"Real Players Win Only 25% of Matches Against Bots": Developer Uncovers Unintended Bot Difficulty in Economic Card Game

“I Looked at My Game’s Stats and Realized People Almost Can’t Beat the Bots”

Andrew Perepechay analyzed his game Rastushiy Gorod’s stats and discovered humans win just 25% of matches against bots. The bots were designed as practice opponents, not final bosses.

Why This Matters

‘The technical reality defies ideal models: Andrew Perepechay’s economic card game Rastushiy Gorod was designed with simple fill-in bots for practice, but analytics revealed a staggering failure scale—humans win just one-quarter of matches. This exposes how even ‘honest’ bot logic can create an unintended skill gap that undermines core gameplay loops.‘

Key Insights

• Players win only 25% of matches against bots in Rastushiy Gorod, a rate far below the developer’s expectations (Perepechay, 2026).
• Difficulty in an economic card game stems from decision-making complexity, not adjustable aim like shooters; lowering bot skill would undermine honest competition (Perepechay, 2026).
• An analytics dashboard rebuild shifted focus from game metrics (e.g., card pick rates) to player behavior metrics (e.g., win rates, drop-offs), providing deeper product insights (Perepechay, 2026).
• Most matches involve one human plus several bots because players start games with bots by choice, reducing the intended multiplayer experience (Perepechay, 2026).

Practical Applications

• • Use case: Analytics dashboards for live games; shifting focus from game-centric metrics to player-centric data reveals unexpected balance issues like bot dominance.
• • Pitfall: Designing bots as mere fillers without testing their comparative difficulty; leads to imbalanced matchmaking where humans lose disproportionately.

References:

• https://dev.to/andrew_perepechay_ee25378/i-looked-at-my-games-stats-and-realized-people-almost-cant-beat-the-bots-1j62