How rakeback affects your poker winrate: the math, variance and hidden profit

Whether you treat poker as entertainment or as a source of income, your bottom-line financial result comes down to three factors:

• your edge over your opponents — your winrate;
• the rake structure (the commission) at the poker table;
• your rakeback percentage — the share of the paid commission that is returned to you.

I. Understanding winrate and the impact of variance

Your edge over the field, expressed as a winrate, is the hardest of the three to pin down. It is shaped by a host of constantly shifting factors, the most important being the strength of the opponents at your table. On top of that, over short samples the impact of variance is enormous: a statistically meaningful winrate can only be estimated after tens of thousands of hands played.

Winrate is conventionally measured in big blinds per 100 hands (bb/100). In tracking software it is most often calculated before the commission is taken (pre-rake). If you don't know your exact winrate, you can roughly estimate it using this classification:

• A negative winrate — you play worse than your opponents at your stake;
• 0 bb/100 — a break-even game, on par with the average opponent;
• 1–10 bb/100 — a moderate edge over the field;
• 10–20 bb/100 and above — a significant edge (usually from regularly playing against weak opponents).

II. Standard deviation and the math of the long run

To analyze the volatility of your results, we use standard deviation (SD). For No-Limit Hold'em (NLH) cash games it runs at roughly 100 bb/100, while in 5-card Pot-Limit Omaha (PLO5) it climbs to 200–250 bb/100 because of the higher density of strong hands.

It is critical to understand one fundamental difference: winrate grows linearly with the number of hands played, whereas standard deviation grows in proportion to the square root of the sample.

The formula for standard deviation over an arbitrary sample looks like this:

where N is the number of 100-hand intervals played. This is an approximation: it assumes the 100-hand segments are independent, but for a practical estimate that is perfectly adequate.

A worked example (before rake)

A sample of 10,000 hands in Hold'em is the volume an active player racks up in a week on 1–2 tables. Since winrate and SD are measured per 100 hands, we treat this sample as 100 intervals (N = 100).

First, let's calculate the standard deviation for the entire sample:

Now let's compare two players (winrate here is before rake):

• A winning regular, +10 bb/100. His expected value over the sample is +1000 BB.
• A losing recreational player, −10 bb/100. His expected value is −1000 BB.

By the confidence-interval rule for the normal distribution:

For the winning regular, there is a 68% probability that his result over 10,000 hands lands in the range:

In other words, even with an outstanding winrate of 10 bb/100 there remains roughly a 16% chance (half of the remaining 32%) of finishing the week in the red. And conversely: the losing player has, with the same 68% probability, a result between −2000 and 0 BB — meaning he too has about a 16% chance of closing the sample in profit. It is precisely this mathematical quirk that keeps recreational players in the poker ecosystem.

But all of this is the picture before rake is deducted. What the commission does to it comes in the next section.

III. Rake load: the silent profit killer

Rake is the commission the poker room takes from every pot that reaches the flop (in some rooms rake is also taken preflop — on a 3-bet-fold scenario). The load is defined by two parameters: the rake percentage and its cap (expressed in BB). The higher these figures, the harder it is to turn a clean profit. If you need the basics — what rake and rakeback are and how they work — start with the guide "Rake and rakeback in poker".

Let's take standard parameters for mid stakes: 5% rake with a 5 BB cap. Depending on playing style (tight or loose), over 100 hands a player pays on average about 10 BB in rake.

Back to our regular. Over 10,000 hands (100 intervals) the rake eats up:

That is exactly all of his expected value. The entire distribution shifts down by 1000 BB: net expected value becomes zero, the real range turns into [−1000; +1000] BB, and the chance of finishing the week in the red jumps from 16% to nearly 50% — effectively a coin flip. Put differently, an excellent technical edge of +10 bb/100 before rake, once the commission is deducted, yields a net winrate of exactly 0 bb/100. This is exactly where the pressing need for effective rakeback arises.

IV. Rakeback as the foundation of financial stability

Rakeback is the return of part of the commission you paid back into your playing account. There are several models:

• Progressive (centralized rooms): the return starts at 10% and can theoretically reach 60–70%. But to reach the high tiers you have to grind an enormous sample while paying commission at a reduced rate — and that erodes your starting expected value.
• Static (decentralized club apps): clubs offer a fixed, transparent return of 20% to 50% from day one.
• Agent-based (for example, NUTS POKER): clubs grant large affiliate terms to agents, and the agents share the rakeback directly with players — fixed deals up to 70%. In this model the agent also acts as a guarantor for the safety of your funds and the payout of your winnings (except in cases of platform rule violations — prohibited software, bots, or team play).

V. A practical example: the price of the right choice

Back to our regular. He pays 10 BB in commission for every 100 hands. With a quality 70% rakeback deal he gets 7 BB back for every 100 hands — which lifts his net winrate from 0 to +7 bb/100. Laid out clearly, the whole chain looks like this:

Over a sample of 10,000 hands the return amounts to 700 BB in clean profit: net expected value swings back into the positive (+700 BB), the distribution shifts upward, and the chance of finishing the week in the red drops from ~50% back to about 24%. Even a recreational player who is losing will recover that same 700 BB and noticeably cut his losses.

If these players are grinding the NL500 stake ($2/$5, where 1 BB equals $5), the weekly payout of net rakeback comes to:

This is exactly why, when building a long-term poker career, you simply cannot ignore rakeback terms. Playing in rooms with a heavy rake load and no commission return is a guaranteed way to deprive yourself of the profit you've earned. You can compare current rakeback deals in the PokerDealsHub catalog.

Frequently asked questions

What is bb/100 winrate and what counts as a good winrate?

Winrate is the average profit in big blinds per 100 hands (bb/100), usually measured before rake is deducted. 0 bb/100 is a break-even game, 1–10 bb/100 is a moderate edge, and 10–20 bb/100 and above is a strong edge, typical of playing against weak opponents.

Why can you lose over a week even with a +10 bb/100 winrate?

Because of variance. Over 10,000 hands the standard deviation equals 1000 BB, so even with a winrate of +10 bb/100 before rake (EV +1000 BB) the result has a 68% probability of landing in the range [0; 2000] BB — about a 16% chance of going into the red. And once rake is deducted, net expected value falls to zero, and without rakeback the chance of finishing the week in the red is already nearly 50%.

How does rakeback affect your bottom-line profit?

Rakeback returns part of the commission you paid. With rake of about 10 BB/100 and a 70% deal, a player gets back 7 BB/100 — which lifts net winrate from 0 to +7 bb/100, that is 700 BB over 10,000 hands, or $3 500 at the NL500 stake. Often it is rakeback that turns a break-even result into a steady profit.

What counts as high rakeback?

In centralized rooms, high return rates (60–70%) require enormous volumes of play. Club apps give a fixed 20–50% right away, and agent-based deals offer up to 70% from day one. Current offers are gathered in the PokerDealsHub catalog.