When you start a grid, you make one very specific call: price is going to bounce around inside the box I just drew. Price leaving that box means the market has rejected that call. The bot is not broken and it is not frozen. It simply has nothing to do out there.
That structure is the same on every platform that offers range-based automation, because it falls out of the design: buy low and sell high inside the range, do nothing outside it. What decides how much this costs you is not the breakout itself but what you do in the next few minutes. And breaking out the top versus falling through the floor leave you in opposite positions, which is why lumping them together — the way most "what to do when price leaves the range" advice does — is not much use.
1. First, separate "out of range" from "terminated"
People see no fills for a day and assume the bot died. Two different states:
- Out of range. The strategy is still in your running list, the funds are still locked with the bot, and price simply is not sitting on any of your levels, so nothing fills. Whenever price re-enters the range, it picks up where it left off.
- Terminated. The strategy is gone from the running list and the funds settle back under the platform's rules. What ends it is usually you hitting stop, a trigger price you set, or a passive cause — the pair delisted, a tick-size rule changed, your spot balance ran short.
The test is blunt: is it still listed as running, and has the money come back to your spot account? Still listed plus money not back = you are only out of range. An actual termination is a different conversation — whether your coins get market sold, how long the transfer back takes, whether a futures position closes along with the bot — and the settings that decide all of that live on the creation screen, not on the stop button.
2. Breaking out the top: you hold cash, and the problem is being left behind
Price climbs past your upper bound, which means the bot has already sold the base asset at every level on the way up. You are sitting on mostly quote currency — USDT, USDC, whatever you priced the pair in. No unrealized loss. The little realized profits in the fill list are even green.
The pain is comparative. Had you done nothing and just held, you would have captured that entire move. Running a grid means handing your position over one slice at a time as price rose, so the higher it went the less you held. That is not a settings mistake, it is what the tool is: you gave up trend upside to buy the round trips inside a range. The full version of that trade-off is in why grid bots lose money.
The saving grace is that this state is not urgent. You are in cash, nothing is compounding against you, and no liquidation price is chasing you, so you can take your time deciding whether and where to re-enter. What you actually need to guard against is the opposite impulse: watching it run and hurriedly opening a new, higher range to chase it. That is buying a full position back at the most expensive price on the chart.
3. Falling through the floor: you hold coins, and the loss is still growing
Price drops below your lower bound, which means every buy level filled, your quote currency is spent, and you are holding base asset bought all the way down at an average cost clearly above the current price.
This is the one to look at now, for three reasons:
- The unrealized loss keeps growing. Price keeps falling, your bag keeps shrinking. That is a different animal from the "earned less" of a breakout to the upside.
- The realized fills will fool you. Those grid profits in the list are real, but they usually do not come close to covering the unrealized loss. The number to watch is total return including unrealized P&L, not that stream of small green rows.
- A futures grid adds a liquidation price. Below the floor on spot, the worst case is a bag you still own. Below the floor on futures, price continuing against you can liquidate the position, and that is real margin gone. The difference in magnitude is laid out in spot grid vs futures grid.
4. Four responses, and what each one costs
There are really only four things you can do here. Each has a price tag. Read the price tag first.
| What you do | Only makes sense if | What it costs |
|---|---|---|
| Nothing — wait for price to come back | You still believe price chops inside this box and has just wandered off | Idle capital earning nothing; if the trend is not done, you get a bigger unrealized loss or a longer stretch on the sidelines |
| Stop it and settle the account | Your original read no longer holds and you want control back | Unrealized becomes realized; if termination triggers a market sell, you pay taker fees and slippage on the way out |
| Stop, then open a new range around the new price | You have a fresh read on where price now sits and are willing to re-bet | A full re-entry: old account closed, entry costs paid again, and the new read can be just as wrong |
| Use the platform's automatic range-following option | You want to stop doing it by hand, not to change your read | It solves who does the work, not whether the direction call is right; a range that followed price up will bag you just the same on the pullback |
On that last row: several platforms offer an automatic shift-the-range-up or shift-it-down option when you create a spot grid, and Binance's spot grid creation screen has settings along those lines. It moves the whole range when price breaks a boundary, saving you the manual stop-and-rebuild. It is not insurance. Once the range has followed price up, a pullback leaves you holding at the new, higher floor. Names, availability and trigger rules vary, so go by what your creation screen shows at the time.
If you do rebuild, the question becomes how to draw the new range: how width, grid count and profit per level constrain one another is in how to set grid parameters. Note that a wide range stretches the percentage per level very unevenly, so arithmetic versus geometric spacing decides whether the levels at one end of your new range even clear their own fees — that is arithmetic vs geometric grid spacing.
5. The most expensive reflex: dragging the range down
Nine times out of ten, the first instinct after breaking below the floor is "just move it down a bit and catch it." This deserves its own section, because it looks like a repair and is actually an add.
Moving the range down means committing more money at a lower price to buy the same asset and average down. That is not automatically wrong — it is fine if you genuinely want more of this asset at this price and the money was planned. The problem is that almost no one does it that way. It gets done under pressure from an unrealized loss, to make a number on screen look better.
Put plainly: the read you started with (it will chop inside a box) has already been rejected once. Shifting the range down re-bets that same read at a lower price. If the trend is not finished, you repeat exactly the same mistake, lower, with more size. The way grids turn into disasters has a shape, and it is: break below, shift down, break below again, shift down again.
6. Match your situation to the response
| Where you are | Ask yourself first | Leaning |
|---|---|---|
| Above the top, holding quote currency | Do I actually have a read on price at this level? | No rush. Redraw only when you have one; otherwise stay in cash rather than chasing a new range up here |
| Below the floor on spot, happy to hold the asset long term | Am I running a grid, or passively accumulating? | Stopping and keeping the coins in spot is reasonable, as long as you admit you have switched activities |
| Below the floor on spot, no interest in holding | Could I sit through another 20% down (illustrative)? | If not, take the loss on plan instead of being dragged along by "it'll come back" |
| Below the floor on futures | How far is liquidation from here? | Deal with liquidation risk first — reduce size or add margin — then decide about the strategy |
| In and out of range several times a month | Did the market change, or did I draw the range too tight? | Go back and review the settings rather than hand-shifting the range forever |
One more thing worth knowing: while you sit outside the range the bot earns nothing, but the annualized return figure on screen keeps being computed as realized grid profit divided by days running. Idle days still land in the denominator, so the headline drifts down — that is a bigger denominator, not a new loss. Where that number comes from and when it stops meaning anything is in where a grid bot's annualized return comes from.
And if you are going to stop with a market sell attached, price the exit first: how far the market has to move to cover a round trip of fees is what the break-even calculator answers, and the fee tiers behind it are in how Binance fees work.
7. FAQ
Does a grid bot stop itself when price leaves the range?
Usually not. Leaving the range and terminating are two different states. Out of range simply means there is nothing left to fill on that side, so the strategy sits there waiting for price to come back. What actually ends a bot is you hitting stop, a take-profit or stop-loss trigger price you set, or a passive cause such as the pair being delisted, a tick-size or minimum-quantity rule changing, or your spot balance running short. So a bot that still shows as running but has not filled anything in days is almost always just outside its range, not stuck or broken. If you cannot tell which state you are in, check whether it is still listed as running and whether the funds have already returned to your spot account.
Which is more urgent, breaking out the top or falling through the floor?
Falling through the floor. When price runs out the top you are holding mostly quote currency, you have no unrealized loss, and what you lost is the upside you did not capture, which is opportunity cost. Waiting another day does not make that hole deeper. When price falls through the floor you are holding the base asset you bought all the way down, and the unrealized loss keeps growing as price keeps falling. On a futures grid there is a liquidation price in the picture too, so price continuing against you can close the position for real. Same event, opposite urgency: up top you can think it over, down below you need to see how much room you have left.
Should I drag the range down to catch a price that already broke below it?
This is the most expensive reflex in grid trading, so be clear about what it actually is before you do it. Moving the range down means putting fresh money in at a lower price to average down, which is adding to a position, not repairing one. Your original read has already been rejected once, and shifting the range only re-bets the same read at a lower price. If the trend is not finished you will repeat the same mistake lower down with a bigger position. The question to answer first is whether you still believe price will chop inside some range. If yes, a new range is a real decision. If no, moving it down only postpones the loss to a more painful level.
If price comes back into the range, does the bot resume on its own?
As long as the strategy is still running and was never terminated, it fills again once price re-enters the range and reaches one of your levels. That is exactly what makes waiting a legitimate option. Two things to keep in mind though. First, the bot earns nothing while it waits, so the capital it holds is idle. Second, whether price comes back depends on your original read of the market, not on the bot. A strategy that has already terminated does not come back to life on its own; that one has to be created again, which means deciding the range and the settings from scratch.
If I stop the bot while out of range, will I sell at the worst possible moment?
You can, which is why the sell-all-base-on-termination setting matters before you stop anything. With it on, the system generally tries to market sell your base asset in one go, and when you are below the floor that is precisely when you hold the most of it at the lowest price. If you are willing to keep the coins, or the pair is thin, turning it off and letting the coins sit in your spot account so you can work out of them gradually is usually calmer than one cut. The exact setting names and where funds land after termination follow whatever your platform shows at the time you do it.
Nothing here is investment advice. Crypto prices are volatile, grid trading does not guarantee a profit, and it carries the risk of being left holding a losing position and — on futures — of liquidation. Judge your own situation and size accordingly. Every price and percentage above is an illustrative assumption for explaining mechanics, not a forecast. Range-following features differ in name, availability and trigger rules between platforms; go by what your own creation screen shows. Official product pages to check against: What is spot grid trading, Introduction to futures grid trading.