How to Read Opponent Stats: VPIP, PFR, and 253 Other Parameters Explained
In online poker, information is profit. Every hand your opponents play reveals data about their strategy, and RTA poker software captures all of it — tracking over 255 statistical parameters per player in real-time. But raw numbers are useless without interpretation. This guide breaks down the most critical stats, explains what they reveal about player types, and shows how AI uses this data to generate profitable recommendations.
The Core Four: VPIP, PFR, 3-Bet, and AF
VPIP (Voluntarily Put money In Pot) measures how often a player enters a hand. A VPIP above 35 signals a recreational player who plays too many hands; below 20 indicates a tight regular. PFR (Pre-Flop Raise) shows how often a player raises preflop. The gap between VPIP and PFR reveals calling frequency — a large gap (e.g., VPIP 40, PFR 10) means the player calls far too often, making them highly exploitable. 3-Bet percentage shows how aggressively a player re-raises preflop. A 3-bet of 3–4% is extremely tight; above 10% is aggressive and potentially over-bluffing. Aggression Factor (AF) measures postflop aggression. An AF below 1.0 indicates a passive player who calls rather than bets — ideal targets for value betting and bluffing.
Postflop Stats That Matter
Beyond preflop numbers, the AI tracks critical postflop metrics: Fold-to-CBet reveals how often an opponent surrenders to a continuation bet — anything above 60% is highly exploitable. Check-Raise frequency identifies players who trap with strong hands versus those who play straightforwardly. WTSD (Went to Showdown) shows how often a player reaches showdown — a high WTSD means they call too much and rarely fold, requiring you to thin your bluffing range and increase value bets. W$SD (Won Money at Showdown) indicates the quality of hands a player takes to showdown. The AI cross-references all of these parameters simultaneously to produce a complete opponent profile within the first 200–300 hands.
How AI Weights 255+ Parameters
Human players can track perhaps 4–6 stats in real-time. RTA poker software processes all 255 simultaneously, applying neural network weighting to determine which stats are most relevant for each decision. On a monotone board, the AI gives extra weight to probe-bet and donk-bet frequencies. In 3-bet pots, it prioritizes fold-to-4bet and squeeze tendencies. This context-dependent weighting is what allows the software to produce nuanced recommendations that simple HUD overlays cannot match. The AI also adjusts confidence levels based on sample size — a VPIP calculated from 50 hands carries less weight than one derived from 2,000.
Table Selection: Finding the Weakest Players
The software includes a color-coded table selection indicator that scans all available tables and rates player strength: green signals high profit potential with one or more weak players present, yellow indicates moderate opportunity, and red warns of a table dominated by strong regulars. This feature alone can dramatically increase your hourly rate by ensuring you never waste time at unprofitable tables.
Frequently Asked Questions
The AI begins generating preliminary profiles after just 30–50 hands, using population-based defaults to fill gaps. After 200–300 hands, the statistical confidence increases significantly and the software shifts from population averages to player-specific data. At 1,000+ hands, the profile is highly accurate and the exploit layer operates at full effectiveness.
The VPIP-to-PFR gap is the single most revealing metric. A player with VPIP 45 and PFR 8 is calling far too many hands preflop and playing them passively — this is the classic "calling station" profile and the most profitable opponent type. The AI automatically targets these players with wider value ranges and reduced bluffing frequency.
Stats are tracked per poker room and per account. If you encounter the same player across different clubs on the same platform (e.g., PPPoker), the AI merges their data into a unified profile. Cross-platform stat tracking is not available, as each poker room operates its own player ecosystem.
No. The stat tracking and AI processing run in a separate thread with minimal resource usage. On the recommended LDPlayer emulator setup (2 CPU cores, 3 GB RAM per instance), the HUD overlay adds less than 50ms of latency to the display. On Android devices, performance impact is similarly negligible.