- The order to set them: range first, then grids, then investment
- Upper and lower bounds: set them off historical swings
- Grid count / density: too dense earns pennies, too sparse fills slow
- Per-grid investment: does it cover two fees
- Geometric grids vs arithmetic grids
- Should you trust the auto-suggested parameters
- The last check before you hit create
- FAQ
A grid bot has no judgment of its own. It just faithfully runs the few numbers you gave it. So whether a grid makes money comes down, maybe eighty percent of the time, not to the market but to what you typed into those boxes on the setup page. Same choppy market: the person who drew the range right and sized the grids well pockets a steady spread, while the person who drew it crooked and packed the grids too tight bleeds out a little more with every fill. This guide walks each box, one at a time: how to set the upper and lower bounds off historical swings, how to weigh grid density, whether each grid covers its fees, what actually separates geometric from arithmetic, and whether that convenient auto-suggested setup is safe to copy.
If you're still fuzzy on what a grid actually earns from, or how spot risk differs from futures, go back through the full grid trading guide first. This one assumes you already know a grid feeds on chop and earns the back-and-forth spread, and jumps straight to parameters.
1. The order to set them: range first, then grids, then investment
A lot of people open the setup page and fill the boxes top to bottom, then realize halfway down that an earlier number has to be scrapped and redone. These parameters actually have an order to them, and following it saves a fair bit of backtracking.
Step one is always the range—it matters most and leans hardest on your read of the market, so think through where price is likely to swing next before anything else. Step two is the grid count; with the range fixed, the count sets how far apart each grid sits, and here you keep one eye on "is the spacing enough to cover fees." Only then, step three, the investment amount: capital divided by grid count is your per-grid investment, and if that comes out too small, you go back to step two and trim a few grids. Last come the optional bits—trigger price, take-profit, stop-loss. The logic behind this order: range and grid count together decide each grid's spread, comparing that spread against fees tells you whether the grid is even worth running, and the amount just gets spread across the grids. Flip the order and agonize over the amount first, and you usually end up redoing it.
2. Upper and lower bounds: set them off historical swings
The range is the band you think price will bounce inside for a while. The upper bound is a high you figure it probably won't reach short-term; the lower bound is a low that still seems reasonable, a level where you'd be happy to keep buying. This is the most important parameter of the set, and the easiest to get wrong by just guessing.
How do you set it without guessing? The most practical approach is to read the historical swings: pull up the last few months of this coin's chart and see the band it keeps grinding inside—roughly where the highs top out, where the lows bottom, how wide the box is. Put the upper bound near that period's highs, the lower bound near its lows, then leave a little cushion outside each. Don't pin a bound exactly on a recent high or low; price briefly poking through a prior high or low is common, and pinning it too tight means one wick can sweep you right out of the range.
Range width is a trade-off, and both extremes cost you:
- Too narrow: price leaves the range easily. Break out the top and your coins get sold off early, leaving you sitting in cash while it runs. Break below the bottom and you're stuck holding the coins you bought up high inside the range. The sneaky part of a narrow range is that in a backtest it "looks" like it fills constantly and the return numbers are lovely—but the moment real price leaves that narrow band, the lovely numbers fall apart. This is the trap most beginners walk into.
- Too wide: the grids get spread thin. Either each grid has a big spread but fills rarely and earns slowly, or you pile on grids to avoid that (and you're back to the fee problem). The more concrete cost is poor capital efficiency—your money is laid out across a huge range, price only swings in a small middle stretch, and the cash out at both ends never moves, so it's just sitting there dead.
So a wider range isn't safer and a narrower one isn't more profitable. What you want is a range that hugs the swing you genuinely expect. If you have no real feel for where this coin will swing next, that's a sign it isn't a good grid candidate right now—forcing one is just a bet.
3. Grid count / density: too dense earns pennies, too sparse fills slow
With the range fixed, the grid count decides how many slices you cut it into—that is, how big the spread between each grid is. Count and density are the same thing: same range, more grids means denser and a smaller spread per grid; fewer grids means sparser and a bigger spread per grid. This is the second spot beginners routinely blow up, because intuition lies to you—it feels like slicing finer and filling more must earn more. The opposite is true.
Lay the trade-off out:
- Dense (many grids): small spread per grid, so the smallest move triggers a grid, fills come constantly, and profit lines pop up one after another—it feels great. But each fill's gross spread is thin, and the two fees (buy and sell) are fixed no matter what. Thin enough and fees eat most of the spread or all of it; at the extreme each grid is actually negative—the more happily the bot trades, the faster you lose.
- Sparse (few grids): big spread per grid, so the net after fees is solid, but price has to travel a real distance to trigger a grid, fills are infrequent, and it earns slowly. In a tight chop it might not move a single grid for a long stretch.
There's one bottom line for grid count: keep each grid's spread clearly larger than the two round-trip fees. Not barely larger—clearly larger, with cushion to spare, because not every fill lands at the maker price. Sometimes you take (taker), sometimes there's a bit of slippage, and a thin cushion gets eaten. To turn that into actual numbers, two calculators working together is the most direct route: first the grid spacing / count calculator shows the spacing percentage when a given range is split into N grids, then plug the range, grid count, capital, and your account's real fee rate into the grid profit simulator, and it tells you straight whether the per-grid net is positive or negative. If the simulator shows a negative net or one hugging zero, drop the grid count until the net is a decent positive.
4. Per-grid investment: does it cover two fees
Per-grid investment = total capital ÷ grid count. It decides the absolute dollars each fill earns. The per-grid net from the last section is a percentage; multiply it by the per-grid investment and that's the money that actually lands in your pocket. So for the same range and grid count, a bigger account means a thicker per-grid investment and more real dollars per fill; a small account means a thin per-grid investment, and the few cents each grid earns may not be worth the bother.
A per-grid investment that's too small causes two concrete problems. First, it's not worth it: even a positive per-grid net percentage, multiplied by a tiny investment, can leave you a few cents—maybe nothing after two fees, maybe negative—while still tying up capital and carrying the risk of getting stuck holding. Poor return for the trouble. Second, it won't create: Binance sets a minimum order size per fill, and if your per-grid investment falls below that floor the grid simply won't launch. That floor varies by trading pair and over time, so go by whatever the page shows when you set it up.
What do you do? With capital fixed, per-grid investment and grid count move in opposite directions—to make each grid thicker you have to trim the count. So when you don't have much to work with, the right move isn't to run a very dense grid on small money for the busy-work thrill; it's to use fewer grids so each one's investment and payoff are meaningful. Matching capital to grid count loops right back to last section's fee trade-off: fewer grids means a thicker per-grid investment and a bigger per-grid spread, so fees take a smaller cut—all three go together.
5. Geometric grids vs arithmetic grids
When you create a grid, Binance asks you to pick "geometric" or "arithmetic." These are two ways of spacing the grids across the range, and plenty of beginners click past it without a second look—but it changes what each grid earns more than you'd think.
| Type | How grids are spaced | What that means |
|---|---|---|
| Geometric grid | Adjacent grids are spaced by the same percentage—e.g. each grid is about 2% above the last | Every grid earns the same percentage, whether price is at the high or low end of the range. Fits assets that move in percentage terms, which is most of the time. |
| Arithmetic grid | Adjacent grids are spaced by the same dollar amount—e.g. each grid is 50 USDT above the last | Lower down, that fixed gap is a bigger percentage and earns more per grid; higher up it's a smaller percentage. The spacing is uneven. |
Here's a concrete example. Range 2,000 to 3,000, split into 20 grids. On a geometric grid each step is a touch over 2%: near 2,000 that 2% is about 40, near 2,900 that 2% is about 58—different dollars, same percentage. On an arithmetic grid every step is a fixed 50: near 2,000 a 50 move is 2.5%, near 2,900 a 50 move is only 1.7%—different percentages. For crypto, which moves in percentage terms, geometric usually matches your intuition better: the per-grid return is steady and easy to estimate, which is why most people leave it on geometric. Arithmetic suits the case where you expect to earn more at the low end and price moves in relatively fixed dollar swings—used less often. When in doubt, use geometric; its per-grid return is obvious at a glance. To see the actual numbers, the grid profit simulator lets you toggle geometric vs arithmetic and compare the per-grid spread for the same range.
6. Should you trust the auto-suggested parameters
When you create a grid, Binance usually gives you two routes: "manual setup" or "smart / suggested parameters." One tap on suggested and it fills in the range and grid count for you—very tempting for beginners, since it saves you from pulling those numbers out of the air. Can you just use them? The answer: fine as a reference, don't copy blindly.
The problem is where the suggestion comes from. It's usually computed from the coin's recent historical volatility, which quietly assumes "it'll keep ranging the way it just did." That assumption holds in a sideways market, but markets change: three months grinding inside a box doesn't mean next month won't break out—up or down—on a single headline. The suggestion can't see the future; it just replays the past volatility. Once price walks out of the history it was built on, a grid set to the suggested range misses the run or gets stuck holding all the same.
Another catch: suggested parameters tend to hand you a range and count that lean narrow and dense, because that produces high fill frequency and pretty return numbers in a backtest—which is exactly what nudges you to hit "create." But as covered above, narrow-and-dense is the combination with the highest fee drag and the highest risk of getting stuck.
So the right way to use them: treat the suggested values as a starting point and a reference—see roughly where the system thinks the range sits and how many grids it proposes—then run those two numbers past your own read of the coin. Especially the upper and lower bounds: you're better off setting those yourself rather than letting an algorithm bet on direction for you. Check the grid count and per-grid investment with the calculators, following the logic from the earlier sections. The suggestion saves you the "start from a blank page" work, but it can't carry the market call for you.
7. The last check before you hit create
Once every box is filled and you're about to hit create, Binance usually makes you confirm again—your last chance to catch a mistake. Spend fifteen seconds running down this list and you'll block most of the dumb ones:
- Are the upper and lower bounds the right order of magnitude—don't flip them, and don't add or drop a zero. This is the most common and most fatal slip of the finger.
- Does the per-grid spacing percentage cover two fees—the page usually shows this percentage once you enter the grid count. Stop and look; if it's too small, go back and cut grids.
- Is the per-grid investment below the platform minimum—too low and it won't create, or each grid is too thin to bother with.
- Have you thought through the trigger price and take-profit / stop-loss—these are optional, but at least settle on a stop-loss line. It's your safety rope for when you're not watching.
- Does this coin, in this market, actually suit a ranging strategy—back to the most basic question. Perfect parameters won't save a grid if you called the direction wrong.
Parameters are ultimately a tool in service of your market call. Set every number right but bet on the wrong market and the grid still loses; call the market right and only then does tuning the parameters mean anything. Which markets you should never run a grid in, and when to shut one down by hand once it's running, is another topic just as important—we wrote it up separately in why grids lose money, and when to shut one off.
8. FAQ
How wide should the grid range be?
There's no single right number, but here's the direction: look at the price band this coin has bounced around in over the past few months, set the upper bound near that period's highs and the lower bound near its lows, then leave a little cushion on each side so you're not sitting right on a recent high or low. Too narrow and price walks out of the range easily—you either miss the run or get stuck holding. Too wide and the grids get spread thin, capital efficiency drops, and each grid's spread shrinks. Most beginners set it too narrow, because a narrow range fills often in a backtest and the return numbers look great—until price leaves that narrow band and the pretty numbers fall apart.
Are more grids always better?
No. More grids means a smaller spread per grid and more frequent fills, but each spread can get so thin it doesn't cover the two fees (buy and sell). Past a certain density each grid is actually negative—the busier the bot, the faster the account shrinks. Fewer grids means a bigger spread per grid and real profit, but slower fills. The bottom line for grid count: keep each grid's spread clearly larger than the two round-trip fees. Use the grid spacing calculator to see the spacing first, then the grid profit simulator to confirm the per-grid net is actually positive.
What happens if the per-grid investment is too small?
Per-grid investment equals total capital divided by grid count. Too small and two things go wrong: first, the absolute dollars each fill earns are so tiny they're not worth it—after two fees there may be almost nothing left, or it goes negative; second, it can fall below Binance's minimum order size, so the grid won't create at all. When you don't have much capital, use fewer grids and make each grid's investment thicker rather than running a very dense grid on small money.
Can I use Binance's auto-suggested parameters as-is?
Use them as a reference, don't copy them blindly. The suggested range is usually computed from the coin's recent volatility, which quietly assumes it'll keep ranging the way it just did. But markets change—three months grinding inside a box doesn't mean next month won't break out in one direction. Treat the suggestion as a starting point, then run the range and grid count past your own read of the coin, especially the upper and lower bounds—those you're better off setting yourself.
This article is not investment advice. Crypto prices are volatile; grid trading can't guarantee a profit and carries the risk of getting stuck holding, plus liquidation risk on futures. Assess your own situation and only risk what you can afford. The price ranges and percentages here are illustrative; real figures and all limits go by whatever Binance's page shows at the time. Reference: Binance Support Center.