Rust crash: what auto cash-out can and cannot do for you

Auto cash-out is the most useful setting in a crash game and the one players misread most often. You type a multiplier, say 2.00x, and the game cashes you out the instant the round reaches it, faster than any click and without a last-second change of heart. That has real value. What the setting cannot do is make the round reach 2.00x more often, and no target you type in changes how much you lose on average.

This guide shows why, using the maths that sits under every provably fair crash game, and then covers what auto cash-out is genuinely good for. We also run the numbers on the martingale, because every crash chat has someone swearing by it.

How a crash round is decided

Each round opens with a short betting window. When it closes, a multiplier starts at 1.00x and climbs until it crashes. Cash out before the crash and you receive your bet times the multiplier at that moment. Still in when it crashes, and the bet is gone.

The crash point is decided before the round starts. On provably fair games it comes from a seed the site committed to in advance, often as one link in a long hash chain, and the climbing line is only the animation that reveals it. The site does not decide to crash early because a large bet is riding, and you cannot see the crash coming either. If you want the details of how a seed becomes a number, our provably fair guide covers it step by step.

Why crash multipliers follow a 1/x curve

Strip out the code and most crash formulas reduce to something close to this:

Crash point = (1 minus house edge) divided by (1 minus r), where r is a random number between 0 and 1.

The consequence is tidy. The chance that a round reaches any multiplier x works out to:

Chance of reaching x = (1 minus house edge) divided by x

Ignore the edge for a moment and the pattern is easy to remember. A perfectly fair game would reach 2x half the time, 10x one time in ten and 100x one time in a hundred. The survival chance is simply 1/x. The house edge then shaves the same small slice off every target.

Real games differ in the details. Some force an instant crash on a fixed share of rounds, some round the multiplier down to two decimals, and some mix the hash with a public seed. That is why we say roughly 1/x. The overall shape stays the same, because it is what lets the operator set one edge that applies equally to every possible target.

The numbers, with an example edge

Each site sets its own edge, and you should look for it on the fairness page. For the table below we use an example edge of 4% so the arithmetic is easy to follow.

Look at the last column. Whatever target you pick, the average return is the same 96% of your stake. We did not pick the targets to make that happen. Multiply the chance of reaching x by the payout x and the x cancels out, leaving 1 minus the edge every single time.

What the target actually changes

If every target has the same expected value, the target only changes the shape of your results.

• Low targets, from 1.10x to 1.50x, win most rounds and pay little. A session feels calm until one loss wipes out a string of wins. At 1.10x, a single lost 10-coin bet takes ten wins of 1 coin each to recover.
• Middle targets around 2x behave like a slightly unfair coin flip. Runs of four or five losses in a row are a normal part of any longer session.
• High targets of 10x and up mean long dry spells with rare big hits. At 100x in our example, you would wait more than 100 rounds on average for a single hit, and plenty of balances run dry long before then.

None of these is the clever choice. Pick the one whose swings you can live with, and accept that the average cost per round stays fixed at the edge times your bet. At 4% and 10 coins a round, that is 0.4 coins per round on average, so a hundred rounds costs about 40 coins in expectation whatever target you set.

One more detail catches people out: even a tiny target is never safe. In our example, a 1.01x cash-out only lands about 95% of the time, which means roughly one round in twenty crashes before you get out.

What auto cash-out is good for

None of this makes the setting pointless. It solves real problems, just not the odds.

• It removes your reaction time. On most crash games the auto cash-out is handled on the server, so the delay between your mouse and the site does not matter. A manual click aimed at 2.00x can land at 2.07x or, worse, after the crash.
• It takes the decision away from you at the worst moment. Watching the line climb past your target is exactly when greed says to hold a little longer. A number typed in before the round is a decision made with a clear head.
• It makes your session measurable. With a fixed target and a fixed bet, you know your hit rate and your expected cost per round. With manual play you are guessing.
• It pairs well with a round limit. If the site's auto-bet panel lets you set a number of rounds, a stop on loss or a stop on profit, use them. If it does not, count rounds yourself and stop at the number you chose.

What it cannot do is change the edge, protect you from instant crashes, or stop you from feeding winnings straight back into the next round. Those parts are on you.

The martingale, run through properly

The martingale is the oldest system in gambling. Bet at 2x. If you lose, double the bet. Keep doubling until you win, then go back to the base bet. Each win covers every earlier loss plus one base bet of profit, so on paper you can never lose a sequence.

Here is how it looks with a 1-coin base bet, a 2x target and our example 4% edge, which gives a 48% chance of winning each round and a 52% chance of losing it.

The arithmetic

With 1,023 coins you can survive nine losses in a row and still place a tenth bet. Ten straight losses happen in about 0.14% of sequences, roughly 1 in 692. That sounds remote. Now weigh it properly:

• About 99.86% of sequences win you 1 coin.
• About 0.14% of sequences lose you 1,023 coins.
• The expected result per sequence is 0.9986 minus 1.479, which is roughly minus 0.48 coins.

The average sequence puts about 12 coins into play, and 4% of 12 is 0.48. The system lost exactly the house edge on everything it wagered. It simply rearranged the losses into a long run of tiny wins followed by one large hit.

That large hit is also less rare than it feels. Over 100 sequences, the chance of at least one ten-loss streak is about 13.5%. Over 500 sequences, which would earn you 500 coins if everything went to plan, it is about 51%. Chasing a profit half the size of your bankroll this way gives you roughly even odds of losing the lot before you get there.

The wall: bet limits and bankrolls

Even players with deep pockets hit a ceiling, because every crash game has a maximum bet and each site sets its own. Say, as an example, the maximum were 250 coins on a 1-coin base. You could place bets of 1, 2, 4 and so on up to 128, and the next doubling to 256 would be refused. The table limit works exactly like a smaller bankroll. Raise the base bet to make the wins feel worthwhile and you reach the wall even sooner.

Other progressions, such as Fibonacci or d'Alembert, fail for the same reason. When every single bet has a negative expected value, no pattern of bet sizes can add up to a positive one. Changing the order and size of your bets only changes when you feel the losses.

Reading the history bar

Crash lobbies show a strip of recent results, and chat fills up with theories about it. Three low rounds means a big one is due. A 50x just hit, so the next few will be short. None of it holds up. If the game uses a hash chain, every crash point in the chain was fixed before its first round was played, long before anyone looked at the strip. Each result is independent of the ones before it.

The history bar is useful for one thing: checking. Pick a round, copy its hash from the fairness panel and confirm that it fits the chain and the published formula. That is a far better use of it than pattern spotting.

A setup we would actually use

If you are going to play crash, a boring setup keeps the cost at the level the edge says it should be and stops a bad run from turning into a chase.

1. Decide a session budget in real money before you open the site, and convert the site currency (coins, gems or credits) back into dollars so the numbers mean something.
2. Pick a fixed bet that is a small slice of that budget. As an example, 1% of the session budget per round leaves room for ordinary losing streaks.
3. Set auto cash-out to a target you are comfortable with and work out its hit rate with the formula above.
4. Set a round limit and, if the auto-bet panel offers them, a stop on loss and a stop on profit.
5. When you hit a stop, withdraw or log off. Do not raise the bet to win back what you lost.

If a site publishes a chance of reaching a given multiplier, you can check its edge yourself: the edge equals 1 minus the target times that chance. If it publishes neither the edge nor the formula, ask support before you deposit.

Where you can play Rust crash

In our current data, crash is listed at RustClash, RustBattle, RustReaper and RustChance. The differences that matter for this guide are the ones you can check yourself: whether the fairness page publishes the formula and the hash chain, what the minimum and maximum bets are, and what the auto-bet panel lets you configure. Our Rust crash sites ranking compares them, and the RustClash review and RustBattle review cover payouts, fairness and the rest of each site. RustChance sits last in our overall ranking, so read its review carefully before you deposit there.

Crash rounds are short, so the edge gets charged many times in a single sitting. Treat the expected cost per round as the price of the entertainment, set that price before you start, and stop when it is spent. If stopping is getting harder than it used to be, our responsible gambling page lists blocking tools and free, confidential helplines.