A stop-loss or win-limit can be a useful rule for controlling the size and length of a craps session. It cannot make the dice remember your target, improve a wager’s payout, or turn a negative-expectation bet into a positive-expectation one. The distinction matters because a good money boundary can become a bad gambling theory when a player starts treating the boundary itself as an edge.
A stop-loss means leaving after losses reach a preselected amount. A win-limit means leaving after profit reaches a target. A time limit and a maximum bet rule are similar session controls. They answer questions such as How much am I willing to lose? How long am I willing to play? How large may one decision become? They do not answer the mathematical question What is the probability of the next roll?
A stopping rule changes when you leave, not the dice
Two fair dice have 36 equally likely ordered combinations. Your decision to quit after losing $150 does not remove any combinations from the next roll. A seven is still produced by six of the 36 combinations; a six is still produced by five. Likewise, the posted rules of the wager determine its long-run expectation. The player cannot edit that expectation by writing a personal exit number on a piece of paper.
This is easiest to see with a single repeated wager. Suppose a player makes the same $10 bet whenever a decision is available. Whether the player plans to leave at +$100, at -$100, after 90 minutes, or when dinner is ready, each new $10 decision is still settled under the same table rules.
The stopping rule can change the shape of session results. A small win target may create many sessions that end with a modest profit and fewer sessions that continue long enough to produce a large loss. A tight stop-loss can cap the amount lost in one sitting. Those are real behavioral effects. They are not evidence that the underlying wager has become favorable.
Why “I win most sessions” can still lose money overall
One of the most persuasive versions of this myth is built around win frequency rather than expected value. A player chooses a small win target and a larger loss limit, then notices that the small target is reached fairly often.
Imagine a simplified session rule:
- leave at +$50 if the win target is reached first;
- leave at -$200 if the loss limit is reached first.
If the player wins four sessions at +$50 and then has one -$200 session, the five-session total is exactly zero before the house edge and any extra action are considered. A fifth small win is required just to move ahead. The player can therefore report a high percentage of “winning sessions” while the occasional larger loss erases many of them.
The reverse can also happen: a player may accept many small losses while waiting for a large win. Session win rate, average winning-session size, average losing-session size, and total action all matter. Counting only how often the cashier cage sees you leave ahead is not enough.
Expected loss follows action and wager quality
For a simple wager with a known house edge, the basic relationship is:
[ \text{Expected loss} = \text{total amount wagered} \times \text{house edge} ]
Suppose a player’s mixture of bets works out to an average house edge of about 1.5% for illustration. If the player generates $600 of resolved action before a stop rule ends the session, long-run expected loss on that action is:
[ $600 \times 0.015 = $9 ]
If the same player ignores the planned stopping point and generates $2,000 of comparable action, the expectation becomes:
[ $2{,}000 \times 0.015 = $30 ]
The stop rule helped in the second comparison only because it prevented additional wagering. It did not reduce 1.5% to a smaller percentage. This is the useful way to understand session limits: less action at the same negative edge generally means less expected cost.
Use the expected loss calculator when you want to compare the effect of action, edge, and session length directly. The variance simulator helps show why an individual session can finish far above or below that expectation.
A target cannot “lock in” a long-run profit
A win-limit can lock in the result of a session only in the ordinary sense that you stop wagering and leave with the chips. If you are up $200, cash out, and do not place another bet, those chips are no longer exposed to the table that night.
The myth appears when that sensible fact is extended into a claim such as “I only need to win one unit every visit, so the casino cannot beat me.” The future visits are new exposure. The player who returns tomorrow has not carried a mathematical protection forward from yesterday’s cash-out.
This is why a sequence of daily +$25 targets is not equivalent to earning a wage. The path to each target is uncertain, and there is always some chance that the loss boundary is reached first. Repeating the plan repeats the negative-expectation decisions as well as the opportunities to reach the target.
A more accurate statement is: a win-limit decides when you stop risking a current profit; it does not guarantee how often that profit will appear.
Moving the limit after the dice turn against you defeats its purpose
A precommitted limit and a moving limit are different things. Consider a player who starts with these rules:
| Rule | Planned boundary |
|---|---|
| Buy-in | $300 |
| Stop-loss | Leave with $150 remaining |
| Win-limit | Leave with $450 in the rack |
| Largest ordinary bet | $30 |
| Proposition-bet budget | $0 |
The structure is simple. The player has defined both the maximum planned loss and the point at which a $150 profit will be taken off the table.
Now imagine that at $155 remaining the player says, “I will move the stop to $100 because the shooter looks good.” At $105, the stop becomes $50. Eventually the original limit exists only as a story about what was supposed to happen.
The same drift occurs on the winning side. A player reaches $450, raises the target to $550, then to $650 because the table feels hot. There is nothing mathematically wrong with choosing to continue if the player can afford the risk. What is wrong is calling the original win-limit a protection after it has been repeatedly abandoned.
The point of a session rule is not to predict the dice. It is to make a decision before noise, celebration, frustration, or a recent streak changes your judgment.
Loss limits and bet limits solve different problems
A $200 stop-loss says how much of the session bankroll may disappear. It does not prevent a player from placing $150 on one decision and reaching the stop in a few seconds. A separate bet-size limit controls concentration risk.
That distinction matters at craps because the layout offers many simultaneous ways to increase exposure. A player can have a Pass Line bet, odds, place bets, come bets, and proposition bets working around the same sequence of rolls. The amount on the felt can grow while the player still thinks in terms of one “$15 table.”
A more complete session plan therefore considers:
- loss boundary: maximum planned session loss;
- profit boundary: a point at which the current win will be cashed out;
- maximum total exposure: how much can be working at once;
- bet-quality rule: which wagers will and will not be used;
- time boundary: when play ends regardless of the rack;
- reload rule: whether another ATM or cage visit is allowed.
A no-proposition-bet rule, for example, may improve the quality of the action by avoiding some high-edge bets. That is different from the stop-loss itself. The craps house-edge material and expected-value guide explain the bet math; the limit decides how much of that math you are willing to buy.
The casino evaluates volume, not your slogan
From the floor’s perspective, “I always quit when I double” is not a special category of player. A casino cares about actual betting volume, average wager, game speed, time, rules, and the resulting theoretical value. A player who reliably leaves early may generate less total action. A player who announces a stop but repeatedly extends the session may generate much more.
That is why a disciplined exit can reduce the casino’s expected revenue from that particular visit without “beating” the game. Fewer priced decisions were purchased.
The same distinction appears in Why Low House Edge Still Loses Money. Choosing lower-edge bets can reduce expected cost per dollar wagered. Choosing a shorter session can reduce the number of dollars wagered. Both can lower expected cost, but neither promises a winning result.
Common stop-rule arguments, tested carefully
“I leave after three losses, so a fourth loss cannot hurt me.” Correct for that session: you are no longer playing. Incorrect as a system claim: the first three wagers still had their original probabilities and expectation.
“I take profit off the table after every win.” That can limit how much current profit is re-exposed. It does not change the expectation of the next chips you choose to wager.
“I use a trailing stop, so I cannot give back a big win.” A trailing stop can preserve some profit if the table turns before the threshold is crossed. It also may end a session that would later have recovered. It is a risk-management preference, not a probability forecast.
“I only play until I win one unit.” A small target may be reached frequently, but the losing paths and their sizes still count. The target cannot be evaluated by win frequency alone.
“The system works because I have used it successfully for ten trips.” Ten trips are a tiny sample compared with the variance of craps. A favorable run shows that favorable runs occur; it does not show that the wager probabilities changed.
A better way to use a stop-loss or win-limit
Use limits for the job they can actually perform. Decide them before the buy-in. Make the loss number small enough that reaching it does not require a rescue plan. Keep the bet size consistent with the bankroll so one ordinary sequence cannot destroy the whole boundary immediately. If the limit is reached, treat the decision as already made rather than reopening negotiations with yourself at the rail.
A win-limit can be equally simple: once the selected rack value is reached, color up and cash out. You do not need to tell yourself the table is “due to turn.” You are leaving because that was your rule.
For bankroll survival, continue with Craps Bankroll Risk. For the danger of increasing action after losses, see Craps Loss Chasing. For lower-cost ways to structure play, compare Reduce the Cost of Playing Craps and Low-Bankroll Craps.
A stop-loss can stop a session. A win-limit can stop a session. Neither can stop probability from being probability. Their real value is much more modest and much more useful: they put a boundary around how much action you are willing to buy.