A 2% Table Cap Lowers Hourly Losses More Than Game Choice
A 2% table cap reduces expected hourly loss more than moving from a 2.5% house-edge game to a 0.5% one, provided the cap is enforced on total wagers rather than on buy-ins. For a player cycling $600 per hour, the difference is $12 versus roughly $6.60 in theoretical loss — the cap wins by nearly a factor of two, and it does so without requiring the player to learn a new game, chase a lower-variance variant, or accept worse comps.
That claim runs against the conventional advice that game selection is the primary lever a player controls. It is not. Table selection and table limits, which most players treat as fixed features of the environment rather than variables they set, dominate game choice once stakes are large enough that the cap actually binds.
The Arithmetic of a Cap
A table cap is a maximum on the house's theoretical win from a player's action over a defined period, usually an hour or a session. The mechanism is straightforward: the operator agrees to stop collecting after a threshold, either by comping back the excess, by restricting the player's total handle, or by adjusting the effective wagering requirement on a promotional credit.
Consider blackjack at 0.5% house edge with perfect basic strategy. At $25 per hand and 60 hands per hour, the handle is $1,500 and theoretical loss is $7.50. Add a 2% cap on total handle — meaning the player's exposure is limited to $1,000 of action per hour — and theoretical loss falls to $5.00, a 33% reduction. The player has not changed games, stakes, or strategy. They have changed the ceiling.
Now compare that to a game switch. Moving from a 2.5% edge game (say, a six-deck shoe with mediocre rules and no surrender) to a 0.5% edge game at the same $1,500 handle reduces theoretical loss from $37.50 to $7.50. That is a larger absolute reduction, but it requires the player to already be playing the worse game. Most players who care enough to read this far are not. They are already in the 0.5% to 1.0% band, where the remaining gains from game selection are small and the gains from capping handle are not.
The crossover matters. A 2% cap on a 0.5% game saves $2.50 per $1,000 of handle. Moving from 1.0% to 0.5% at the same handle saves $5.00. So game choice still wins at very low edges. But move to a 1.5% game — common in six-deck shoes with 6:5 payouts on blackjack, which now appear in roughly a third of Las Vegas Strip tables — and the cap saves $2.50 while the game switch saves $10.00. Game choice wins again.
The cap only dominates when the player cannot or will not switch games. That is most players, most of the time.
Why the Cap Binds More Often Than Players Expect
Caps are usually expressed as a percentage of handle, not of buy-in. A player who buys in for $500 and cycles it three times has a $1,500 handle. A 2% cap on that handle is $30. If the player's theoretical loss at a 1.0% game is $15, the cap never binds. If the player is at a 2.0% game, theoretical loss is $30 and the cap binds exactly at the limit.
This is why caps are more valuable to players in higher-edge games and to players who cycle their bankroll aggressively. A player who buys in for $500 and plays one hand has a $500 handle and a $10 cap — irrelevant. A player who plays 300 hands at $25 has a $7,500 handle and a $150 cap, which is meaningful at any edge above 2%.
The practical implication: caps reward volume, not stakes. A $5 player at 600 hands per hour has a $3,000 handle and a $60 cap. A $100 player at 60 hands per hour has a $6,000 handle and a $120 cap. The low-stakes grinder gets more absolute protection per dollar wagered, which is the opposite of how most players think about table limits.
Where the Cap Comes From
Table caps are not a standard casino product. They appear in three forms:
Promotional caps. A deposit bonus with a maximum cashout tied to wagering. The cap is implicit: the player's expected loss is bounded by the bonus terms, not by the game.
Table maximums. A $500 maximum bet on a $5 minimum table is a crude cap. It limits the player's ability to press into a high-variance sequence, but it does not limit total handle.
Voluntary caps. A player sets a loss limit or a session time limit. This is the only form the player fully controls, and it is the one that produces the arithmetic above.
The third form is the one worth modeling. A loss limit of $200 per session, combined with a 2% effective cap on handle, means the player stops at $200 regardless of game. The game choice determines how long that $200 lasts, not how much is lost. At a 0.5% edge and $25 hands, $200 lasts roughly 1,600 hands. At a 2.0% edge, it lasts 400 hands. The cap does not change the loss; it changes the duration.
That is the real trade. A cap converts game choice from a loss-reduction tool into a time-management tool. The player who caps losses and then chooses a low-edge game gets both: a bounded loss and a longer session. The player who chooses a low-edge game without a cap gets a longer expected session but no bound on the tail.
The Tail Matters More Than the Mean
Expected hourly loss is a mean. Players do not experience means; they experience sessions. A 2% cap truncates the upper tail of the loss distribution, which is where the damage lives.
At a 1.0% edge and $1,500 handle, the standard deviation of session loss is roughly $150 for a game with a 1.15 variance-to-mean ratio. A 2% cap at $30 cuts the mean but does not change the variance. The player still has a 16% chance of losing more than $180 in a session before the cap is applied. After the cap, that probability falls to roughly 2%.
That is the argument for caps over game choice. Game choice moves the mean. Caps move the tail. For a player with a fixed bankroll and a fixed tolerance for loss, the tail is what determines whether they can keep playing.
What This Means for Operators
If caps reduce player losses without reducing handle, operators should be indifferent or better off. The player who loses less per hour plays longer. The player who plays longer generates more handle. The operator's theoretical win per session may fall, but the number of sessions rises.
The empirical question is whether that substitution is one-for-one. It probably is not. A player who hits a cap and stops has a different return rate than a player who hits a loss limit and stops. The cap is a signal that the operator is managing the player's exposure; the loss limit is a signal that the player is managing their own.
That difference is testable. A casino that offers a 2% cap on handle to a cohort of mid-stakes players and compares retention, handle, and theoretical win against a control group would have the data within a quarter. The arithmetic says the cap should win. The behavioral question is whether players notice, and whether noticing changes how they play.