Craps.team Gamingencyclopedia

Craps strategy: systems ranked by what they cost

No strategy beats the casino — but the price gap between the best and the worst is close to a hundred to one. We break down eight systems: what each costs per $1,000 wagered, what bankroll it needs, and who it actually suits.

Analysis · no promises8 systems priced in dollarsAn honest Martingale verdictA method for testing it yourself

What a strategy can and cannot do

Start with the honest warning. No craps strategy lets you beat the casino. The house edge is built into the payouts themselves, and the order in which you push chips forward does not touch it — nor does the size of your bet, the length of a run, or a hunch.

The benefit is different, and it is real: the right system lowers the cost of an hour of play. The bankroll drains more slowly, the same budget lasts longer, and the money goes predictably instead of in lurches. The difference between disciplined play at 0.37% and a system running at 3.9% is a tenfold difference in the price of the same evening.

What each strategy costs per $1,000 wagered

The average the house keeps out of every thousand dollars that crosses the table.

$1.20 Don't Pass + 10× lay odds
$3.70 Pass Line + odds, and 3 Point Molly
$15.20 Place 6 and 8
$39 Iron Cross
$90–$111 Hardway 'systems'
$0 Saved by the Martingale

The simplest way to understand what a strategy does is through a single number: the cost of $1,000 wagered. It tells you how much the house keeps, on average, out of every thousand dollars that passes through your bets — and it sorts the systems into an honest order with no marketing attached.

The spread is close to a hundred to one: from $1.20 to $111 out of the same thousand. That, and not “your chances of winning”, is what actually separates one system from another.

Craps strategies ranked by real cost

SystemHouse edge (with odds)Cost per $1,000BankrollWho it is for
Don’t Pass + max lay odds0.12–0.18% (10×)≈$1.20–$1.8020–30 unitsSolitary mathematicians
Pass Line + max odds0.37% (3-4-5×)≈$3.7020–30 unitsThe baseline for everyone
3 Point Molly≈0.37%≈$3.7040–50 unitsThe gold standard
Dark Molly (DP + 2 DC + lay)≈0.14–0.37%≈$1.40–$3.7040–50 unitsExperienced dark-siders
Place 6 and 81.52%≈$15.2030 unitsSimplicity and pace
Iron Cross≈3.9% (per roll)≈$3950+ unitsShort fun sessions
Hardway “systems”9.09–11.11%$90–$111Nobody (entertainment)
Martingale1.41% (unchanged)$14.10 + ruin riskNo bankroll is enoughNobody
What a unit is

A unit is one betting increment — a chip of a single denomination you can wager independently of the others. A bankroll of “40 units” at a $10 bet means $400. Counting in units beats counting in dollars: the recommendations do not have to be recalculated for your stake level.

The one rule that genuinely works

Craps is math, and the player’s job is to find the least expensive positions mathematically. Everything hinges on one thing: odds bets, which the casino settles at a 0% house edge.

Odds do not exist on their own — they attach to a line bet. Which gives the working principle: a modest line bet plus maximum odds. The more money sits in the free part, the lower the combined house edge across both bets1.

Odds multiple behind the Pass LineCombined house edge
no odds1.41%
0.85%
0.61%
3-4-5×0.37%
10×0.18%

On the dark side the figure drops further still — to 0.12% with 10× lay odds behind Don’t Pass. That is the absolute floor available to a casino player.

Long run does not mean tonight

All of these percentages describe thousands of rolls. Within one session, extended winning and losing runs are entirely normal. A low house edge protects your bankroll from draining fast; it does not protect you from a bad evening.

System by system: what each one actually does

There are no right and wrong systems here — only ones whose logic you understand and ones you expect the impossible from.

3 Point Molly — the gold standard

Widely treated as the benchmark: it is simple, runs at about 0.37% with 5× odds, and needs a bankroll of 40–50 units.

  1. Bet the Pass Line, then add maximum odds once the point is set.
  2. Make a Come bet; when it establishes its own point, put maximum odds behind that too.
  3. Repeat step 2 once more. Three points working — hence the name.

The long-run math is no different from a plain Pass Line with maximum odds. The point of the system is elsewhere: in a long roll the Come bets pay repeatedly, and a seven-out wipes out all three at once. That is the price of more frequent payouts — the spread of outcomes widens in both directions.

The full breakdown with a roll-by-roll example is in the dedicated 3 Point Molly guide.

Don’t Pass + Lay Odds — the dark side

Mathematically the best system in craps. In a land-based pit, betting against the table earns sideways looks; online that argument disappears entirely.

  1. Put a small amount on Don’t Pass.
  2. Once the point is set, add the largest lay bet allowed — the dark side’s version of odds.
  3. At 10× lay you reach 0.12–0.18% house edge.

Most players never use it simply because they do not know it is marginally better than the popular mirror image, Pass Line + odds. The mechanics of Don’t Pass and Lay are covered in detail in the bet reference.

Iron Cross — a payout on nearly every roll

A system for players who care about how often they win rather than how much.

  • Once the point is set, take Place 5, Place 6 and Place 8.
  • Add a bet on the Field.
  • Typical proportions: Field and Place 5 at 5 units each, Place 6 and Place 8 at 6 each.

The coverage is nearly complete: the five, six and eight pay through Place, and every other number except the seven pays through the Field. You lose only on the seven — 6 outcomes in 36.

30 out of 36 — and still negative

Winning on 83% of rolls sounds like a cheat code, but the payouts are small and the occasional seven-out takes down all four bets at once. The per-roll house edge reaches 3.9% — ten times the price of a Pass Line with odds2. The Iron Cross works beautifully as a short diversion and very badly as a plan for the evening.

Place 6 and 8 — the fewest decisions

The simplest fast-moving system: keep two bets working on the numbers that appear most often after the seven.

  • The six and the eight are made 5 ways each out of 36 — 10 of 36 together.
  • The payout is 7:6 and the house edge 1.52%.
  • After a win the bet stays up and can hit again in the same round.

Recommended bankroll: 30 units. Bets must divide by six, or the fractional 7:6 payout rounds down and the house edge grows.

Bankroll management — not a system, but hygiene

Strictly speaking this is not a betting strategy but a set of rules that saves money under any bets. Three pillars, none of them optional.

  1. Choose low-edge bets. The choice is always available and it costs nothing.
  2. Fix the session bankroll in advance — in units — and do not go past it, whether you are running hot or chasing.
  3. Set a stop-loss. Do not play down to zero: some of the money has to survive into the next session.

You cannot calculate a house edge for this approach — it depends entirely on which bets you choose. But it is the only one that works every single time.

The Martingale — why it does not work

The Martingale is sometimes sold as a reliable way not to lose. It is not, and here is why.

  1. The player sets a base bet.
  2. After a loss the bet doubles — so the first win covers the whole run and leaves one unit of profit.
  3. After a win, back to the base.

The problem is geometry. From a $10 base, seven losses in a row make the next bet $1,280, against $1,270 already lost in the run. A win produces a pitiful $10 of profit; a loss makes the running total −$2,550.

That run is not rare

Seven consecutive Pass Line losses happen roughly once in 116 rounds, eight in 230. An average session covers that, and the run can land at the very start of it. Sooner or later you hit either the table maximum or the floor of your own bankroll — and the Martingale has not moved the house edge by a hundredth of a percent3.

Dice control — an honest assessment

For decades there has been a belief that a practised player can throw the dice to a planned result. Nobody has ever demonstrated that ability under controlled conditions.

At a regulation table both dice have to reach the far wall, whose pyramid rubber surface scatters them unpredictably. The one technique that could theoretically work — dice sliding, sliding rather than throwing — is against the rules in every land-based casino.

Online the question is closed completely:

  • in live craps the dice are thrown by a mechanical arm, not the dealer;
  • in Bubble Craps a mechanism shakes them under a clear dome, and it does not “know” which combination suits the casino;
  • in RNG craps there is no physical throw at all — the result comes from an algorithm certified by eCOGRA, GLI or iTech Labs.
The player's only lever

The dice are random — that is not an opinion, it is the condition under which the game exists at all. The only things you control are which bets you make and how big your bankroll is. Anything sold as “throwing technique” controls nothing but your wallet.

Which strategy suits you

Systems differ less in value than in character: some ask for patience through rare big wins, others hand out frequent small ones.

Player profileSystemWhy
BeginnerPass Line + max oddsOne decision per round, minimal rules
Loss minimiserDon’t Pass + max lay oddsThe lowest edge on the table
After a month of play3 Point MollyThree numbers working at the price of Pass + odds
Wants frequent payoutsIron Cross (short sessions)Something pays on 30 rolls out of 36
Wants fewest decisionsPlace 6 and 8Two bets and a payout you can read at a glance

Don't Pass + 10× lay

0.12%
Pays
1:1 + true odds
Type
Multi-roll

The absolute minimum house edge. Requires a table with a high odds limit.

Pass Line + 3-4-5× odds

0.37%
Pays
1:1 + true odds
Type
Multi-roll

The standard available at any table. Start here, and come back to it.

Place 6 and 8

1.52%
Pays
7:6
Type
Multi-roll

Simple and fast. The bet must divide by six.

Iron Cross

≈3.9%
Pays
mixed
Type
Four bets

Frequent payouts bought at a high price. Short sessions only.

A session plan: three numbers written down in advance

“Set a stop-loss” is useless advice until someone tells you where the number comes from. It comes from your bankroll for the evening — the amount you would not mind losing in full, and which is not committed to anything else. Everything else is derived from it.

The unit is the bankroll divided by the number of units the chosen system requires (the “Bankroll” column in the table above). The stop-loss is the portion of the bankroll after which you stand up. The target is the amount of profit after which you do exactly the same thing.

Bankroll for the eveningUnit for 3 Point MollyStop-lossWin target
$400$10−$200+$200
$1,000$25−$500+$500
$2,000$50−$1,000+$1,000

Half the bankroll in each direction is a working reference point, not a law: some players set the stop tighter, some count in time rather than money. What matters is not the exact fraction but that both numbers are written down before the first bet, while the decision is being made by a clear head rather than by the current score.

The win target matters no less than the stop-loss. Without it there is exactly one way for a session to end — with the bankroll gone: as long as money is left, a reason to stay will always present itself.

Three prohibitions that hold the plan together

Do not reload. Hit the stop-loss and the evening is over; a second deposit does not happen under any circumstances. Do not raise the unit after a loss. A bigger bet does not recover what is gone, it only speeds up the next result. Do not extend the session past the target. “A couple more rounds with the winnings” is a new session, started without a plan.

How to test a strategy properly

Do not take anyone’s word for it — run the system yourself at the free table. The procedure:

  1. Pick one strategy and write its rules down.
  2. Play a session of at least 200 rounds, sticking to those rules exactly.
  3. Record the bankroll after every round, not the wins — you need a curve, not a list of events.
  4. Compare the result with the strategy’s price from the table above.
What to expect from the result

In a short session the actual result will almost certainly diverge from the average, in either direction. That does not mean the strategy “works” or “doesn’t”. The longer the run, the closer the curve gets to the calculated price. That is the whole point of the exercise: watching randomness dissolve into math.

Picked a system and ready to play for money? See the craps casino rankings — and check the table’s odds limit before anything else: it decides how far down you can push the house edge.

Turning a strategy into discipline

A system only works when the decisions are made before the session, not during it:

  1. 1

    Work out your unit

    The unit is your base bet. Divide the session bankroll into units: 40 units at a $10 bet means $400, and not a dollar more.

  2. 2

    Pick one system

    Switching systems mid-session is a reliable way to get the worst of each. Write the rules down before the first bet.

  3. 3

    Set a stop-loss

    The point where you leave the table regardless of how it feels. Half the session bankroll is the usual figure.

  4. 4

    Set a stop-profit

    Exactly the same thing from the winning side: a 30–50% rise in the bankroll is a good reason to bank the result.

  5. 5

    Keep a record

    After the session, compare your actual result with the calculated cost of the system. It is the only way to see what is really happening.

Spinbetter
Editor's pick

Welcome bonus

25% to $250 + 400 free spins

Spinbetter

Is playing to a system worth it at all

What a strategy gives you

  • Cuts the cost of an hour several times over, with no loss of enjoyment
  • Removes impulsive decisions: the rules were agreed in advance
  • Makes the bankroll drain predictable and the session plannable
  • Forces you to learn what each bet actually costs

What it does not

  • Does not change the house edge on a single bet
  • Does not protect you from a bad session — the variance is unchanged
  • Progressions like the Martingale only redistribute risk, raising its peak

Common questions about craps strategy

What is the best craps strategy?
By house edge, Don't Pass with maximum lay odds (0.12–0.18% at 10×) and the Pass Line with maximum odds (0.37% at 3-4-5×, 0.18% at 10×). Both need discipline and a table with a high odds limit. A good compromise is 3 Point Molly: the same 0.37% price, but three points working and more frequent payouts.
Can a strategy make you win at craps?
Over distance, no. No system changes the probability of the dice or touches the house edge on a given bet. All strategies do is redistribute risk: some produce frequent small wins and rare large losses, others the reverse. Finishing an individual session ahead is entirely possible — that is luck, not the system.
Does the Martingale work in craps?
No. The Martingale creates an illusion of protection but runs into two walls: the table maximum and a finite bankroll. From a $10 base the eighth bet in a run is $1,280 for a $10 profit, and the total loss if it fails is $2,550. Seven consecutive Pass Line losses happen roughly once every 116 rounds — regularly, in other words.
What is 3 Point Molly?
A system that keeps three points working at once: the Pass Line and two Come bets, each backed with maximum odds. Mathematically it is identical to a plain Pass Line with maximum odds (0.37% at 3-4-5×), but the wins arrive more often and a seven-out hits harder, taking down three bets at once. It needs a 40–50 unit bankroll.
Is the Iron Cross a good strategy?
As entertainment, yes; as a plan for the evening, no. Place 5, 6 and 8 plus the Field cover 30 of the 36 combinations, so something pays on almost every roll. But the payouts are small, the occasional seven-out takes down all four bets, and the per-roll house edge reaches 3.9%. That is ten times the price of a Pass Line with odds.
Is dice control possible in online craps?
No, and for a different reason in each format. In live craps a mechanical arm throws the dice, not a person. In bubble machines they are shaken by a mechanism under a dome. In RNG craps there is no physical throw at all — the result comes from an algorithm certified by eCOGRA, GLI or iTech Labs. Checking that the certificate exists is worth doing before you deposit.
How much money do you need for a session?
Count in units, not dollars. Pass Line with odds needs 20–30 units, 3 Point Molly needs 40–50 (three bets at once), the Iron Cross needs 50 or more. At a $10 unit that is $200–$500. A bankroll below the recommendation means an ordinary losing run knocks you out before the system has a chance to show anything.
Can you combine strategies?
Technically yes, practically no. Switching systems mid-session almost always happens on emotion: after a loss you want to catch up faster, after a win you want to push harder. The result is the high house edge of the aggressive system without the frequent payouts you took it on for.

Related reading

Notes

  1. Combined house edge for the Pass Line with odds: 0.85% at 1×, 0.61% at 2×, 0.37% at 3-4-5× and 0.18% at 10×. For Don't Pass with lay odds it is lower still — down to 0.12% at 10×. The odds portion itself always costs 0%: the more money sits in it relative to the line bet, the lower the blended figure.
  2. The blended Iron Cross edge of ≈3.9% is per roll, for the standard $22 layout ($5 Field + $5 Place 5 + $6 Place 6 + $6 Place 8) with the Field paying 2:1 on the twelve. Published estimates run as low as ~1.1% because of a different base: per roll versus per resolved bet — after most rolls the bets do not lose, they stay in action.
  3. The Martingale changes the house edge on nothing: the Pass Line stays at 1.41%. Seven consecutive losses from a $10 base require an eighth bet of $1,280 against $1,270 already lost in the run — for a potential profit of $10. The table maximum and a finite bankroll make such a run a guaranteed loser over a long enough distance.

Choose your language