Every gambler keeps some version of a running total. Up $200, down $150, broke even over the weekend. Dollar figures feel concrete and honest, which is precisely why they're so misleading. Dollars won or lost are the output of a process, and evaluating a process by its output—without accounting for the inputs that produced it—gives you noise dressed up as information.
The two numbers that actually describe gambling performance are cost per decision and risk-adjusted return. Neither requires advanced math. Both require only that you stop letting raw dollar totals substitute for real analysis.
Why Dollar Results Mislead
Consider two players. Player A sits at a $10 blackjack table for three hours, plays basic strategy, and walks away down $40. Player B plays $25 slots for ninety minutes and walks away up $60.
By dollar result, Player B had the better session. By any meaningful performance measure, Player A did. The slot player ran lucky against a machine extracting 8–12% per spin at roughly 600 decisions per hour. The blackjack player ran unlucky against a game costing less than 0.5% per hand at 60 decisions per hour.
Dollar results can't tell you this. They absorb luck, game selection, bet sizing, and session length into a single figure that obscures all of them. You can't improve what you can't isolate.
Cost Per Decision: The Real Unit of Measurement
Cost per decision is exactly what it sounds like: the average amount you should expect to lose on each individual bet, given the game's house edge and your typical bet size.
The formula is straightforward:
Cost per decision = House edge × Average bet size
At a full-pay video poker machine (roughly 0.5% house edge) with a $1.25 five-coin bet, cost per decision is about $0.006—less than a cent per hand. At a roulette wheel (5.26% on a double-zero layout) with a $15 bet, it's $0.79 per spin.
Once you have cost per decision, multiply it by your pace—decisions per hour—and you have a reliable estimate of expected hourly cost:
Expected hourly cost = Cost per decision × Decisions per hour
The video poker player pushing through 600 hands per hour expects to lose roughly $3.75 per hour at optimal play. The roulette player at 40 spins per hour expects to lose about $31.60 per hour.
This number, not your dollar result after a session, is the baseline against which everything else should be measured. If you lost more than expected hourly cost × hours played, you ran bad. If you lost less—or won—you ran good. The direction and magnitude of variance become visible only when you have a cost baseline to measure against.
Risk-Adjusted Return: What You're Actually Getting for Your Exposure
Cost per decision handles the edge side of the equation. Risk-adjusted return handles the volatility side.
Two games can have identical house edges and identical expected costs per hour while exposing you to entirely different levels of swings. A $5 baccarat bet (house edge ~1.06%) and a $5 slot spin (house edge ~5%, but with a standard deviation per spin that can be 10–20× the bet) are not the same wager scaled by edge alone. The slot machine can produce sessions where you're up or down many multiples of your stake through variance alone—all while extracting more of your money on average.
Risk-adjusted return asks: how much volatility am I accepting per dollar of expected cost? The formal version is a ratio of expected loss to standard deviation, but you don't need precise standard deviation figures to use this concept practically. You need to know roughly where your game sits on the volatility spectrum and whether your bankroll can absorb swings at that level.
Here's the practical test: if the standard deviation of a typical session is large relative to your session bankroll, your actual results will be dominated by variance. The expected cost is still being extracted—it just gets buried under noise. Players in high-volatility games often feel like they're getting away with something when they win and getting robbed when they lose, because neither result reflects the underlying cost clearly.
Risk-adjusted return is lowest—best, from a player's perspective—in low-volatility games with low house edges. High-volatility, high-edge games produce exciting swings and extract money efficiently from players who mistake those swings for evidence that the game is beatable.
How to Apply Both Numbers Before You Sit Down
Before any session, run two quick estimates.
First, calculate expected hourly cost for the game you're considering: edge × average bet × decisions per hour. This tells you the minimum you should expect to pay for your time at the table, in the same way you'd estimate what any other form of entertainment costs per hour.
Second, compare that hourly cost against your session bankroll. If your expected hourly cost is $30 and you're sitting down with $100, you have a rough statistical runway of about three hours before expected losses alone exhaust your bankroll—before accounting for volatility, which can accelerate that timeline significantly in a high-variance game.
If the math produces a runway shorter than the session you're planning, you have three options: reduce bet size, choose a lower-edge game, or reduce expected decisions per hour by playing more deliberately or choosing slower games. What you should not do is ignore the calculation and hope the dollar result works out.
What Changes When You Use These Numbers
The immediate practical change is that you stop evaluating sessions by whether you won or lost money. You start evaluating them by whether you controlled cost per decision and managed volatility relative to your bankroll. A session where you lost $50 against an expected loss of $80 is a good session by this measure. A session where you won $120 against an expected loss of $15 required significant positive variance to produce that result and tells you very little about whether your approach is sound.
The deeper change is that game selection and bet sizing become primary decisions rather than background details. The house edge is fixed; the number of decisions you take and the size of each bet are within your control. Cost per decision and risk-adjusted return make the consequences of those choices legible in advance rather than only visible in hindsight through dollar results that mix skill, structure, and luck into a single uninterpretable number.
Dollar results answer the question of what happened. These two numbers answer the question of why—and why is the only figure worth tracking.