Originally published at pokerhack.org
Introduction and Definition
Spotting a poker bot at the table is a skill that blends quantitative observation with an understanding of table dynamics. In 2026, players increasingly encounter automated agents operating in online poker rooms and at virtual tables, which prompts a structured approach to detection. This article defines seven statistical tells that help distinguish human from bot behavior without relying on rumor or speculation. While no single tell is conclusive, a convergence of patterns over multiple hands raises the probability that a bot is present, prompting cautious engagement and further analysis.
From a research perspective, bot-like behavior can arise from scripted decision logic, macro-level timing strategies, and response consistency that diverge from typical human variability. In the context of online poker, recognizing these patterns requires careful data collection, normalization across table states, and a disciplined comparison to baseline human behavior. The goal is to empower players to make better-informed decisions and to understand the limitations of pattern-based detection in a complex, real-time environment.
As with any discussion of online poker fairness and automation, it is important to acknowledge that platform operators implement security measures and policies designed to deter abuse, and regulatory bodies require ongoing oversight of software that interacts with real-money play. This article focuses on statistical tells and does not claim definitive identification of any individual player as a bot.
Core Content
Tell 1: Uniform Bet Sizing Across Diverse Scenarios — Bots tend to apply highly regular bet sizing across a wide range of hand types and board textures. When a player consistently mixes aggression, bluff frequency, and value bets with mechanical precision that ignores common human pitfalls (e.g., overbetting with marginal hands after a scare card), it can indicate scripted logic. Compare their sizing distribution across similar spots, and look for low variance in sizing choices that would normally reflect table psychology and dynamic reads.
Tell 2: Minimal Table Talk or Predictable Communication — While humans respond to action and table dynamics with talk and dynamics, bots typically minimize chat or respond with highly constrained phrases. A paucity of spontaneous, varied verbal or chat-based interaction, even in casual rooms, can be a statistical signal when combined with other tells. Track the ratio of non-action messages to action decisions and examine whether communication behavior remains within a narrow, pre-programmed envelope.
Tell 3: Reaction Time Consistency — Bots often exhibit highly uniform decision latencies, with many actions occurring within a narrow time window after the board cards are revealed or a bet is faced. Human players display more drift due to cognitive load, distractions, and strategy adjustments. Construct a distribution of decision times and flag bodies of hands where action timing clusters unnaturally around specific values.
Tell 4: Action Sequencing and Card-Reading Rhythm — Some bots operate with deterministic action sequences (e.g., always calling with bottom pair on a flush draw) that align with a fixed codebook. When a table shows recurrent, structurally identical lineups of decisions across random hands, especially in early positions or blinds, this suggests algorithmic patterning. Normalize for standard variance in post-flop lines and check for overrepresented hand classes.
Tell 5: Hand-Strength Perception Versus Equity Realization — Bots can exhibit a misalignment between perceived hand strength and actual board texture, deploying value bets in spots where human players typically pause. Analyze the correlation between declared hand strength categories and real-time equity estimates using typical ranges; a persistent mismatch across multiple sessions may indicate automated decision logic rather than adaptive human strategy.
Tell 6: Consistent Exploitation of Tells Across Tables — A bot system can scale across multiple tables, exploiting similar board textures or bet-calling patterns with little variation. If a single account or a rare subset of accounts demonstrates parallel patterns across different table contexts (stakes, player pools, or platforms), it warrants deeper statistical scrutiny and cross-table comparisons.
Tell 7: Rake and Action Ecology Alignment — Bots may adapt to ecological pressures such as rake brackets by adjusting aggression to maintain win rate ceilings. When a player’s frequencies align tightly with known ecology-driven distributions (e.g., higher aggression at higher rake environments with little deviation), this could reflect preset optimization against house economics rather than organic table adaptation.
In practice, detecting bots requires convergence: combine timing data, bet-sizing distributions, chat activity, and board texture responses. For p
Read the full analysis: Spotting a Poker Bot at the Table: 7 Statistical Tells in 2026
Top comments (0)