19 HUD Stats & Online Reads
Live poker gives you a face, a voice, a trembling hand, the way a stack gets cut for a bet. Online poker takes all of that away and replaces it with something both colder and, in the right hands, far more powerful: data. Every hand your opponent has ever played at your tables can be recorded, aggregated, and projected back onto the felt as a small panel of numbers floating beside their name. That panel is a Heads-Up Display, or HUD, and learning to read it is the online equivalent of learning to read a face — except the numbers do not bluff, do not get tired, and do not vary with the lighting.
This chapter is about turning that data stream into decisions. We will define the core statistics, give the typical value ranges that separate a tight player from a loose one and a passive player from an aggressive one, and — most importantly — turn each stat into a concrete exploit. We will be disciplined about sample size, because a HUD lies most dangerously when it is confidently displaying a number built from nine hands. And we will close with the parts of online reading that are not in the database at all: timing tells, note-taking, and the brutal information-management problem of playing six, twelve, or twenty-four tables at once.
A HUD does not replace the thinking in the rest of this book. It is a hypothesis generator. It tells you where an opponent is likely to deviate from a sound baseline, and the exploit chapters that surround this one tell you what to do about it. Reads online are still probabilistic. A 70% fold-to-c-bet means he folds most of the time, not every time — and the one time he doesn’t is not evidence the HUD was wrong.
19.1 What a HUD actually is
Tracking software — the major desktop trackers and their equivalents — imports the hand histories your poker client writes to disk and stores them in a database. The HUD is the live overlay: as you play, it queries that database for each opponent and displays their stats in real time. Critically, a HUD only knows hands it has seen. If you have 40 hands on a player, every number on screen is computed from those 40 hands. There is no global database of all humanity; you are reading your own sample of this player.
Most stats are expressed as percentages of the opportunities to take an action. VPIP of 24% does not mean “24% of hands”; it means “on 24% of the hands where he had the option, he voluntarily put money in.” This denominator matters enormously for sample size, which we return to below.
19.2 The core stats
The art is not memorizing every stat a tracker can compute — there are hundreds — but choosing a compact, high-signal layout and knowing each number cold. Below are the workhorses. The value ranges given are typical archetypes for full-ring and 6-max online cash; treat them as rough population landmarks, not laws. Live games and micro-stakes pools run looser and more passive; high-stakes regulars cluster tighter and more aggressive.
Preflop: how often and how hard they enter
VPIP — Voluntarily Put In Pot. The percentage of hands where the player put money in by calling or raising (blinds posted don’t count). This is the master dial of looseness.
- Very tight (“nit”): ~12–17%
- Solid 6-max reg: ~20–26%
- Loose / recreational: ~30–45%+
PFR — Pre-Flop Raise. The percentage of hands raised preflop. Read it together with VPIP, never alone.
The VPIP/PFR gap is the single most informative preflop read you have:
| VPIP / PFR | Profile | What it means |
|---|---|---|
| 22 / 19 | Tight-aggressive (TAG) | Tight range, almost always raised — the standard tough reg |
| 16 / 3 | Passive nit / “calling nit” | Enters rarely and mostly by calling — limps and flat-calls |
| 40 / 8 | Loose-passive (“fish”) | Plays everything, raises little — the classic recreational |
| 38 / 31 | Loose-aggressive (LAG) | Wide and aggressive — dangerous, high-variance |
A gap of 3–5 points is a tight-aggressive player who raises almost everything they play. A gap of 20+ points (e.g., 35/10) is a caller — they see a lot of flops with a wide, weak, face-up range.
3Bet. The percentage of times the player re-raised a preflop open. Typical regs sit around 6–10%; a 3Bet under ~4% means their reraises are nearly all premiums (fold unless you have a real hand), while 11%+ signals a light, aggressive 3-bettor whose range is full of bluffs you can call or 4-bet wider against.
Fold to 3Bet. How often they fold their open to a 3-bet. Around 50–60% is normal. Above ~65% they are surrendering too often — 3-bet them light relentlessly. Below ~40% they defend their opens stubbornly, so 3-bet for value and cut the bluffs.
Steal (open-raising from the cutoff, button, or small blind when folded to) and Fold BB to Steal are the blind-battle pair. A high stealer (button steal north of ~45%) is attacking your blinds; if his steal is high and your read says he folds to 3-bets, your blind is a printing press. Conversely, Fold BB to Steal above ~75% marks a big blind you can open-raise wider against from late position with near-impunity.
Postflop: aggression and showdowns
C-bet (flop continuation bet). How often, as the preflop raiser, they bet the flop. The population baseline is roughly 55–70%. The more useful partner stat is the opponent’s Fold to C-bet: around 40–55% is typical. A fold-to-c-bet of 60%+ means you can c-bet any two cards profitably and fire again; under ~35% means stop bluffing and value-bet relentlessly because he calls down light.
Aggression Factor (AF) and Aggression Frequency (AFq). Two ways to measure postflop aggression.
- AF = (bets + raises) / calls. A purely descriptive ratio. ~1 is passive (a calling station), ~2–3 is balanced-to-aggressive, 5+ is a maniac — but AF is distorted by a few big calls, so read it cautiously.
- AFq = aggressive actions / total actions, expressed as a percentage. More robust than AF and easier to compare across players. ~35–50% is a healthy aggressive range; under ~25% is passive.
A low AF/AFq player who is firing a second barrel is rarely bluffing — believe them. A high-AF player betting means much less.
WTSD — Went To ShowDown. Of flops seen, how often the player reached showdown. Typical is ~25–30%. High WTSD (33%+) is a station who chases and calls — value-bet thin, bluff almost never. Low WTSD (~20%) folds too much postflop — barrel them off marginal hands.
**W\(SD — Won money at ShowDown.** Of showdowns reached, how often they won the pot. Typical is **~48–54%**. Read it against WTSD: a player with high WTSD and *low* W\)SD is getting to showdown with garbage (a true station — print value); a player with low WTSD and high W$SD only shows down the nuts (their showdowns are strong, but they fold so much earlier that you bluff their non-showdown lines).
Stats hunt in pairs. VPIP is meaningless without PFR. Fold-to-c-bet matters because it sits next to your ability to c-bet. WTSD only makes sense beside W$SD. A single number is a rumor; a pair of numbers is a read. Train yourself to glance at the partner stat before you act.
19.3 Sample size: which stats you can trust, and when
This is where most HUD users light money on fire. Every stat needs a different number of hands before it means anything, because each has a different denominator and a different natural frequency. VPIP gets an opportunity nearly every hand, so it stabilizes fastest. 3Bet, fold-to-3bet, and especially river or turn stats get an opportunity rarely, so they stabilize slowest.
A practical hierarchy of how quickly stats become trustworthy:
| Stabilizes | Stats | Rough usable sample |
|---|---|---|
| Fast | VPIP, PFR | ~30–50 hands for a useful read; ~100 to trust |
| Medium | Steal, C-bet, Fold to C-bet | a few hundred hands |
| Slow | 3Bet, Fold to 3Bet, WTSD, W$SD, AF | many hundreds to low thousands |
| Very slow | Turn/river-specific, fold-to-double-barrel, check-raise | rarely trustworthy from in-session play alone |
The reason is opportunity frequency. You can see 80 hands against someone and have observed zero of their 3-bet opportunities that they chose to act on, leaving 3Bet displaying a misleading 0% or a wild 33% off one attempt. Postflop stats are worse: a player must (a) see a flop, (b) be in the right role, and (c) face the right action, so a “Fold to Turn C-bet” stat needs an enormous raw sample before its denominator is large enough to mean anything.
Acting on a small-sample postflop stat as if it were gospel. “His fold-to-c-bet is 100%, I’ll barrel” — over four opportunities. That is not a read; it is noise wearing a number’s clothing. Below a usable sample, fall back on population defaults (the previous chapter) and your range reading, and let the HUD only break ties. Configure your tracker to gray out or hide stats below a minimum sample so you are not seduced by a confident-looking 0% or 100%.
A useful mental rule: trust VPIP/PFR early, trust 3Bet/fold-to-3bet only once you have seen real volume, and treat any granular postflop stat as a tiebreaker, never a trigger unless you have thousands of hands on a regular you face constantly.
19.4 Turning each stat into an exploit
A HUD number is worthless until it changes a decision. Here is the translation table — the heart of this chapter. Each row is “when you see this, do that.”
| Read | Likely leak | Your exploit |
|---|---|---|
| High Fold to 3Bet (65%+) | Opens wide, folds to pressure | 3-bet light relentlessly, especially in position; weight toward non-value blockers |
| Low Fold to 3Bet (<40%) | Defends opens stubbornly | 3-bet only for value; flat your bluffy hands instead |
| High Fold to C-bet (60%+) | Over-folds flops | C-bet any two on dry boards; double-barrel scare cards |
| Low Fold to C-bet (<35%) | Floats and calls light | Cut bluff c-bets; value-bet thinner and bigger; give up turns with air |
| High WTSD + low W$SD | Calling station | Value-bet relentlessly (thin and large); never bluff multi-street |
| Low WTSD (~20%) | Over-folds postflop | Barrel multiple streets; attack with air on later streets |
| High Steal (BTN 45%+) | Attacks your blinds wide | Defend wider; 3-bet from the blinds; flat and float in position |
| High Fold BB to Steal (75%+) | Surrenders blinds | Open any two from the button/CO; size down to risk less |
| Passive AF/AFq (AFq <25%) | Bets/raises only with it | Fold to their aggression; bet your medium hands for value without fear of a raise |
| High 3Bet (11%+) | Light, bluffy reraiser | Call wider in position; 4-bet (or 4-bet bluff) more; flat and let them barrel |
| Big VPIP/PFR gap (e.g. 35/8) | Loose-passive caller | Value-bet wide and thin; almost never bluff; isolate their limps |
The two highest-EV HUD exploits at most stakes are dead simple: against high-fold-to-c-bet players, bet more; against calling stations (high WTSD / low W$SD), bluff less and value-bet more. You do not need fifteen stats to crush a typical online pool. You need to find the stations and stop bluffing them, and find the over-folders and never stop betting.
19.5 Popups, color-coding, and notes
The main panel should hold only a handful of stats — clutter slows you down and, across many tables, costs you decisions. Everything else lives one click away.
Popups are the expanded stat windows you open by clicking a player’s HUD. This is where the positional and street-specific breakdowns live: 3Bet by position (someone 3-bets 3% from early but 14% from the button — a very different read than the blended number), c-bet by board texture, fold-to-turn-barrel, check-raise frequency. In a key spot, pause and open the popup before a big decision; the blended top-line stat hides exactly the asymmetries that matter.
Color-coding turns the database into instant gestalt. Configure your tracker to tint a player’s name or stat by value — for example, red for high VPIP (recreational target), blue for tight-aggressive regs, a bright background for very high VPIP. The goal is that on a fresh table you can spot the fish before you read a single number, just from the colors. This is invaluable for table selection and seat selection: you want the loose-passive money on your right (so you act after them) and the dangerous LAG on your left under control.
Notes are where you store everything a stat cannot capture, and over time they become more valuable than the HUD itself. Good notes are specific and behavioral, not vague:
- Bad note: “bad player.”
- Good note: “Limp-called UTG with KQo 100bb, stacked off top pair on K-high turn-checked board. Overvalues top pair, no check-raise bluffs.”
- Good note: “Min-3bets only AA/KK so far (3 of 3). Over-folds turn to second barrel on broadway cards.”
Tag the showdowns that surprised you and the sizing tells you observed. A standardized shorthand (e.g., “L/C” for limp-call, “OF-T” for over-folds turn) keeps notes fast enough to take mid-session. Notes outlast samples: a stat decays toward the population mean over months of pool churn, but “stacks off light, no bluffs” is a behavioral truth that holds.
19.6 Online timing tells
The HUD is your structural read; timing is your real-time, hand-specific read. Online players reveal information through how long they take to act, and most of it is involuntary. Treat timing tells exactly like live physical tells — probabilistic, opponent-dependent, and worth far more against recreationals than against multi-tabling regs whose timing is contaminated by their other tables.
General patterns at typical online tempo (calibrate to the player and remember that auto-time-bank clicks and lag distort everything):
- Instant check — usually weakness or disengagement. He clicked check-fold or pre-selected “check”; he has given up on the hand and is not planning to fight for the pot. A bet often takes it down.
- Instant call — frequently a draw or a medium-strength hand that wants to continue but had nothing to think about. A snap-call rarely contains a monster (a monster considers raising, which takes time).
- Instant bet / snap-bet — often a pre-planned action: a made hand following through, or a draw firing a bet it decided on the previous street. Less likely to be a tank-and-bluff.
- The long tank, then bet — genuinely polarized and player-dependent. Against thinking regs a tank-then-large-bet can be a real value hand weighing sizing; against recreationals a long delay then a big bet skews toward strength (“Hollywooding” — exaggerated hesitation — is rarer online than live but does exist).
- The long tank, then check — often a real decision between betting and checking a medium hand, i.e., a genuinely marginal holding. Frequently exploitable by betting.
- Delay then minimum-size action — sometimes a disconnected player whose client auto-folded or who is distracted; flag it but do not over-read a single instance.
Reading timing tells off a multi-tabling regular. A grinder on 12 tables tanks because another table demanded a decision, not because your spot is hard. Their timing is noise. Timing tells are sharpest against players who are clearly on one or two tables and emotionally engaged — the recreational who just lost a big pot and is now taking three seconds to hero-call. Weight the read by how likely the delay is about your hand.
19.7 Information management while multitabling
A HUD is a firehose. One table, you can read every popup and think deeply. At twelve or twenty-four tables, the constraint is no longer information — it is attention. The skill becomes triage.
Principles for staying sane and profitable across many tables:
- Minimal main panel. Six to eight stats, no more: VPIP, PFR, 3Bet, Fold-to-3Bet, C-bet, Fold-to-C-bet, WTSD, and a steal stat cover most decisions. Everything else lives in popups for the rare deep spot.
- Color-code so reads are pre-attentive. When tables are popping, you cannot read numbers on all of them. Colors let you classify a table — fish present? regs only? — in a single glance and decide where to invest attention.
- Table-cap honestly. Add tables only up to the point where your worst decisions are still good. Most players are more profitable on fewer tables than ego admits; volume that degrades your win rate below break-even is negative volume. The marginal table that turns a thoughtful 3-bet-or-fold spot into a reflexive fold is costing you more than it earns in rake-back.
- Triage by pot size and difficulty. Let the small, standard spots play fast and near-default; spend your scarce thinking time on the big pots and the close decisions. A stack-off decision deserves a popup and a pause even if four other tables are blinking.
- Use the time bank deliberately. It exists for exactly the big-pot, multi-tabling collision where two tables demand a real decision at once. Burning a few seconds of time bank on the pot that matters is correct; do not let table count force a snap-fold in a stack-off.
- Default to sound baseline strategy. With limited attention, your fallback is not “guess” — it is the GTO-leaning, population-aware baseline from the earlier chapters. The HUD then refines that baseline only where a strong, well-sampled read justifies the deviation. Exploits are the seasoning, not the meal.
19.8 Worked example: assembling a read
You are 100bb deep, 6-max online cash. It folds to the cutoff, who opens to 2.5bb. You are on the button with A J. Before you act, you read the cutoff’s HUD: **VPIP 29 / PFR 22, 3Bet 7, Fold to 3Bet 68%, Fold to C-bet 41%, WTSD 31, W\(SD 47** — built over ~600 hands, so the preflop stats and steal/c-bet stats are trustworthy and WTSD/W\)SD are getting there.
Let us read it. A 29/22 cutoff is opening a fairly wide, aggressive range — wider than a nit, so A J is well ahead of his opening range and a clear continue. The decisive number is Fold to 3Bet 68% over a solid sample: he opens wide and then folds two-thirds of it to a re-raise. That is a glaring leak. A J also carries useful blocker value — an ace and a jack remove some of the very hands (AJ, AA, JJ, AK, AQ) that would continue against a 3-bet.
Decision: 3-bet to ~8bb. Two-thirds of the time he simply folds and we win 2.5bb + blinds uncontested with a hand that was never in trouble. The exploit is driven almost entirely by one well-sampled stat over a default flat-call.
Now suppose he calls. We see a flop of K 8 4 rainbow and he checks. His Fold to C-bet is 41% — middling, not a fold-button — and his WTSD 31 / W$SD 47 flags a tendency to get to showdown a bit too often with hands that aren’t winning, i.e., a mild station. So on this dry king-high board, where our 3-bet range hits the king hard and he is capped (he 4-bets or folds his kings preflop more often than he flats them), a small c-bet prints: he over-continues relative to how often he holds a king. But because he is showdown-bound, our plan against a call is value, not bluff — barrel A J only as a genuine value hand when we improve, and shut down the air on the turn rather than firing a second bluff into a player whose own stats scream “I call too much.”
That is the whole loop in one hand: the HUD generated the hypothesis (wide opener who over-folds to 3-bets but over-calls postflop), and the surrounding theory chose the action (3-bet light preflop for the fold, then value-lean postflop rather than bluff).
Over your next 1,000 online hands, run this routine. (1) Before every 3-bet decision against an opener with 200+ hands on them, glance at their Fold to 3Bet and note whether your action matched the read. (2) Tag every hand where you bluffed into a player with WTSD 30%+, and at session’s end count how many you won — most players are horrified to find this number near zero. (3) Pick your two most-faced regulars and write one specific behavioral note on each from a real showdown. Repeat until reading the partner-stat before acting is automatic, not effortful.
19.9 Summary
A HUD converts the anonymity of online poker into the richest read in the game — but only for a player disciplined about sample size and ruthless about turning numbers into actions. Read stats in pairs, never alone. Trust VPIP and PFR early, 3-bet and fold-to-3bet only with volume, and treat granular postflop stats as tiebreakers rather than triggers. Keep the main panel lean, push detail into popups, color-code so reads are instant, and let your notes capture what numbers cannot. Layer timing tells on top as a real-time signal, weighted by how engaged and single-tabling the opponent is. And above all, remember that the HUD is a hypothesis generator: it points at where an opponent likely deviates from sound play, and the exploitative judgment you have built in the rest of this book decides what to do about it. The numbers do not bluff — but they are still only probabilities, and the player who treats a 68% as a 100% will, eventually, pay for the other 32%.