A blackjack betting strategy decides how much you put on each hand. Basic strategy decides how you play it. Only the second one moves the house edge. Martingale, Paroli and 1-3-2-6 change when you win and when you lose, never how much the casino keeps per dollar in action. What a staking plan can do is match your bankroll to the session you want and keep you inside the table limits. This page works through both, with the numbers.
On this page
- Why no blackjack betting strategy beats the edge
- Martingale: double after every loss
- Table limits: where the progression stops
- Paroli: press the wins, three at most
- 1-3-2-6: four hands, then reset
- A bankroll for flat betting: the calculation that matters
- A blackjack betting strategy checklist
- Where to put the plan to work
Why no blackjack betting strategy beats the edge
Every hand is a separate bet. Its expected value per dollar is the same whether you raised, lowered or held your stake after the last result. Add the hands up and the expected cost of a session is:
expected loss = total amount wagered × house edge
A system can only touch the first factor, and most progressions make it bigger. Wizard of Odds checked this on roulette with three bettors: one flat, one raising $1 after every win, one doubling after losses up to a $200 cap. They lost 5.267%, 5.266% and 5.269% of what they wagered, against an edge of 5.263%. Its pass-line Martingale test landed on the pass line's own edge of 7/495. A table with good rules keeps about 0.45% of every stake; the order of your bet sizes leaves that number where it is.
Martingale: double after every loss
The rule is simple. Bet one unit. After a loss, double. After a win, go back to one unit. Each finished run wins exactly one unit. The trouble is the run that does not finish.
Blackjack is not a coin flip here. On Wizard of Odds' table for a six-deck game played with basic strategy, a hand ends ahead 42.43% of the time, level 8.48% and behind 49.09%. A push just repeats the bet, so leave pushes out: 53.6% of decided hands are losses. That share is what drives a Martingale.
| Losses in a row | Chance, pushes left out | About 1 run in | Bankroll needed, $5 base | The next bet |
|---|---|---|---|---|
| 5 | 4.4% | 23 | $155 | $160 |
| 6 | 2.4% | 42 | $315 | $320 |
| 7 | 1.3% | 78 | $635 | $640 |
| 8 | 0.69% | 146 | $1,275 | $1,280 |
| 10 | 0.20% | 507 | $5,115 | $5,120 |
The bankroll formula: to survive n losses from a base bet b you need b × (2ⁿ − 1). Each extra loss you want to cover doubles the money at risk and cuts the chance of meeting that streak by a little under half.
Three things make it worse at a real table:
- Doubles and splits. By the hand counts in the same Wizard of Odds table, your first move is a double on about 9.5% of hands and a split on about 2.5%. When one of those arrives at step six, you either put up twice the progression bet or play the hand worse than basic strategy.
- Streaks come round. A 1-in-78 streak sounds rare. Over 100 finished runs, roughly 240 hands, the chance of meeting at least one seven-loss streak is about 72%.
- The payoff is lopsided. One failed seven-step run costs $635, the profit of 127 good runs at $5.
Wizard of Odds' verdict is short: the Martingale "won't even dent" the house edge.
Table limits: where the progression stops
Every table shows a minimum and a maximum, on the betting panel or in the info screen. Any blackjack betting strategy built on doubling runs into the maximum, because the ratio of the two fixes how many bets a run can contain:
bets in the run = whole part of log₂(max ÷ min) + 1
| Max ÷ min | Bets in the run | Bankroll in units | Chance of losing them all |
|---|---|---|---|
| 100× | 7 | 127 | 1.3% |
| 200× | 8 | 255 | 0.69% |
| 500× | 9 | 511 | 0.37% |
| 1,000× | 10 | 1,023 | 0.20% |
A $5 to $500 table therefore stops you after seven bets: the eighth would be $640. A wider spread makes the failure rarer and larger by the same factor; it never makes the run safe. Live tables and RNG tables set their limits separately, so read them for the table you actually sit at.
Paroli: press the wins, three at most
Paroli turns the Martingale round. Double after a win, return to one unit after a loss, and also return to one unit after three wins in a row. On even-money hands a cycle can only end two ways:
- a loss at any step costs 1 unit: −1 at step one, +1 − 2 at step two, +1 + 2 − 4 at step three;
- three wins bank 7 units: 1 + 2 + 4.
With 46.4% of decided hands won, three in a row comes up in about one cycle in ten. As a blackjack betting strategy its appeal is clear: your risk per cycle never exceeds one unit. Blackjacks at 3:2 and won doubles make some steps pay more than this, which is why the long-run cost has to come from the edge, not from counting cycles: you still pay about 0.45% of every unit in action. The open-ended version of the same idea, the anti-Martingale, reached a 32-unit target in about 2.9% of baccarat sessions in Wizard of Odds' simulation, and showed a loss for every target size.
Play live and RNG blackjack with crypto.
1-3-2-6: four hands, then reset
Bet 1, 3, 2 and 6 units on four hands in a row, moving to the next step only after a win. Any loss resets to the start. The net result depends on where the sequence ends. Wizard of Odds published the odds for the baccarat Player bet, which is close to a coin flip:
| Sequence ends | Net, units | Chance |
|---|---|---|
| Loss on hand 1 | −1 | 50.7% |
| Loss on hand 2 | −2 | 25.0% |
| Loss on hand 3 | +2 | 12.3% |
| Loss on hand 4 | 0 | 6.1% |
| Four wins | +12 | 5.9% |
The loss per unit wagered came out at 1.235% for the Player bet and 1.414% for the craps pass line: in each case the edge of the bet itself. A sequence puts about 4.07 units in action on average. Played at blackjack with a $10 unit, that is about $41 per sequence, and at 0.45% about 18 cents of expected cost. Flat betting $10 for four hands carries the same price.
A bankroll for flat betting: the calculation that matters
Flat betting is the plainest blackjack betting strategy, and with a flat stake the swing of a session follows from one number. Wizard of Odds gives a standard deviation of 1.142 units per hand for its six-deck table. Over N hands the swing is 1.142 × √N units, and the expected loss is N × 0.45%. A bankroll of expected loss plus two standard deviations covers the downswing in all but about two sessions in a hundred (a normal approximation).
| Hands in the session | Expected loss, units | One standard deviation | Bankroll, units | At $5 a hand | At $25 a hand |
|---|---|---|---|---|---|
| 100 | 0.45 | 11.4 | 23.3 | $117 | $583 |
| 400 | 1.8 | 22.8 | 47.5 | $237 | $1,187 |
| 1,000 | 4.5 | 36.1 | 76.7 | $384 | $1,918 |
The standard deviation is for Wizard of Odds' rule set; treat it as an approximation for others. Compare the columns with the Martingale table: a thousand flat hands need about 77 units, while a seven-step Martingale ties up 127 units for a single run and still fails at the rate shown above.
A blackjack betting strategy checklist
- Choose the session length first, then divide your bankroll by the units in the table above to set your stake.
- Read the minimum and maximum before the first hand.
- Keep the stake flat within a session. Change it between sessions, never in reply to the last result.
- Hold back money for doubles and splits: together they come up on roughly 12% of hands.
- Decline insurance and side bets; each carries its own, higher edge.
- Play every hand by the basic strategy chart. It is worth more than any staking plan.
- Check the payout first: a 6:5 table costs more than any betting system can hide.
The only bet sizing that does move the numbers reacts to the cards left in the shoe, not to past results. That is card counting, and it depends on how deep the shoe is dealt.
Where to put the plan to work
A good staking plan starts with a good table: 3:2 on blackjack, dealer standing on soft 17 and limits wide enough for your unit. The casinos below are ranked on our own score; open a table, read its info screen and set your unit before the first hand. The full comparison is on our list of crypto blackjack casinos.
- 1
Vave
4.9/5Casino and sportsbook on one balance, with around 90 coins in the cashier.
Latcas B.V. · since 2022 · Licence Curaçao OGL/2024/1676/0905 · 90 coins: BTC, ETH, USDT, LTC · Sportsbook on the same account
- 2
Shuffle
4.8/5Provably fair originals, a sportsbook and Solana deposits.
Natural Nine B.V. · since 2023 · Licence Curaçao OGL/2024/1337/0628 · 20 coins: BTC, ETH, LTC, SOL · Provably fair originals · Sportsbook on the same account
- 3
Wolf.bet
4.8/5Thirteen coins, two stablecoins and provably fair dice.
Cipher Games Ltd · Licence Anjouan ALSI-132405048-FI3 · 13 coins: USDT, USDC, BTC, ETH · Provably fair originals · Sportsbook on the same account
- 4
Stake
4.3/5The best-known name in crypto gambling, with in-house originals and a full sportsbook.
Medium Rare N.V. · Licence Curaçao OGL/2024/1451/0918 · Provably fair originals · Sportsbook on the same account
- 5
Cloudbet
4.3/5In crypto betting since 2013, with a sportsbook as deep as the casino.
Halcyon Super Holdings B.V. · since 2013 · Licence Curaçao OGL/2024/328/0599 · 30 coins: BTC, ETH, USDT, LTC · Sportsbook on the same account
The score counts what each casino publishes about licence, ID checks and limits, not how much we like the lobby.
Questions and answers
Does any blackjack betting system lower the house edge?
No. Each hand has the same expected value per dollar whatever you bet before it, so the long-run loss is the amount wagered times the edge. Simulations of Martingale, anti-Martingale and 1-3-2-6 all land on the edge of the underlying bet.
How much money do I need for a Martingale at blackjack?
Base bet × (2 to the power of n − 1) to survive n losses in a row. With a $5 base that is $635 for seven losses, and a run of seven losses still turns up about once in 78 runs.
Is Paroli safer than Martingale?
It risks far less per cycle, one unit, because it raises only after wins. It does not cost less per dollar wagered; the edge is the same.
What is the best way to size bets at blackjack?
Flat betting with a unit sized from your bankroll and planned session length, plus basic strategy. For 400 hands, about 48 units covers a normal downswing.
Do table limits matter if I flat bet?
Only the minimum. The maximum matters to doubling systems: it fixes how many losses in a row the progression can absorb.