Gambling odds and probability basics
Gambling odds show the price of a bet, while probability shows how often the event should happen. The house edge appears when the payout is shorter than the true odds.
A bet is priced, not predicted
Casino gambling odds are a price, not a prediction. On a single-zero roulette straight-up bet, the true odds are 36:1, but the table pays 35:1. That missing unit is where the house edge sits.
Probability in gambling tells you how often an event should occur over a long run. Payout odds tell you what the game pays if it occurs. If payout odds are shorter than true odds, expected value is negative. That does not mean every session loses.
It means the average result over repeated bets favours the rules, not the player. The useful question is not whether a bet can win once. It can. The useful question is what the price says about repeated play. The wider framework sits in the gambling maths guide.
Probability and odds are not the same thing
Probability: how likely
Probability is the share of all possible outcomes that win. A two-dice total of 7 has 6 winning ways out of 36, so its probability is 16.67%. A single number on a single-zero roulette wheel has a probability of 2.70%.
Odds: the ratio
Odds are a ratio, not a percentage. Odds against compare losing outcomes with winning outcomes. Odds for reverse that order. This is why 1 in 37 and 36:1 against can describe the same roulette event while looking completely different. The wording matters because players often read a ratio as if it were a chance.
Odds for vs odds against
Most gambling prices are payout odds. They tell you the reward, not the true chance. Implied probability is the break-even chance suggested by the payout. If the real probability is lower than the implied probability, the bet has a negative expectation.
Conversion | Formula | Worked conversion |
|---|---|---|
Probability to odds against | (1 - p) : p | A single-zero roulette number at 2.70% is 36:1 against |
Odds against to probability | Winning side ÷ total sides | Rolling a 7 with true odds of 5:1 against is 16.67% |
Payout odds to implied probability | 1 ÷ decimal odds | A 35:1 payout is decimal 36.00, implying 2.78% |
Decimal odds to implied probability | 1 ÷ decimal odds | Decimal 1.91 implies 52.36% |
Moneyline to implied probability | Price converted to break-even chance | Moneyline -110 implies 52.38% |
Moneyline to decimal | Positive price converted to total return | Moneyline +150 converts to decimal 2.50 |
The four odds formats and how to convert them
Fractional
Fractional odds show net profit compared with stake. Fractional 5/1 has implied probability of 16.67%. A fraction such as 7/2 is profit-to-stake notation: the first figure is the profit side and the second is the stake side.
To compare prices such as 4/1, 9/2, 10/3, 5/2 and 3/1, put them into one format first. The higher price gives the larger payout for the same stake, but it does not prove the event is more likely.
Decimal
Decimal odds include the returned stake. Fractional 1/1 is decimal 2.00. Decimal 4.00 has implied probability of 25.00%. Formula: decimal odds = net profit odds plus stake. Implied probability = 1 ÷ decimal odds. Decimal format is often the cleanest way to compare unlike fractions because every price becomes a total return per stake unit.
Moneyline (American)
Moneyline odds use plus and minus signs. A plus sign marks a positive price, while a minus sign marks an odds-on price. American odds +300 convert to decimal 4.00. American odds -200 imply 66.67%. A price of +200 is the same payout shape as 2/1, which is decimal 3.00 and implies 33.33%.
Implied probability %
Implied probability turns a price into the chance required to break even before considering whether the true chance is lower. It is not the same as the real probability. On a single-zero roulette red bet, the payout implies 50.00%, but the actual probability is 48.65%.
Event or payout example | Fractional | Decimal | Moneyline | Implied probability | Actual probability |
|---|---|---|---|---|---|
Coin landing heads | 1/1 | 2.00 | +100 | 50.00% | 50.00% |
Red on a single-zero wheel | 1/1 | 2.00 | +100 | 50.00% | 48.65% |
Dozen on a single-zero wheel | 2/1 | 3.00 | +200 | 33.33% | 32.43% |
Straight-up roulette number, single-zero wheel | 35/1 | 36.00 | +3500 | 2.78% | 2.70% |
Split bet, single-zero wheel | 17/1 | 18.00 | +1700 | 5.56% | 5.41% |
Rolling a 7 with two dice, if paid 4/1 | 4/1 | 5.00 | +400 | 20.00% | 16.67% |
True odds vs payout odds: where the house edge lives
Roulette worked example
True odds vs payout odds is the core split. On a double-zero wheel, a straight-up number has true odds of 37:1 against, but the payout is 35:1. The house edge is 5.26%. On a single-zero wheel, the same payout meets true odds of 36:1 and the house edge is 2.70%.
The payout looks almost fair because the missing part is small on one spin. Repetition makes it visible. The table is not trying to predict which number appears next. It is paying less than the fair price whenever the chosen number does appear. For a full roulette breakdown by bet type, see the roulette odds and payouts table.
Craps worked example
The same pattern appears in dice games. The craps any seven bet pays 4:1, while rolling a 7 with two dice has true odds of 5:1 against. That gap produces a 16.67% house edge. A craps place 6 or 8 bet pays 7:6 and carries a 1.52% house edge.
The general formula
If true odds against are L:W and the payout is P:W, the edge comes from paying P instead of the fair L. In ratio form, house edge = (L - P) ÷ (L + W), expressed as a percentage of the stake.
This only works cleanly when the payout and the true odds are written on matching stake units. If they are not, convert them first, then compare.
Bet | True odds | Payout odds | House edge |
|---|---|---|---|
European roulette straight-up number | 36:1 | 35:1 | 2.70% |
American roulette straight-up number | 37:1 | 35:1 | 5.26% |
European roulette red or black | 19:18 | 1:1 | 2.70% |
American roulette red or black | 20:18 | 1:1 | 5.26% |
Craps any seven | 5:1 | 4:1 | 16.67% |
Craps place 6 or 8 | – | 7:6 | 1.52% |
Calculating the probability of a casino event
A single event
For one event, divide winning outcomes by all outcomes. With two dice, a total of 7 has 6 ways out of 36. The arithmetic is 6 ÷ 36 = 16.67%. For a single-zero roulette number, the rounded probability is 2.70%, while the same chance written more precisely is 2.7027%.
Several events in a row
For independent events in a row, multiply the probabilities. A single-zero roulette red result has probability 48.65%. Three reds in a row are calculated as 48.65% × 48.65% × 48.65% = 11.514%. Five reds in a row have probability 2.725%. The streak may look unusual, but it is still just multiplication of independent events.
At least one (the complement rule)
For at least one hit, calculate the chance of no hit, then subtract it from certainty. Missing a chosen single number on every spin of 37 spins on a single-zero wheel has probability 36.29%. So hitting it at least once is 63.71%.
The same method works with dice. The probability of rolling no 6 in four rolls of one die is 48.23%. So the probability of rolling at least one 6 in those rolls is 51.77%. This is easier and safer than trying to add separate hit patterns.
Rounding matters. The same single-number roulette chance can be written as 2.7027% or 1 in 37. Use the exact rate for the calculation, then round the final result. The phrase 1 in 37 is useful for reading, but it hides the steps when you multiply several events or use the complement rule.
Independent vs dependent events
The roulette wheel has no memory
Independent events do not remember previous outcomes. A roulette spin does not become more likely to hit red because black has just appeared. Slot RNG rounds follow the same independence principle. This is the error covered in the guide to the gambler's fallacy.
The blackjack shoe does
Dependent events change as cards are removed. In an infinite-deck model, the dealer showing a ten-value card has probability 30.77%. In a real shoe, exposed cards change the composition. That can affect decisions, but it does not turn any guess into an advantage.
Game | Independent or dependent | What it means for strategy |
|---|---|---|
Roulette | Independent | Previous spins do not change the next spin |
Slots | Independent | Past results do not make a payout due |
Blackjack shoe | Dependent | Card removal changes composition and can affect correct play |
Baccarat shoe | Dependent | Card removal exists, but normal betting choices remain rule-bound |
Reference table: odds for common casino bets
This table gives casino odds in the form most useful before betting: probability where the event is specific, payout where the price is fixed, true odds where they are defined, and house edge where the game maths gives one.
No row matches your search.
Bet | Probability | Payout | True odds | House edge |
|---|---|---|---|---|
European roulette straight-up number | 2.70% | 35:1 | 36:1 | 2.70% |
European roulette split | 5.41% | 17:1 | 35:2 | 2.70% |
European roulette corner | 10.81% | 8:1 | 33:4 | 2.70% |
European roulette dozen or column | 32.43% | 2:1 | 25:12 | 2.70% |
European roulette red, black, odd or even | 48.65% | 1:1 | 19:18 | 2.70% |
American roulette straight-up number | 2.63% | 35:1 | 37:1 | 5.26% |
American roulette red, black, odd or even | 47.37% | 1:1 | 20:18 | 5.26% |
French roulette even-money bet with la partage | – | 1:1 | – | 1.35% |
Blackjack, basic strategy model with 3:2 blackjack | – | – | – | 0.48% |
Baccarat banker, 8-deck, 5% commission | – | – | – | 1.06% |
Baccarat player, 8-deck | – | – | – | 1.24% |
Baccarat tie, 8-deck, pays 8:1 | – | 8:1 | – | 14.36% |
Baccarat tie, 8-deck, pays 9:1 | – | 9:1 | – | 4.84% |
Craps pass line or come | – | – | – | 1.41% |
Craps don't pass or don't come, 12 pushes | – | – | – | 1.36% |
Craps place 6 or 8 | – | 7:6 | – | 1.52% |
Craps any seven | 16.67% | 4:1 | 5:1 | 16.67% |
Sic Bo small or big, loses on any triple | – | 1:1 | – | 2.78% |
Big Six wheel $1 segment, 54-segment wheel | – | 1:1 | – | 11.11% |
Big Six wheel $10 segment, 54-segment wheel | – | 10:1 | – | 18.52% |
Slot or pay-table game at 96.5% RTP | – | – | – | 3.5% |
Read the table across the row, not by one column alone. A frequent hit rate can still be attached to a negative price, and a large payout can still be too short for the true chance. For return-based games, use the RTP guide before comparing them with table-game bets. A pay table or exact return figure is needed before video poker can be ranked honestly.
Five mistakes people make reading odds
Most odds mistakes come from mixing up chance, price and long-run return. The corrections below are simple, but expensive to ignore.
Mistake | Why it is wrong | Correct reading |
|---|---|---|
Reading odds against as probability | 36:1 against is a ratio, not a percentage | Use probability for likelihood and odds for ratios |
Treating RTP as a session promise | A 96% RTP slot has expected value of -$0.04 per $1, not a guaranteed short-term return | Expected loss on $1,000 total wagered at 96% RTP is $40 |
Confusing frequency with payout size | Red wins more often than a straight-up roulette number, but both single-zero bets carry 2.70% house edge | Compare edge, not only hit rate |
Assuming a 1 in 37 event appears once every 37 tries | A chosen single number can miss every spin of 37 spins with probability 36.29% | Use the complement rule for at least one hit |
Adding probabilities that should be multiplied | Three independent red results in a row are not found by addition | Multiply: 48.65% × 48.65% × 48.65% = 11.514% |
The first mistake is a language problem. Odds against tell you how many losing outcomes stand against the winning outcomes. Probability tells you the share of all outcomes that win. Moving between the two is simple, but skipping the conversion leads to bad comparisons.
The second mistake is a timing problem. RTP and house edge are long-run averages across total wagering, not a promise for a session. The expected value of $1 on a 96% RTP slot is -$0.04, but any single result can be above or below that figure.
The last mistakes are pattern problems. Independent events do not owe you balance, and streaks do not repair themselves. When the question asks for several outcomes in a row, multiply. When it asks for at least one result, use the complement rule.
What this means for you
Before betting, compare true odds with payout odds, not just how large the prize looks. Use the house edge guide to rank games and expected value examples to price a stake in US dollars. Treat streaks as noise in independent games, and stop if the maths no longer fits your budget.
Frequently asked questions
What is the difference between odds and probability?
They describe the same event in two notations. Probability is the share of outcomes that win, such as wins divided by all outcomes. Odds against are losing outcomes to winning outcomes. Convert probability to odds with (1 - p):p, and convert odds L:W to probability with W ÷ (L + W).
What does 'true odds' mean?
True odds are the payout that would make a bet break even over repeated trials. Any shorter payout is the house edge. On a single-zero roulette straight-up bet, true odds are 36:1, but the payout is 35:1, creating a 2.70% house edge.
Do the odds change after a losing streak?
No, for independent games such as roulette spins or slot RNG rounds. Yes, in a limited way, for games with card removal, because the remaining cards have changed. Losing streaks alone do not predict a reversal. That mistake is the gambler's fallacy, while card counting belongs to dependent shoe games.
Which casino bet has the best odds?
The lowest figure here is blackjack at 0.48% house edge, but only with basic strategy and the listed rules: dealer stands on soft 17, double after split, late surrender, and 3:2 blackjack. Baccarat banker at 1.06% and French roulette even-money with la partage at 1.35% are also low-edge examples.