Real-Time Assistance (RTA): Poker’s Fastest-Growing Gray-Zone Cheating

Solvers changed modern poker: every top player uses them to study the game. But the same tool, moved from post-session review into a live hand, changes nature entirely — from learning to cheating. That is RTA (Real-Time Assistance), the fastest-growing and most contested cheating category of recent years.

Where the line sits

The industry consensus line is actually clear:

• Allowed: reviewing hands with a solver after the session; studying strategy away from the table; memorizing and applying what you learned
• Forbidden: consulting any tool during a live hand for advice on the current decision; software that reads the table and displays prompts in real time; sending your current hand to an outside person or program for input

In one sentence: knowledge in your head is fair game; outside help during a hand is cheating. Nearly every serious platform’s terms say exactly this.

Why RTA is harder to catch than bots

Bots run unattended with highly mechanical patterns, which ironically makes them easier to flag (see the bots explainer). An RTA user is a real human:

• Mouse movement, chat, and emotional reactions are all genuine, so the first layer of behavioral detection slides right off
• On any single hand, a strong player’s normal line and a solver-copied line are nearly indistinguishable
• Detection must lean on statistics: abnormal decision-time distributions (even trivial decisions take long), and long-run alignment with solver output beyond human consistency

Which is why RTA convictions usually come from reports and investigation — device checks, screen recordings, community evidence — rather than pure algorithms.

Signals players can observe

• Fixed, lengthy delays even in trivial spots (obvious folds, obvious calls) — because every hand gets “looked up”
• Extremely balanced play with none of the usual human leaks, sustained with zero drift over long periods
• Decision quality that stays flat across many tables — humans tire when multi-tabling; a lookup table doesn’t

No single signal convicts; only long-run accumulation means anything. If suspicious, save records and use the report channel — same principles as the general anti-cheating guide.

What platforms and architecture can do

Platform level: decision-time statistics, divergence analysis against solver output, device environment checks, human review of key tournament spots. Architecture level: Fair Poker again goes after the economic incentive — play-chips-only games make “hiring a solver operator” pointless; the social setting of private games among friends means an RTA user faces people who know them, and social cost is its own deterrent. Meanwhile provably fair dealing guarantees that however the player-side cat-and-mouse evolves, the dealing layer stays clean and checkable forever.