Rake CalculatorBETA

How this works

We ran each game structure through RocketSolver at the stack depth it is normally played at, testing many combinations of rake percentage and cap. Rake is money that leaves the table, so if you add up what every seat wins and loses, whatever has gone missing is what the house took. Divide that by the number of seats and you have what the average player pays. We then worked out a formula that matches those results, which lets the calculator handle structures we never ran.

Rake does not climb in a straight line as you raise the percentage or the cap. At 5% with a 0.20bb cap, the cap is reached once the pot hits 4bb, so beyond that point moving the rake to 6% changes nothing, because the house is already taking the most it can. Players also change how they play. When the rake is low they can afford to enter more pots, and those extra pots give back part of what the lower percentage saved them. The formula follows both effects.

The number covers the average seat and assumes everyone at the table plays close to perfectly at the depth we solved. A loose player pays more than the table average and a tight player pays less, so use this to compare one structure against another rather than to predict what your own database will show.

Finding your rake and cap

Open the Don't know your rake? lookup above, then choose your site, game, stakes and players dealt. It will fill in the published rake percentage and cap for you.

If your site or game is not listed, use its published rake table and enter the values manually. Select $ when the cap is quoted in money, and use the cap for the number of players dealt into the hand. Rake structures can change, so always confirm the values against the site's latest published table.

What the number leaves out

This is what the house takes, before anything it gives back. It does not account for:

• Rakeback
• Bonuses
• Leaderboards and promotions
• Jackpot and cash-drop deductions

Some sites deduct extra money from pots to fund player payouts, including GGPoker's Bad Beat Jackpot and CoinPoker's Splash Pots. The calculator leaves these deductions out because they cannot be estimated from the published rake percentage and cap, and the collected money is paid back to players. This is the main reason the calculator may show less rake than your database or tracker on those sites: a tracker can count the deduction as rake while recording the payout separately. Across the player pool, these programs are approximately EV-neutral.

To find what you actually pay, work out what your rewards are worth in bb/100 and subtract them. If the calculator says 5.0 and your rewards come to 1.2 bb/100, you are paying 3.8. That figure is the one that decides whether a game is worth sitting in, and Where Your Win Rate Comes From works through how it combines with the players around you.

Arriving at that 1.2 is easier on some sites than others. Plenty of reward schemes are vague on purpose, with tiers you climb at an unpublished rate, points that convert at a ratio the site can move, and prizes that land unevenly across a month. When the terms will not give you a number, measure one instead: total what you were actually paid over a stretch you have already played and divide by the hands. An estimate drawn from your own history is worth more than a precise figure from a page written to be hard to pin down. If you can manage neither, treat the calculator as your worst case.

The seven formats

• Standard 6-Max. Six-handed, 100bb, no ante. The house rakes only pots that see a flop.
• Standard 9-Max. Nine-handed, 100bb, no ante. A nine-handed table pays slightly more rake in total than a six-handed one, but it splits across nine seats, so each player pays about 30% less.
• HU Standard. Heads-up, 125bb, no ante. Heads-up games generally use lower caps than 6-max, which offsets some of the extra rake from playing two-handed.
• GGPoker 6-Max. Six-handed, 100bb, no ante, but the house also rakes 3-bet-or-larger pots that end before the flop. At a 5% rake that comes to about 25% more than standard 6-max.
• CoinPoker 6-Max Ante. Six-handed, 100bb, and every player posts a 0.16bb ante. The antes build bigger pots and pull more players into them, so it costs about 60% more than standard 6-max at the same rake.
• HU CoinPoker Ante. Heads-up, 125bb, and both players post a 0.16bb ante.
• WPTGold 8-Max Ante+Straddle. Eight-handed. Every player posts a 0.5bb ante, and one player posts a 2bb straddle. This puts 7.5bb in the pot before anyone acts. Players have 200bb stacks, equal to 100 straddles. Results are reported in original big blinds, which makes the rake look larger than it would in straddles.

Comparing rake across formats

Comparing WPTGold with a standard game in bb/100 needs extra care. A standard table has 1.5bb in the pot before anyone acts. WPTGold has 7.5bb: 4bb in antes, 1.5bb in blinds and a 2bb straddle. Each player starts with 200bb, which is 100 straddles. The calculator still reports the result in original big blinds, so the number is twice as large as it would be in straddles.

The antes and straddle create more dead money, which raises both the rake and the possible win rates. We did not test whether they rise at the same pace. Comparing rake structures inside one format stays reliable because the units and dead money are the same on both sides.

Table composition

A recreational player enters more pots and builds bigger ones, which lifts the table's average rake by about 20%. This setting rests on weaker evidence than the rest of the calculator: the all-regulars figure comes from the solver, while the 20% is a rule of thumb drawn from online hand databases and applied to the full-table formats. We leave heads-up at the solver estimate. Tables differ, so a tight recreational player moves the number by less than 20% and a very loose one moves it by more. The 20% is probably too high when the cap is small, because once a pot has hit the cap the house cannot take any more from it however big it grows.

What we tested

The 100bb formats were solved at rake percentages from 2% to 6%, with caps from 0.06bb up to 16bb. Both heads-up formats were solved at 125bb from 2% to 6%. Standard heads-up covers caps from 0.03bb to 12bb, while CoinPoker heads-up covers 0.12bb to 12bb. WPTGold was solved with 200bb stacks and a 2bb straddle, so each player had 100 straddles. The results are reported in original big blinds. We tested rake from 2% to 5%, with caps from 0.2bb up to 8bb.

Raising the cap keeps costing you long after it looks like it should stop. In standard 6-max at 5%, moving from a 4bb cap to a 16bb one costs a full big blind per 100 hands.

If a site quotes its cap in dollars, divide by the big blind, so a $3 cap at $2/$5 comes to 0.6bb. The $ toggle above does this for you.

For the full-table formats, the formula sits within 0.03 to 0.06 bb/100 of the solver on average, and its worst miss is 0.17. The standard heads-up curve misses by about 0.12 bb/100 on average, while the CoinPoker heads-up curve misses by about 0.24. These figures show how closely the formula tracks the solver, not how closely either one tracks your own results.

The formula

If you want to check the work, this is the whole model:

R is the rake per player in bb/100, p is the rake percentage and C is the cap in big blinds. K, H, S and B are chosen for each format to match the solver results, and all seven share the 0.56 exponent.