Pax Historia Lets You Rewrite History and Watch It Play Out in Real Time

The Macro: Grand Strategy Games Have Not Changed in Twenty Years

The grand strategy genre has a devoted audience and almost no innovation. Paradox Interactive dominates with Crusader Kings, Europa Universalis, Hearts of Iron, and Victoria. These are brilliant games with deep simulation systems, and they have not meaningfully evolved their core design in two decades. You pick a country, you manage resources, you make decisions on a map, and the AI opponents follow scripted behavior patterns.

The problem is that scripted AI creates predictable games. After enough hours, experienced players learn the patterns. They know that France will always ally with Castile, that the Ottomans will always push into the Balkans, that the AI will make the same mistakes on the same difficulty settings. The historical sandbox becomes less of a sandbox and more of a puzzle with known solutions.

AI changes this equation entirely. Instead of scripted opponents following decision trees, you get opponents (and allies) that reason about situations, adapt to player actions, and produce outcomes that nobody, including the developers, can predict. The simulation stops being a game and starts being a genuine experiment in counterfactual history.

The broader AI gaming market is heating up. Companies like Altera and NVIDIA are building AI-driven NPCs for open-world games. But grand strategy is a particularly interesting application because the genre already depends on emergent storytelling and complex simulation. The AI does not need to animate a character or render a face. It needs to model geopolitics, economics, and military strategy. That is a great fit for current LLM capabilities.

The Micro: Choose a Country, Change a Variable, Watch History Unfold

Pax Historia is an alternate-history sandbox game where you choose a country, pick a moment in time, and change something. What if Rome never fell? What if the Soviet Union survived? What if Genghis Khan ruled for another generation? The AI simulates what happens next, and the result is a unique, procedurally generated historical narrative that plays out in real time.

The platform goes beyond strict historical scenarios. Want to simulate an alien invasion? A tech startup? A fictional country? The system supports it. The ability to create, edit, and customize scenarios suggests an underlying simulation engine that is general-purpose enough to handle inputs outside the historical dataset.

The product is web-based with user authentication and what appears to be a game library where you can save and return to scenarios. The interface supports dark and light themes and works on mobile, which suggests the team is thinking about accessibility beyond the hardcore strategy gaming audience.

Eli Bullock-Papa and Ryan Zhang are the cofounders. Zhang’s bio on the company’s YC page reads simply: “I like history.” That kind of founder-product fit matters in a space where understanding the domain deeply is essential to building something that resonates with the audience. The company went through Y Combinator’s W26 batch.

The competitive positioning is interesting because Pax Historia is not really competing with Paradox for the same players. Paradox games require hundreds of hours of learning and dozens of dollars to buy (plus DLC). Pax Historia seems to be going after a broader audience: people who are fascinated by counterfactual history but do not want to master a complex 4X strategy game to explore it.

The marketplace tag on YC’s profile suggests there may be a community element, where players share scenarios, trade custom setups, or build on each other’s alternate histories. That kind of user-generated content loop could be a major retention driver.

The Verdict

Pax Historia sits at a genuinely interesting intersection of AI capabilities and an underserved audience. The alternate history community is large, passionate, and currently served mostly by Reddit threads, books, and the occasional YouTube video. A playable, AI-driven alternate history sandbox is something this audience has been waiting for.

At 30 days: simulation depth. Does the AI produce outcomes that feel historically plausible, or does it generate random noise dressed up as geopolitics? The credibility of the simulation is everything for this audience.

At 60 days: replayability. Can you run the same scenario ten times and get ten genuinely different outcomes? If the AI converges on similar results regardless of the starting conditions, the sandbox loses its appeal.

At 90 days: community formation. Grand strategy games live or die on their communities. If Pax Historia can build a community of scenario creators and alternate history enthusiasts, the content flywheel becomes self-sustaining.

I think the timing is perfect for this product. LLMs are good enough to simulate complex strategic reasoning. The audience exists and is underserved. And the “what if” framing is inherently shareable, which gives the product a natural distribution advantage. Whether this becomes the next Paradox or stays a niche experiment depends entirely on the depth of the simulation engine.