Regression betting in craps means starting with more money on the layout, then reducing the bets after one or more hits. It can lower exposure after an early win, but it does not change the house edge of the bets. Regression manages volatility and psychology. It does not turn craps into a beatable game.
Quick Facts
- Regression means reducing a bet after a win or after a planned trigger.
- It is common with place 6 and 8, inside numbers, and across setups.
- It can protect some early profit after a hit.
- It can also create larger first-roll exposure.
- Regression does not improve the mathematical edge of the underlying bets.
- The main risk is seven-out before the first planned hit.
- It is a bankroll-control method, not a winning system.
Plain Talk
Regression betting is the opposite of pressing. Instead of increasing after a win, you pull money down.
A player may start with $66 inside, catch one hit, then drop to $32 inside. Another may start with $30 each on the 6 and 8, take the first hit, then reduce both to $12. The goal is simple: get some money off the table before the seven arrives.
That sounds disciplined. Sometimes it is. But the first part of the system is aggressive. You start with extra money exposed, and craps does not promise you the first hit before the seven.
This page is about regression as a betting behavior. For the cost of the bets themselves, read craps house edge and place bet house edge. For raw probability, use the craps odds calculator.
For outside reference, the Wizard of Odds craps basics explains standard craps bets and house-edge framing, the Wizard of Odds craps edge derivations show how expected value is calculated, and the Massachusetts craps and mini-craps rules show how live-table wagers, dice handling, and dealer procedure are formally controlled.
How It Works
A regression plan has three parts.
| Step | What the player does | What it really changes |
|---|---|---|
| Start larger | Put more money on numbers early | Higher first-roll exposure |
| Catch a hit | Collect or partially collect a payout | Some bankroll recovery |
| Regress | Reduce bets to a smaller setup | Lower future exposure |
A common example:
| Setup | Amount exposed | Trigger | After regression |
|---|---|---|---|
| $66 inside | $66 | One inside hit | $32 inside |
| $30 each on 6 and 8 | $60 | One hit | $12 each on 6 and 8 |
| $110 inside | $110 | Two hits | $44 inside |
The house edge on each individual bet stays the same. Regression only changes how much money is in action at different moments.
The dangerous part is the first roll after the setup. If the seven appears before the regression trigger, the larger opening exposure is gone. That is why players who talk about regression usually remember the sessions where the first hit came quickly and forget the sessions where the seven came immediately.
Craps Table Example
You buy in for $400 at a $15 table. The point is 9. You tell the dealer:
“Sixty-six inside.”
The dealer places:
| Number | Bet |
|---|---|
| 5 | $15 |
| 6 | $18 |
| 8 | $18 |
| 9 | $15 |
Total exposure: $66.
The next roll is 6. Your $18 place 6 pays $21. You tell the dealer:
“Take me down to thirty-two inside.”
Now the dealer leaves smaller bets working and returns chips to you. You had one hit, reduced exposure, and kept some action. That is the clean version.
The ugly version is easier: same $66 inside, first roll is 7, all $66 loses before the system gets to breathe.
From the Casino Side:
Dealers do not care whether a regression plan is clever. They care whether the instructions are clear, timely, and paid correctly.
A regression player can slow the game if he gives vague commands after every roll. “Take me down some,” “make it look good,” or “same thing but smaller” forces the dealer to clarify. On a busy craps table, vague regression talk creates payout delays and possible disputes.
The boxman and floor supervisor care about clean chip movement. Large regressions mean chips come off numbers, move through the dealer’s hands, and return to the rail. Surveillance wants the sequence visible: original bet, payout, instruction, reduction, returned chips.
Good players say the amount clearly. Bad players negotiate with the layout after the dice are already moving.
Common Mistakes
- Thinking regression lowers the house edge.
- Starting too large for the bankroll.
- Regressing after one hit but immediately pressing again without a plan.
- Giving unclear dealer instructions.
- Ignoring the chance of an immediate seven-out.
- Using regression to justify too many numbers at once.
- Counting returned chips as profit before subtracting the original risk.
Hard Truth
Regression feels safe because it takes money off the table after something good happens. The casino edge was already there before the good thing happened.
FAQ
Is regression betting a good craps strategy?
It can be a reasonable bankroll-control method if the starting exposure fits your bankroll. It is not a mathematical advantage.
Does regression reduce the house edge?
No. It changes bet size and timing. The house edge of the underlying bets stays the same.
What is the biggest risk with regression betting?
The seven can roll before your first hit. Then the larger starting setup loses immediately.
Is regression better than pressing?
Regression is usually less volatile after the first hit. Pressing is more aggressive. Neither removes negative expectation.
Which bets are common for regression?
Place 6 and 8, inside numbers, and across setups are common because they can be adjusted by the dealer after hits.
Should beginners use regression betting?
Beginners should first understand craps bets explained and table commands. Regression adds moving parts.
Can regression protect a winning session?
It can help reduce future exposure, but it cannot protect money already lost or prevent a later seven-out.
Deeper Insight
Regression is popular because it solves an emotional problem: players hate watching a decent hit disappear on the next seven. Pulling down gives a feeling of control.
That control is real in one narrow way. You decide how much money remains exposed. But the dice outcome is still random. The system does not know when the seven is coming.
A sharper way to judge regression is by total action. If you start with $66 inside for ten shooters instead of $32 inside, your session has more money exposed early. Even if you regress on some shooters, the first-roll losses matter.
A regression player should track three numbers:
| Number to track | Why it matters |
|---|---|
| Initial exposure | Shows how much can vanish before the plan works |
| Amount pulled down | Shows actual risk reduction |
| Total action per hour | Drives expected loss |
The expected loss calculator is more useful than a win goal here. It forces the player to look at total money wagered, not just the elegance of the system.
Formula / Calculation
Expected Loss = Total Amount Wagered × House Edge
Example:
$2,000 total action × 1.52% average edge = $30.40 expected loss
If regression causes you to wager $3,200 in total action instead:
$3,200 × 1.52% = $48.64 expected loss
Formula Explanation in Plain English
The system name does not matter as much as the money pushed through the layout. If regression starts with bigger bets and creates more total action, the expected loss can rise even if the later bets look safer.
Related Reading
Start with the craps guide if you want the full course path. Use craps odds to understand why the seven matters, then compare the cost on craps house edge. For the most common regression base, read Place 6 and Place 8 Strategy. If you are tempted to turn regression into a guaranteed pattern, read why betting systems fail.