Casino Comps Lift Rebuy Frequency 31% in 400-Session Logs

Four hundred logged sessions across 47 regular players at a mid-tier European-facing operator show a 31% increase in rebuy frequency when casino comps—free-play credits, cashback, and tier-point multipliers—are active versus inactive. The logs, captured between March and November 2024, control for session length, game type, and deposit method, isolating the comp effect from natural variance. The finding complicates the standard assumption that comps primarily extend session duration; here, they compress the interval between deposits, suggesting a distinct behavioral mechanism tied to perceived sunk value rather than time-on-device.

Method and Data Structure

The dataset originates from a voluntary telemetry program where participants agreed to share anonymized deposit timestamps, game selection, and comp accrual states. Each session was tagged with a binary comp flag: 1 if the player had an active comp balance exceeding €5 or a pending tier-point multiplier, 0 otherwise. Rebuy frequency was defined as the number of deposits made within 60 minutes of a session’s start, divided by total sessions in that comp state. A rebuy did not require a new game round; it merely marked a fresh monetary injection into the same sitting.

The 47 players were not high rollers. Median deposit size was €23, and median session length was 41 minutes. The group skewed toward slot-heavy play (72% of logged rounds), with the remainder split between blackjack and roulette. Crucially, the comp flag was not randomly assigned; players opted into comp-earning modes. To mitigate self-selection bias, the analysis matched each comp-active session with a comp-inactive session from the same player within a 14-day window, controlling for weekday and hour. This paired design reduced the raw 38% difference down to the reported 31%, still significant at p < 0.01 via a Wilcoxon signed-rank test.

One numerical anchor deserves emphasis: the median comp balance at rebuy time was €12.40, not the larger promotional bonuses often studied. This suggests the effect operates at a micro-incentive level, far below the threshold of a "free spin package" or "deposit match." Players were not chasing a jackpot-sized carrot; they were responding to a small, already-earned credit that made the next deposit feel less like a loss.

The Mechanism: Sunk Cost Inversion

Standard loss-chasing models predict that rebuys cluster after losses, as players attempt to recover. In these logs, the opposite pattern emerges under comp-active conditions. The average pre-rebuy loss in comp-active sessions was €8.20, compared to €19.70 in comp-inactive sessions. Players rebought earlier and with smaller deficits when comps were present. This inverts the classic sunk-cost narrative: instead of throwing good money after bad, they threw modest money after a perceived discount.

The comp balance appears to function as a mental accounting buffer. A €12.40 comp credit, when applied to a €23 deposit, reduces the effective stake to €10.60. The player is not depositing €23; they are depositing €10.60 plus converting a non-cash asset. Rebuy frequency rises because the psychological cost of the second deposit drops below a threshold that typically deters continuation. This aligns with prospect theory’s loss aversion asymmetry—the comp shifts the reference point from "I am down €20" to "I am down €8, but I have €12 in free play."

Notably, the effect decays within 90 minutes. Rebups occurring after the 60-minute window show no significant comp-related difference. This temporal boundary suggests the comp’s influence is tied to active session attention, not a lingering account balance. Players who log out and return later do not treat the comp as a live subsidy; it becomes inert ledger data. The comp works only when it is visually present and mentally linked to the game in progress.

Game-Specific Variance

The 31% aggregate figure hides substantial divergence by game type. Slot sessions showed a 44% rebuy lift under comp-active conditions, while blackjack sessions showed only 12% and roulette showed a non-significant 7%. The slot effect is not surprising given the game’s rapid feedback loops and frequent small wins, which pair naturally with comp accrual. Every 20-spin cycle that returns €0.50 in comp credits creates a micro-reinforcement schedule, effectively gamifying the deposit decision itself.

Blackjack’s weaker response likely stems from the game’s slower pace and higher skill salience. Players at the table track cards, dealer upcards, and basic strategy; a comp credit is background noise compared to the decision architecture. Roulette’s null result may reflect its low comp-earning rate—the house edge on a single-zero wheel generates comps at roughly half the speed of a 96% RTP slot. The comp balance simply never reaches the €12.40 anchor point during a typical roulette session.

This divergence has operational implications for comp allocation. Operators currently distribute comps uniformly across game categories, often as a flat percentage of turnover. The logs suggest that shifting a portion of comp spend toward high-frequency slot players would maximize rebuy lift, while maintaining table-game comps as retention tools rather than rebuy triggers. A 5% reallocation from blackjack comps to slot comps would, per this dataset, increase aggregate rebuy frequency by an estimated 6-8% without increasing total promotional cost.

Session Length and Deposit Intervals

The comp effect does not extend session duration—a counterintuitive result given the industry’s historical focus on time-on-device metrics. Median session length in comp-active states was 44 minutes, versus 43 minutes in comp-inactive states, a difference well within noise. What changed was the deposit cadence. The median interval between first and second deposit shrank from 22 minutes to 13 minutes under comp-active conditions. Players were not playing longer; they were committing more capital per unit of time.

This distinction matters because session length is a lagging indicator of engagement, while rebuy frequency is a leading indicator. A player who stays for 60 minutes without rebuying may simply be grinding a small balance. A player who rebuys at minute 13 has signaled an intention to continue, and their subsequent session is nearly 80% likely to include a third deposit if comps remain active. The logs show a compounding effect: each active comp dollar at rebuy time increases the probability of a subsequent rebuy by 0.9%, up to a ceiling of roughly four rebuys per session.

The practical takeaway is that comps should be timed to pre-rebuy moments, not session starts. A comp credit awarded at login has no measurable effect on rebuy frequency. A comp credit that appears after a player’s balance drops below €5, or after a 10-minute dry spell on slots, triggers the 31% lift. This suggests operators should experiment with dynamic comp issuance tied to loss thresholds rather than static accrual rates. The data does not yet show whether such dynamic issuance would be perceived as manipulative, but the behavioral response is clear.

Open Questions on Player Segmentation

The 47-player sample, while internally consistent, skews toward players who already engage with comp mechanics. Players who ignore comp balances entirely—estimated at 22% of the operator’s active base—showed no rebuy response, diluting any population-level effect. The 31% figure applies to comp-responsive players, not the average depositor. Scaling this finding requires identifying comp-responsive players early, perhaps through a two-session test where a €5 comp is offered after a small loss. Those who rebuy within 15 minutes are likely candidates; those who do not may be better served by other retention tools.

A more troubling question emerges from the data’s temporal decay. If comps lose their rebuy-triggering power after 90 minutes, what prevents players from habituating to comp presence over weeks? The logs cover only four months, and the comp-active sessions are spread unevenly across that period. No player maintained a comp-active state for more than 11 consecutive sessions, suggesting either natural churn or operator-side cap on comp accrual. Longitudinal studies are needed to determine whether the 31% lift persists beyond a player’s first 20 comp-active sessions, or whether it erodes as the comp balance becomes a normalized feature of the deposit environment. The answer will determine whether this is a sustainable lever or a short-term novelty effect—and that distinction carries very different implications for responsible gambling frameworks, which typically assume that any increase in deposit frequency is a risk factor rather than a neutral behavioral shift.