Bankroll management for craps: sizing a session so variance doesn't end it

Most craps sessions don't end because the house edge ground you down. They end because a run of seven-outs hit a rail that was too small for the bets on it. Here is how to size a session so the normal swings of the game don't knock you out before the good hands arrive.

Think in units, not dollars

A unit is your base bet, the amount you put on the pass line or don't pass. At a $5 table your unit is $5. At a $10 table it's $10. Every other number in this article is a multiple of that unit, which lets the same rules work whether your rail holds $200 or $5,000.

A $10 pass bet with full odds behaves the same way whether you have $300 or $3,000 behind it. The only question is how many of those swings you can absorb before you're done. Divide your money by your unit and you have the honest answer.

Why craps swings harder than blackjack

At a $10 blackjack table, a hand risks $10, occasionally $20 on a double or split. At a $10 craps table playing the pass line with full 3-4-5x odds, a single point risks $10 flat plus $30, $40, or $50 in odds. That's up to $60 on one decision, and a 3-Point Molly can have three of those working at once, so $180 can vanish on a single 7.

The odds are still the best bet in the building. They cost nothing. But zero house edge is not zero variance, and every dollar you put behind the line multiplies how far your rail moves per hand. This is the trade you're making: lower expected loss, bigger swings. Bankroll management is how you afford the second half of that trade. The math of why swings happen is in the 36 combinations: a 7 shows one time in six, and four seven-outs in a row happens all the time.

A rule of thumb for session size

These are guidelines, not laws, and they assume you want a session of an hour or two that can survive a cold stretch.

The logic: a line bet with full odds puts up to 6 units at risk per point. Fifty units lets you lose eight or nine full points in a row and still be at the table, and a streak that bad is rare enough that you'll usually see hands go your way first. The Molly triples the exposure, so it needs roughly double the rail, since not every number is loaded at once. Placing across puts 6 units out on every roll after the comeout, and a couple of quick seven-outs eat it fast, which is why it needs the deepest rail of the flat-bet approaches. Compare the approaches themselves in strategies compared.

What a $1,000 rail actually supports

The free table starts you with a $1,000 practice rail, and it's a good number to think through because it's what a lot of people bring to a real casino.

At a $5 table, $1,000 is 200 units. That comfortably supports a 3-Point Molly with full odds, where each number carries up to $30 and a full board is $90. You could absorb ten fully loaded seven-outs in a row and still be playing. It also supports $32 across with room to spare, and place 6 and 8 for $6 each is almost impossible to bust with.

At a $10 table, $1,000 is 100 units. A single line bet with full odds ($60 per point) is fine. The full Molly gets tight: a loaded board is $180, so five or six bad hands in a row can take more than half your rail. Most players are better off either running two come bets instead of three, or taking odds only on the 6 and 8 where the point is most likely to hit. And $32 across at a $10 table isn't $32 anymore; place minimums scale with the table, so the honest spread is $64 across, and $1,000 is only 15 of those.

At a $25 table, $1,000 is 40 units, which is the bare minimum for pass with full odds and not enough for a Molly. If that's the only table open, play the line with odds on the 6 and 8 only and skip the come bets.

Loss limits and win goals that actually work

A loss limit is the amount you're willing to lose in a session before you walk. A win goal is the amount ahead at which you lock in a profit. Both only work if you decide them before the first roll and treat them as final.

The loss limit is easy: it's the session bankroll from the table above. Bring that amount, and when it's gone the session is over. Don't reach for the wallet, don't hit the ATM, don't tell yourself the table is due. There's no debt the dice owe you. Decide the session size in advance and leave the rest somewhere it can't be reached.

Win goals are trickier because a hot hand can blow past any target, and quitting in the middle of it feels wrong. A rule that holds up is the half-back rule: once you're up by half your session bankroll, set aside your original buy-in plus half the profit, and only play with the rest. If you bought in for $500 and you're up to $750, put $625 in your pocket and play with $125. If that $125 goes, you leave up $125. If the shooter keeps rolling, you keep pressing with money you've already decided is play money.

Press on wins, never chase

Pressing means increasing a bet with its winnings. When your $6 place 8 hits and pays $7, you tell the dealer “press it” and the bet becomes $12, with $1 back to you. It's the correct way to get bigger during a hot hand because the extra exposure is money the table just gave you, not money from your rail.

Chasing is the opposite: increasing bets after losses to win it back faster. It feels similar but it's the exact thing that empties rails. The dice don't know you're behind. Doubling a $10 pass bet to $20 after three losses doesn't change the 244-in-495 chance of the next one winning; it just puts twice as much on the same coin flip. A useful discipline: your base unit only goes up after a session where you finished ahead, and only by one step. It never goes up during a session in response to losses.

The same rule applies to the center of the table. The props pay 15 to 1 and 30 to 1, and when you're down it's tempting to let one big hit fix everything. At 11% to 16.67% house edge, the props fix nothing. If you like them, budget them: $10 of a $500 session, gone when it's gone.

Practice sizing before it costs anything

The one part of bankroll management that reading can't teach is what the swings feel like. A $180 loaded board disappearing on a seven-out reads like a number on this page. At a real table it's a physical sensation, and it's what makes people abandon their plan. So rehearse it.

Sit at the free table with the $1,000 practice rail, pick a unit and a strategy, and play a few hundred rolls with the loss limit and win goal written down beside you. Watch how often you hit the loss limit. Try $5 units and then $10 units and feel the difference in how fast the rail moves. Try the Molly with full odds and see how quickly $1,000 can be $600, then $1,300, then $700. When you've hit your loss limit twice in practice and walked away both times without reaching for more, you've learned the discipline that matters. It's the same discipline the strategy page comes back to: the bets you choose set how much you'll lose per dollar, but the sizing you choose sets whether you'll still be standing when the good hands come.