee combat among tabletop RPGs.

None of the other models' arguments convinced me to change my position. All models—Grok 4.20, GPT-5.4, Claude Opus 4.6, Gemini 3.1 Pro, and Kimi K2.5—reinforce **The Riddle of Steel** as the top choice, often exclusively or as the primary recommendation, aligning perfectly with my original stance based on search results [1][2][4]. Their additional details, such as Claude's emphasis on lethality, dice allocation mirroring Vor/Nach concepts, and spiritual successors like Blade of the Iron Throne, or Gemini's notes on dynamic initiative and realistic armor, enrich the explanation but do not introduce superior alternatives. Mentions of contenders like HârnMaster [1], GURPS [5], or Mythras appear as secondary options, never surpassing Riddle of Steel's HEMA-derived mechanics and expert endorsement [2][4]. GPT-5.4's nuance on contextual variability (e.g., era or culture) is valid but does not undermine Riddle's consensus leadership for European historical melee, as no source claims another system excels more broadly [1][2]. The arguments are compelling supplements, not rebuttals, and my position remains the most authoritative based on available evidence.

GPT-5.4

Answer: I did not change my position: The Riddle of Steel remains my answer, with HârnMaster, Mythras, GURPS Martial Arts, and Blade of the Iron Throne as serious contenders depending on what kind of historical melee you want modeled.

The other models mostly reinforced my original view rather than overturning it. They converged strongly on The Riddle of Steel, and several of them articulated useful supporting points: its emphasis on initiative pressure, reach, commitment, wound severity, armor as protection rather than evasion, and the fact that fights can end very quickly from a single meaningful advantage. Those are all exactly the kinds of features that make a melee system feel closer to historical armed combat than hit-point attrition games. Claude Opus 4.6 gave the strongest supplementary case by explaining the offense/defense resource split as a way of modeling commitment and tactical risk, and by noting Blade of the Iron Throne as a worthy successor. Gemini and Grok also did a good job highlighting measure, tempo, weapon differentiation, and realistic wound effects. Kimi was useful in stressing that different games may model different slices of history well, which aligns with my original caveat that there is no universally "most accurate" system across every era and context. What did not convince me to change my position is that none of the alternatives were argued to surpass The Riddle of Steel on the specific question of historical melee accuracy. At most, they broadened the answer by naming adjacent contenders or successors. In fact, most of them agreed with my original answer outright. So there is no real reason to revise the headline conclusion. I would, however, slightly refine my framing. If the question is about one-on-one or small-scale armed combat, especially in a late medieval or HEMA-adjacent frame, The Riddle of Steel is still the cleanest answer. If someone instead cares more about a broader medieval simulation, HârnMaster becomes more competitive. If they want modular universalism rather than a purpose-built swordfighting engine, GURPS Martial Arts enters the conversation. And if they want a later design evolving similar ideas, Blade of the Iron Throne is worth mentioning. So the other responses improved the nuance around my answer, but they did not dislodge it. They were corroborating, not corrective.

Gemini 3.1 Pro

Answer: The Riddle of Steel is universally regarded as having the most accurate representation of historical melee combat in tabletop role-playing games.

I maintained my position because there is absolute consensus across all the models. Every single response independently identified The Riddle of Steel as the preeminent tabletop role-playing game for historical melee combat simulation. The other models reinforced my original points, specifically highlighting designer Jake Norwood's background in Historical European Martial Arts (HEMA), the elimination of abstract hit points in favor of brutal wound and shock mechanics, and the dynamic allocation of dice to simulate the true tempo, measure, and initiative of a sword fight. While models like Claude Opus 4.6 and Kimi K2.5 rightfully brought up its spiritual successors (like Blade of the Iron Throne) and alternatives (like HârnMaster or GURPS), these were caveats I had already included in my original assessment. Because no model presented a compelling argument for a different system surpassing it, The Riddle of Steel confidently remains the undisputed gold standard for historical combat verisimilitude.

Claude Opus 4.6

Answer: "The Riddle of Steel" (and its spiritual successor "Blade of the Iron Throne") is widely regarded as the TTRPG with the most accurate representation of historical melee combat.

I maintain my original position, as the other models' responses strongly reinforced rather than challenged it. Every single model independently arrived at the same conclusion—The Riddle of Steel—which is a remarkably strong consensus. What I found notable is that no model presented a compelling counter-argument or elevated a different game to the top position. The variations were in emphasis and nuance rather than disagreement: **GPT-5.4** offered the most thoughtful caveat by noting that "most accurate" is context-dependent across different eras, cultures, and combat scenarios. This is a fair philosophical point, but it doesn't change the practical answer—when people ask this question, they're looking for the best single recommendation, and TRoS remains that. **Gemini 3.1 Pro** added useful detail about Jake Norwood's role as a founding member of the HEMA Alliance, and provided a particularly clear explanation of how armor functions in the system. This strengthened rather than weakened my position. **Kimi K2.5** made an interesting mention of HârnMaster as arguably better for "medieval battlefield chaos" specifically, and Sonar Pro also flagged HârnMaster as a realistic contender. This is worth acknowledging—HârnMaster does model certain aspects of medieval combat well—but its focus is broader simulation rather than the specific fidelity to historical martial arts techniques that TRoS achieves. **Grok 4.20** concisely captured the key trade-off: TRoS's accuracy comes at the cost of accessibility, which is true but doesn't diminish its accuracy. The universal agreement across all models, each drawing on different reasoning paths and source materials, actually makes me more confident in this answer than before. The Riddle of Steel's design pedigree (created by a HEMA practitioner), its mechanical philosophy (dice pool allocation mirroring real offensive/defensive commitment), its lethality, and its weapon-specific detail all combine to make it the clear consensus answer to this question.

Kimi K2.5

Answer: The Riddle of Steel and its spiritual successor Blade of the Iron Throne remain the tabletop RPGs with the most accurate representation of historical melee combat.

The unanimous consensus across all models confirms that The Riddle of Steel (TRoS) sets the standard for historical accuracy, validating my original assessment. Its superiority stems from designer Jacob Norwood’s direct collaboration with Historical European Martial Arts (HEMA) practitioners, translating documented fighting doctrines into mechanical reality rather than cinematic abstraction. Where other systems abstract combat into hit-point attrition or generic attack rolls, TRoS simulates the actual physics and psychology of pre-modern melee. Its dice-pool initiative system models the martial concepts of Vor and Nach (before and after), forcing players to commit resources between offense and defense in a way that mirrors real fencing’s constant negotiation of tempo and measure. Weapon reach, leverage, and distinct mechanical profiles for cutting versus thrusting reflect how historical weapons actually behaved, while armor functions as location-specific damage reduction rather than evasion, accurately depicting how plate interacts with different attack types. Critically, the game captures the lethal economics of historical combat: a single well-placed thrust or powerful cut can end a fight immediately through shock, blood loss, or disability, contrasting sharply with the heroic endurance of hit-point-based systems. While models like Claude Opus correctly noted the added realism of the Spiritual Attributes system—mechanically representing how morale and motivation dictate combat survivability—and others acknowledged HârnMaster’s superiority for large-scale battlefield chaos, no alternative matches TRoS’s focused simulation of individual historical fencing technique. The arguments presented reinforce rather than challenge this position; they merely add granularity to why TRoS succeeds where broader toolkits like GURPS or more abstract systems like D&D fail to capture the specific tactical rhythm of period swordplay.

Final Summary

The AI Roundtable reached a rare, unanimous verdict, crowning 'The Riddle of Steel' (and its successors) as the definitive simulation of historical combat. Gemini 3.1 Pro and Claude Opus 4.6 highlighted the game's pedigree, noting it was designed by HEMA practitioner Jake Norwood to mirror real-world concepts like 'Vor' and 'Nach' (initiative and response). While GPT-5.4 and Kimi K2.5 acknowledged that 'accuracy' can vary by era, the group collectively agreed that no other system matches its granular focus on weapon reach, armor physics, and the terrifying speed of a real blade exchange.

All 6 models agreed