Cash game profitability is usually discussed as a single number: win rate. A player at 5bb/100 is doing better than one at 2bb/100. As a measure of poker skill, that is true. As a measure of income, it is incomplete.
Real profitability is a function of several interacting variables: win rate, volume, rake paid, rakeback percentage, game quality, variance and how much your performance deteriorates as you add tables. That creates a central optimisation problem in online poker, volume versus precision, and the answer changes depending on how much of your rake comes back.
The right variable is total expected value
Win rate measures the quality of your decisions per hundred hands. It does not measure what you earn per hour.
Consider two players at the same stake:
| Tables | Win rate | Hands/hour | Hourly EV | |
|---|---|---|---|---|
| Player A | 4 | 4 bb/100 | 300 | 12 bb |
| Player B | 8 | 2.5 bb/100 | 600 | 15 bb |
Player B has a visibly worse win rate and makes more money. That is the basic argument for volume. Poker pays for aggregate expected value, not for the elegance of individual decisions.
Hourly EV = hands per hour × EV per hand + rakeback per hour
Maximising EV per decision and maximising EV per hour are different problems. Rakeback shifts the optimal point between them.
Why precision declines with volume
Extra tables are not free. They consume attention. At four tables you can reconstruct ranges, notice unusual sizings, take notes and find thin value. At twelve you are running on heuristics, pre-flop systems and pattern recognition. Marginal spots get compressed into simple rules, and exploitative adjustments disappear.
The result is a volume-induced win-rate decay curve. A strong regular might see something like:
| Tables | Win rate |
|---|---|
| 4 | 5.0 bb/100 |
| 6 | 4.5 bb/100 |
| 8 | 3.5 bb/100 |
| 10 | 2.0 bb/100 |
| 12 | 0.5 bb/100 |
The decline is rarely linear. Cognitive overload has thresholds. The economically relevant question is not how many tables you can play but at what point the marginal value of extra hands drops below the EV lost through worse decisions.
Rake is a tax on volume; rakeback is a subsidy
Rake makes poker a negative-sum game for the players collectively. Your observed win rate is already net of it:
Gross edge − rake paid = net table win rate
Rakeback returns part of that rake. If you pay 8bb/100 in rake and receive 25% back, that is worth about 2bb/100. A player showing 1.5bb/100 at the tables is effectively earning 3.5bb/100. At 50% rakeback the same rake produces 4bb/100 in returned value.
This is why rakeback behaves mathematically as a volume subsidy. It is not free money, since you still lose the share that is not returned, but it reduces the cost of participation and raises the value of every additional hand.
How the percentage changes the equation
At 0% rakeback, your decisions carry the entire economic burden. Precision dominates. Doubling volume while halving win rate is a bad trade unless the hourly maths clearly works.
At 10 to 20%, volume becomes somewhat more valuable, but the basic structure holds. Table selection and decision quality still beat mass multi-tabling.
At 30 to 40%, things change materially. Suppose Player A pays 7bb/100 in rake, plays 300 hands an hour and wins 4bb/100 at the tables. With 40% rakeback the rewards add 2.8bb/100, for 6.8bb/100 total and 20.4bb an hour. Player B doubles volume to 600 hands an hour and drops to 2bb/100 at the tables. Rakeback still adds 2.8bb/100, for 4.8bb/100 total. Per hundred hands B looks worse. Per hour, B makes 28.8bb, roughly 41% more.
At 50% and above, the incentive for volume grows stronger still, provided the underlying strategy remains at least modestly profitable.
There is a hard boundary. Rakeback can compensate for a lower win rate. It cannot compensate for losing poker. If extra tables turn you from a small winner into a large loser, rakeback only slows the bleeding. The number that matters is your effective all-in win rate after rewards.
The hidden variable: variance
Volume has a second benefit: faster statistical convergence. A player logging 100,000 hands a month learns about their strategy five times faster than one logging 20,000. Leaks show up sooner. Changes in game quality are easier to spot.
But volume that degrades decision quality produces a bigger sample of worse decisions. The professional objective is high-quality scalable volume: a table count where decision quality has started to dip slightly but has not collapsed.
Bankroll considerations follow. A player winning 0.5bb/100 at the tables and relying on rewards has far less margin over break-even than one winning 5bb/100. Both can be profitable, but the first carries deeper, longer downswings and needs a bigger bankroll and more conservative assumptions.
Maximise EV per hour, not hands
Imagine testing table counts and recording effective win rate including rakeback:
| Tables | Effective win rate | Hands/hour | Hourly EV |
|---|---|---|---|
| 4 | 7.0 bb/100 | 300 | 21.0 bb |
| 6 | 6.0 bb/100 | 450 | 27.0 bb |
| 8 | 4.8 bb/100 | 600 | 28.8 bb |
| 10 | 3.0 bb/100 | 700 | 21.0 bb |
Eight tables is the optimum. It has neither the best win rate nor the highest volume. It has the highest product of the two. Higher rakeback pushes that optimum toward more tables.
This has a training implication. Low-volume players should keep sharpening precision, because each improvement affects a small number of hands. High-volume players should work on strategic compression: internalised ranges, instant responses to common textures, automatic reactions to sizing. Precision and volume are not permanently opposed. Training converts precision into automation, and a player who once held 5bb/100 at four tables and 2bb/100 at eight can, with work, hold 4bb/100 at eight.
Conclusion
The central mistake in the volume versus precision debate is treating win rate as the final objective. It is one input into an economic system.
With little rakeback, protect your table win rate because nearly all your profit comes from opponents. As rakeback rises, the value of volume rises with it, and accepting a moderate decline in win rate for a large increase in hands can be mathematically correct. The key word is moderate. Rakeback should amplify a viable strategy, not rescue an unprofitable one.
The most profitable player is not the most precise or the most prolific. It is the one who finds the highest sustainable point on the volume-precision curve, and as rakeback increases, that point moves toward volume.
