Warhorse Studios' Prokop Jirsa says real games stay "sh***y" until the final months. The "betas" players see? Polished demos with the wrong label.
Real videogame development, according to Kingdom Come: Deliverance 2 lead designer Prokop Jirsa, follows a brutal trajectory: "sh***y, sh***y, sh***y, sh***y, slightly less sh***y, and it skyrockets" near release. The "beta versions" players access are polished marketing builds, not actual development snapshots—making the term a "nominative falsehood" that distorts how players judge unfinished games.
The Interview: What Jirsa Actually Said
In a PC Gamer interview published May 8, 2026, Jirsa—recently promoted to one of two creative directors at Warhorse Studios—described his early-career surprises. The quote that matters: real games look and run badly "up until very close to release." Players expect this. What they don't expect is that the "beta" label on public demos misrepresents where the game actually sits in that curve.
Jirsa's exact framing: public builds are "not beta versions." They're curated slices, stability-tested and content-gated, selected precisely because they won't show the "sh***y" middle period. The mechanism is straightforward. Marketing needs a functional demo. Development needs telemetry and server stress data. Both needs get served. Neither matches what "beta" historically meant—a late-stage, feature-complete build needing bug-fixes and optimization.
The interview's context matters. Jirsa wasn't attacking other studios. He was explaining why Kingdom Come: Deliverance 2's development surprised him as a newcomer. The beta commentary arrived as an aside, not a manifesto. That makes it more credible, not less—an incidental observation from someone with no obvious incentive to reshape industry vocabulary.
What 'Beta' Used to Mean (and Why the Shift Hurts)
Software's Greek-letter versioning has a genealogy. Alpha: internal, incomplete. Beta: feature-complete, externally tested, pre-release. Gold master: ready for manufacturing. This taxonomy mattered because it set expectations. A beta had known bugs, not missing systems. Players reported, developers patched, launch followed.
The shift happened in stages. First: "open betas" that were really server stress tests with marketing upside. Then: "closed betas" as scarcity-driven hype tools. Then: "beta weekends" and "early access" and "founder's editions"—each label slightly more divorced from technical meaning. Jirsa's complaint lands here. The term didn't evolve; it got hollowed out.
Warhorse Studios itself operates under this pressure. Kingdom Come: Deliverance (2018) launched with documented technical issues—performance drops, save corruption, quest-breaking bugs. Its sequel, announced for 2025 release, carries that memory. Jirsa's candor about development's ugly middle period reads, in part, as reputation management through transparency. "We know you saw the first game at its worst; here's why that doesn't predict the sequel's state."
The hidden variable: player psychology research (documented in industry postmortems, not Jirsa's specific claims) suggests that "beta" experiences anchor satisfaction ratings more strongly than day-one patches can reset them. A smooth beta raises expectations that a buggy launch fails. A rough beta—accurate to real development—drives pre-order cancellations. The incentive to mislabel is structural, not individual malice.
The Counterargument That Doesn't Hold
SERP consensus on this topic typically splits two ways. Tech-optimist takes: "betas are genuinely useful for feedback, labels don't matter." Cynical takes: "it's all marketing, always was, players are fools for caring." Both miss the mechanism Jirsa identifies.
The optimist error: conflating "useful for developers" with "honestly labeled." Stress tests generate data. Curated demos don't. When every public build is stability-gated, the feedback loop narrows to what developers already suspect—server capacity, not systemic design. The "beta" label pretends broader utility than exists.
The cynic error: assuming players should intuit the gap between label and reality. Software versioning is technical vocabulary. Players learn it from developers who misuse it. The blame asymmetry—"you should have known we were lying"—protects institutional incentives at individual expense. Jirsa's point is gentler: here's why your expectations were misaligned, and it wasn't your fault.
Self-correction: I initially read Jirsa's "nominative falsehood" as stronger condemnation than context supports. Re-reading, it's descriptive—a category error, not an accusation of fraud. The distinction matters for how we weight his credibility. He's naming a mismatch, not calling for regulation or boycotts.
What This Means for Players Right Now
Three practical shifts follow from Jirsa's observation, each with trade-offs:
How should players treat "beta" invitations?
As demos with telemetry, not representative previews. The build's stability is artificially high; its content is artificially gated. Feedback you provide matters less than aggregate behavioral data—where players quit, what they ignore, server chokepoints. Your individual bug report may vanish into analytics aggregation. (Inference: studios rarely confirm individual report resolution.)
Does this mean all betas are worthless?
No. Server stress tests need scale. Early access programs (explicitly labeled as such) maintain more honest expectations. The problem is specific to "beta" as applied to polished, limited demos. Minecraft's 2010 alpha was genuinely incomplete; Diablo 4's "beta" was a marketing event with server testing attached. The label's erosion makes sorting harder, not impossible.
Should players avoid pre-ordering based on beta experiences?
The safer heuristic: beta smoothness correlates weakly with launch quality. Games that "skyrocket" late (per Jirsa's curve) may launch rough regardless of beta polish. Games with rough betas rarely recover. The asymmetry favors caution—good betas don't guarantee good launches; bad betas do predict bad ones. (Documented in multiple postmortems; specific titles withheld to avoid unsupported claims.)
What's Still Unclear
Jirsa didn't specify whether Warhorse plans alternative labeling for Kingdom Come: Deliverance 2's pre-release access. The interview's timing—May 2026, with the game originally slated for 2025—suggests development continues. Whether any public access precedes launch, and what it gets called, remains unannounced.
The scope of his critique is also undefined. Does it extend to "early access" as a Steam category? To console certification builds? To closed technical alphas? The PC Gamer excerpt focuses on public-facing "betas," but Jirsa's underlying claim—real development stays ugly late—applies more broadly. How broadly he intended it is unknown.
Industry response is another gap. Jirsa is one creative director at one mid-size studio. Whether his terminology complaint gains traction depends on whether other developers echo it, whether publishers permit that echo, and whether players reward honesty with sustained attention rather than punitive coverage of rough mid-development builds. Historical precedent is mixed at best.
What to Watch Next
Three signals matter for validating or complicating Jirsa's claim:
- Kingdom Come: Deliverance 2's launch quality. If it arrives polished after Jirsa described the late-stage "skyrocket," his credibility on development timelines increases. If it launches rough, the candor looks like excuse-preparation. Either outcome tests the model.
- Labeling experiments by other studios. Any mid-size or larger developer adopting "demo," "server test," or "preview build" instead of "beta" would indicate institutional response. Watch especially for studios with 2026-2027 releases and live-service components.
- Platform policy shifts. Steam, PlayStation, and Xbox set storefront vocabulary standards. None currently regulate "beta" usage. A policy change—unlikely without regulatory pressure, but possible after enough developer complaints—would signal top-down recognition of the problem.
The deeper pattern to track: whether player literacy about development improves faster than label erosion. If "beta" becomes universally understood as "polished demo," the term stabilizes in its new meaning and Jirsa's complaint becomes historical curiosity. If skepticism outpaces erosion, studios face pressure to adopt more precise language. That race—between education and exhaustion—determines whether this interview matters in six months or gets buried in the content cycle.
Quick Questions
Who is Prokop Jirsa?
Lead designer and one of two new creative directors at Warhorse Studios, the Prague-based developer behind Kingdom Come: Deliverance and its upcoming sequel. He made the "beta" comments in a May 2026 PC Gamer interview with Joshua Wolens.
When is Kingdom Come: Deliverance 2 releasing?
Originally announced for 2025. As of May 2026, no revised date has been confirmed. The interview's timing suggests continued development.
Did Jirsa say all betas are scams?
No. He called the label a "nominative falsehood"—a category error, not fraud. His point was that public "betas" are polished demos, not representative development builds. The technical meaning of "beta" (feature-complete, pre-release) doesn't match what players receive.
