How Poker Rake Works: The Complete Data Guide

Rake is the single largest recurring cost in poker, and it is the one most players never measure. A recreational player who tracks bad beats obsessively will rarely be able to say how many big blinds per 100 hands they pay the house. Yet rake decides whether a break-even player is a small winner or a slow loser, it changes which hands are profitable to open, and it explains why the same strategy that prints money at higher stakes bleeds chips at micro stakes.

This guide explains how poker rake works across all three environments where it is collected: online cash games, tournaments, and live card rooms. It uses real, sourced figures throughout — WSOP fee schedules, published solver data on how rake shifts optimal play, and current Las Vegas drop structures — and translates each into the metric that actually matters: big blinds per 100 hands.

Key Findings

Online cash game rake is typically 5% of the pot with a cap that falls from around 4bb at micro stakes to under 1bb at high stakes, meaning micro-stakes players pay up to seven times more relative rake than high-stakes players. Documented tracker samples show players paying 10.8bb/100 in rake against a 13.8bb/100 pre-rake win rate. The 2025 WSOP Main Event returned 93.0% of its $97.35m in entries to the prize pool, and Las Vegas card rooms now commonly take 10% capped at $5–$6 plus a $1–$2 promotional drop.

What Rake Actually Is

Rake is the fee a poker operator charges for hosting the game. Unlike a casino table game, poker is played between players, so the house has no mathematical edge to profit from. It sells a service — dealing, security, software, liquidity, payment processing — and charges for it. Every profitable poker room in the world, online or live, is funded by rake.

Because poker is a negative-sum game once rake is applied, the total amount won by all players is always less than the total amount lost. The gap is the rake. Understanding its size is the difference between a realistic expectation and a fantasy one, which is why it belongs alongside pot odds and position in any serious poker fundamentals curriculum.

Three ways operators collect it

1. Pot rake — a percentage of each pot, capped at a fixed amount. Standard in online and most live cash games.
2. Tournament fee — a surcharge added to, or withheld from, the buy-in. Advertised as "$100 + $10" or withheld as a percentage of an all-in figure.
3. Time collection — a flat charge per player per half hour, common in high-stakes live games where pot rake would be punitive.

The Three Variables That Define Cash Game Rake

Comparing rake between two games is harder than it looks, because three separate variables interact:

1. The percentage

Online rake percentages cluster between 2.5% and 10%, with 5% by far the most common figure. Live rooms commonly run 10%. A higher percentage only matters in pots small enough that the cap is never reached — which, in practice, is most pots at low stakes.

2. The cap

The cap is the maximum rake taken from a single pot, expressed in currency but best understood in big blinds. This is where the real difference between stakes lives. A $3 cap is 6 big blinds at $0.25/$0.50 and 0.6 big blinds at $2.50/$5. The same nominal cap produces a tenfold difference in relative cost.

3. The drop rule

The drop rule determines when rake is taken at all. "No flop, no drop" means unresolved preflop pots are raked nothing. Most rooms also scale the cap by the number of players dealt in, taking less at heads-up and three-handed tables. A minority of operators additionally rake preflop-only pots such as 3-bet pots that end before a flop, which measurably increases total cost for aggressive players.

Typical online cash game rake by stake level

Stake (NLHE)	Typical rake %	Typical cap (6-max)	Cap in big blinds	Relative burden
NL10 ($0.05/$0.10)	5%	$0.40–$0.50	4.0–5.0bb	Highest
NL25 ($0.10/$0.25)	5%	$1.00–$1.25	4.0–5.0bb	Very high
NL50 ($0.25/$0.50)	5%	$2.00–$3.00	4.0–6.0bb	High
NL200 ($1/$2)	5%	$3.00	1.5bb	Moderate
NL500 ($2.50/$5)	5%	$3.00	0.6bb	Low
NL1000+ ($5/$10)	4–5%	$3.00–$5.00	0.3–0.5bb	Lowest

Figures represent common industry structures rather than any single operator's schedule; caps vary by table size and are usually reduced at short-handed tables. The pattern, however, is universal across the market: the cap in big blinds shrinks as stakes rise.

Weighted Contributed, Dealt, and Average Contributed

Once the total rake for a pot is calculated, the operator has to decide who paid it — a question that matters enormously for rakeback calculations and for players who fold often.

• Weighted contributed — each player is attributed rake in proportion to the money they put in the pot. Fold preflop and you are attributed nothing. This is the fairest and most common method, and it favours tight players.
• Dealt — rake is split evenly among everyone dealt into the hand, whether they played or not. This favours loose players and inflates the apparent rake contribution of nits.
• Average contributed — rake is split evenly among players who contributed anything at all, regardless of amount. A sit-in-between of the two.

The method matters most when a room's rewards programme pays back a percentage of "your" rake. Under a dealt model, a tight player receives credit for rake they did not economically pay; under weighted contributed, they do not. Neither model changes the total rake removed from the ecosystem — only its attribution.

Rake in Big Blinds: The Only Metric That Matters

Currency amounts are meaningless without context. Convert everything to big blinds per 100 hands (bb/100) and rake becomes directly comparable to your win rate.

Published analysis from solver developers illustrates the scale. At NL50, with a 5% rake and a cap of roughly 4 big blinds, the big blind is priced out of defending a large share of its range; at NL500, with the same 5% but a cap of just 0.6 big blinds, the same seat can profitably continue with meaningfully more hands. Solver output shows the big blind folding around 62% of the time at NL50 against roughly 59% at NL500 — a direct consequence of relative rake, not of opponent quality.

Modelled rake cost by stake

Stake	Approx. rake paid (bb/100)	Pre-rake win rate needed to earn 3bb/100	Realistic outcome for a break-even player
NL10	8–11	11–14bb/100	Substantial loser
NL50	5–7	8–10bb/100	Moderate loser
NL200	3–5	6–8bb/100	Small loser
NL500	2–3	5–6bb/100	Marginal loser

Ranges are modelled from published rake structures and typical pot sizes; individual results vary with table size, game type, and playing style. Looser, more aggressive players build bigger pots and therefore pay more rake in absolute terms.

Two documented real-world data points anchor these models. A widely cited tracker sample shows a player with a pre-rake win rate of +13.8bb/100 paying 10.8bb/100 in rake, leaving a post-rake win rate of +3.0bb/100 — rake consumed 78% of the edge. A second common illustration used in strategy education has a player beating the game for 7bb/100 while paying 5bb/100, netting 2bb/100. In both cases, more than half of all value generated at the table went to the house.

How Rake Changes Correct Strategy

Rake does not merely subtract from winnings; it changes which plays are correct. Because rake is only taken from contested pots, any line that wins the pot before rake is triggered — or that avoids building a raked pot with a marginal holding — gains value relative to an unraked model.

Solver comparisons of raked and unraked solutions show measurable behavioural shifts. In one published comparison, an unraked solution bets its trash hands 36% of the time on a given flop, while the equivalent raked solution bets trash 45% of the time. Rake increases the value of taking the pot down immediately, which pushes optimal strategy toward more aggression with low-equity hands and more folding from the big blind.

The practical consequences for low-stakes players are consistent:

1. Defend the big blind less. Marginal calls that break even in an unraked game lose money once the pot is taxed.
2. Open tighter from early position. Hands that realise equity through multi-street, multi-way pots pay the most rake.
3. Value bet thinner is not always right. Thin value that generates a large raked pot for a small edge can be net negative.
4. Steal more. Uncontested pots are the cheapest pots.

These adjustments are exactly why studying with rake-aware tools matters, and why a solver's default unraked output can mislead at micro stakes. Players working through GTO versus exploitative frameworks should treat rake as a structural input, not a rounding error, and check marginal spots with an equity calculator before assuming a call is close enough to make.

Tournament Rake: A Different Model

Tournaments charge a fee rather than a pot percentage, and that fee is disclosed up front. The convention is either explicit ("$100 + $10", where $10 is the fee) or withheld from a round buy-in, which is how most major live series operate.

WSOP fee structure by buy-in

Event buy-in	Entry fee withheld	Dealer/staff withholding	Total withheld	To prize pool
$10,000 Main Event	4.2%	1.8%	6.0%	94.0%
$1,000 events	7.0%	3.0%	10.0%	90.0%

The pattern is the same as in cash games: larger buy-ins are proportionally cheaper. A $1,000 entrant pays ten cents of fee per dollar of prize pool equity; a $10,000 entrant pays six.

The absolute sums are large. The 2025 WSOP Main Event drew 9,735 entries, representing $97,350,000 in gross entries, and produced an official prize pool of $90,535,500 — meaning 93.0% reached the players and roughly $6.8m was retained. That single tournament paid 1,461 places, with a $15,000 minimum cash and $10,000,000 to the champion. Across an entire summer of live events, WSOP rake has been reported at more than $40 million.

For tournament players, the practical implication is that fee percentage should be treated as a fixed drag on ROI. A player with a 20% ROI in a 10%-fee field would have roughly a 32% ROI in the same field with no fee. Field quality being equal, moving from high-fee to low-fee structures is one of the few free improvements available.

Live Poker Rake in 2026

Live rake is higher than online rake by a wide margin, because a live table deals roughly 25–30 hands per hour against 60–100 or more online, while carrying dealer wages, floor staff, and real estate costs.

Current Las Vegas structures illustrate the range. Of the 18 poker rooms operating in Las Vegas as of early 2026, most take 10% of the pot capped at $5, with several rooms having moved to a $6 maximum; one Strip room now takes 10% up to $6 plus a $2 promotional drop for a total of $8 per raked pot. Only three rooms in the city take no jackpot or promotional drop at all.

Live versus online rake burden

Environment	Typical structure	Hands per hour	Approx. rake per hour (per table)
Live $1/$2 NLHE	10% to $5 + $1–$2 drop	25–30	$120–$180
Live $2/$5 NLHE	10% to $6–$8 total	25–30	$150–$210
Online NL50 (6-max)	5% to $2–$3	60–100	$45–$90

Live figures assume most pots reach the cap, which is typical at $1/$2 and above, and exclude dealer tokes — an additional real cost of roughly $1 per pot won.

At a live $1/$2 game, $150 per hour leaving a nine-handed table means the average seat loses roughly $16.50 per hour to the house before anyone plays a hand well or badly. That is the mathematical reason small-stakes live poker is so difficult to beat, and why the games remain playable only because the average opponent's error rate is high enough to overcome it.

Rakeback: How Much Comes Back

Most operators return a portion of rake through rewards programmes, and the effective rate is a legitimate part of any comparison between online poker sites. Reward structures generally take one of three shapes:

• Flat rakeback — a fixed percentage of attributed rake returned, typically 20–40%, usually via an affiliate or VIP arrangement.
• Tiered loyalty — a percentage that scales with volume, often starting near 5–10% and reaching 30%+ only at high volume.
• Points-for-prizes — points redeemable for tournament tickets, bonuses, or merchandise, where the effective cash value is opaque and frequently lower than the headline rate implies.

The arithmetic is worth doing explicitly. A player paying 6bb/100 in rake who receives 30% back is effectively paying 4.2bb/100. For a player whose pre-rake edge is 7bb/100, that is the difference between a 1bb/100 win rate and a 2.8bb/100 win rate — nearly a tripling of net earnings from a change that has nothing to do with how well they play. Terminology around these programmes is covered in more depth in our poker glossary.

Methodology

This guide combines three categories of evidence. Published fee schedules — WSOP entry fee and staff withholding percentages, official 2025 Main Event prize pool and entry figures, and Las Vegas room drop structures reported in 2026 — are taken directly from operator announcements and industry reporting and are stated as published. Solver-derived strategy effects — big blind fold frequencies at NL50 versus NL500, and raked-versus-unraked betting frequencies with weak holdings — are drawn from published comparisons by solver developers.

The third category, modelled rake burden by stake, is our own calculation. We took the common industry structure of 5% rake with the caps shown in the first table, applied typical pot-size distributions for six-max no-limit hold'em, and converted results to big blinds per 100 hands. These are ranges, not precise measurements: actual rake paid depends on table size, VPIP, aggression, and the specific drop rule in force. Where a figure comes from a single tracker sample rather than a population study — most notably the 13.8bb/100 pre-rake versus 10.8bb/100 rake example — we label it as an individual data point and do not generalise it. No figure in this article has been estimated where a sourced figure exists.

Frequently Asked Questions

What is a normal poker rake percentage?

Online cash games almost always use 5% of the pot, with some formats between 2.5% and 6%. Live card rooms typically take 10%. The percentage alone is not a meaningful comparison, because the cap determines how often that percentage is actually reached — a 10% rake capped at $5 can be cheaper than a 5% rake capped at $10 in large pots.

How much rake does the average player pay per 100 hands?

Modelled from standard structures, roughly 8–11bb/100 at NL10, 5–7bb/100 at NL50, and 2–3bb/100 at NL500. Individual results vary substantially with playing style: loose-aggressive players build larger pots and pay more rake in absolute terms than tight players at the same stake.

Does rake make micro stakes unbeatable?

No, but it raises the bar considerably. At NL10, a player needs a pre-rake win rate above roughly 11bb/100 to clear 3bb/100 net. That is achievable against weak opposition but leaves little margin for error, which is why moving up in stakes — where the cap in big blinds shrinks — often improves results even against stronger opponents.

Is tournament rake higher or lower than cash game rake?

Tournament fees are lower per dollar risked but are charged on the full buy-in regardless of how long you last. A $1,000 event with 10% total withholding takes $100 whether you bust in level two or win. Cash game rake is charged only on pots that are actually played, so a player who folds often pays less.

How do I calculate what rake is really costing me?

Use a tracking program to display total rake paid alongside your win rate, and express both in bb/100. If your database shows a pre-rake win rate and a post-rake win rate, the gap is your true cost. Then check your rewards programme's effective return rate and subtract it. Detailed worked examples appear throughout our poker strategy library.

Sources

• WSOP — 2025 Main Event Official Prize Pool and Total Entries
• PokerNews — World Series of Poker Rake and Statistics
• GTO Wizard — Rake & Rakeback Explained
• GTO Wizard — Customizable Raked Solutions
• Vegas Advantage — Las Vegas Poker Room Rake and Drop Structures
• PokerNews — Comparing Rake and Comps at Las Vegas Poker Rooms

Cite This Article

If you use data from this article, please link back to https://www.deucescracked.com/blog/how-poker-rake-works-complete-data-guide — DeucesCracked, "How Poker Rake Works: The Complete Data Guide" (2026).