On this page 11
Key takeaways
- A grid bot splits a price range into levels, buys at each lower level and sells one level higher. Profit per round trip is the gap between levels minus two trading fees.
- It earns in choppy, sideways markets. If price leaves the range, the bot stops trading and either sits in cash (breakout up) or holds coins bought on the way down (breakdown).
- Grid lines must be spaced well above the round-trip fee. At 0.1% per order, steps under about 0.6% give most of the profit to the exchange.
- Pionex and KuCoin build grid bots into the exchange; Bitsgap, 3Commas v2, WunderTrading and TradeSanta connect to your exchange by API; Freqtrade lets you code your own.
- No list of "best pairs" stays right for long. Choose pairs by liquidity, recent volatility, range behaviour and fees, and re-check before every launch.
A crypto grid trading bot is one of the oldest and most honest forms of trading automation. It does not predict anything. It assumes price will bounce around inside a band you choose and quietly harvests those bounces. That simple idea is why grid bots are the default product on Pionex, Bitsgap and most exchange bot menus, and also why so many beginners lose money with them: the strategy works beautifully until the market stops cooperating.
Below: a worked example, the platforms compared, and criteria for picking pairs and ranges. For every bot type, see our ranking of the best AI crypto trading bots.
What is a grid bot in crypto trading?
You give the bot four inputs: a lower price, an upper price, a number of grids and the capital to use. The bot divides the range into evenly spaced levels and places a limit buy order at each level below the current price and a limit sell order at each level above it. When a buy fills, the bot immediately places a sell one level higher. When that sell fills, it re-places the buy one level lower. Each completed buy-then-sell pair is a "round trip" and books a small, fixed profit.
That is the whole engine. There is no forecasting, no indicator and no learning. Some platforms add an "AI" range suggestion based on recent volatility, which is a parameter helper, not a trading brain. Our guide on how AI crypto trading bots work explains where real machine learning starts.
Because a grid bot buys as price falls and sells as it rises, it is a mean-reversion strategy. It does well when price keeps returning toward the middle of a band and poorly when price trends hard in one direction.
A worked example with real numbers
Here is a small, fully computed example. The numbers are illustrative, not a forecast or a backtest. Assume $1,000 on a BTC/USDT spot grid from $60,000 to $66,000 with 6 grids, so levels sit every $1,000 and each grid gets about $166.67. The fee is 0.1% per order. At launch BTC is at $63,000, so the bot buys coins for the three grids above the price and places buy orders at $62,000, $61,000 and $60,000.
Illustrative 6-grid BTC bot, $1,000, 0.1% fee per order (our own calculation)
| Price moves to | What fills | Profit booked | Running total |
|---|---|---|---|
| $61,800 | Buy at $62,000 | – | $0.00 |
| $63,100 | Sell at $63,000 | +$2.35 | $2.35 |
| $60,500 | Buys at $62,000 and $61,000 | – | $2.35 |
| $62,200 | Sell at $62,000 | +$2.40 | $4.75 |
| $64,300 | Sells at $63,000 and $64,000 | +$2.35, +$2.31 | $9.41 |
| $67,000 | Sells at $65,000 and $66,000 | +$4.88, +$7.36 | $21.65 |
Three things stand out. First, a clean round trip inside one grid earns about $2.35 to $2.40 on $166.67, roughly 1.4% after fees. Second, the last two sells earn more because they close the coins bought at launch, which is really just profit from holding. Third, look at the final row: price is now above $66,000, the bot holds only USDT and stops trading. Someone who simply bought $1,000 of BTC at $63,000 would be up about $63.49 before fees, almost three times the bot's $21.65.
Now run it the other way. If BTC drops straight from $63,000 to $55,000, every buy order fills and the bot ends up holding about 0.01601 BTC worth roughly $880.66, a loss of about $119 (11.9%). Plain holding would have lost about $127 (12.7%). The grid softened the fall a little because half the money was deployed lower, but it did not protect you.
Arithmetic vs geometric grids
Arithmetic grids space levels by the same dollar amount. Geometric grids space them by the same percentage. Over a narrow range the difference is small. Over a wide range it matters, because a fixed $1,500 step is a bigger percentage at the bottom of the range than at the top.
| Range $55,000–$70,000, 10 grids | Arithmetic (equal $) | Geometric (equal %) |
|---|---|---|
| Step between levels | $1,500 every time | 2.44% every time |
| Gross profit per grid, bottom | 2.73% | 2.44% |
| Gross profit per grid, top | 2.19% | 2.44% |
| Net per grid at 0.1% fee | 1.99%–2.53% | 2.24% |
| Best for | Narrow, stable ranges | Wide ranges and volatile coins |
A rule of thumb: if your upper price is more than about 30% above your lower price, use geometric spacing, so the grids near the top are not starved of profit. The grid simulator below lets you switch between the two and see profit per grid change.
Spot, futures, infinity and reverse grids
Exchanges and bot platforms now ship several grid variants. They share the same ladder logic but differ in what you hold and how much you can lose.
- Spot grid. The classic version. You hold the coin and stablecoins, with no leverage, so the worst case is holding a coin that has fallen. Always the right place to start.
- Futures grid. Runs on perpetual futures in long, short or neutral mode, usually with leverage. You can profit from a falling range, but a breakout against you can hit the liquidation price and wipe the margin. Our futures trading bots guide covers liquidation math.
- Infinity grid. Has a lower limit but no upper limit. As price rises it keeps selling small slices so the value of your coin position stays roughly constant. You stay invested in a long uptrend, but each grid earns less and a fall still hurts.
- Reverse grid. Starts holding the coin instead of stablecoins and aims to end with more coins, by selling into rallies and buying back lower. It is meant for people who expect a decline but want to keep their coin exposure. Profit is counted in the coin, not dollars.
Pionex lists the widest menu we found: spot, reverse, infinity, leveraged, margin, futures, Coin-M and cross-margin futures grids, plus a leveraged reverse grid. KuCoin offers Spot Grid, Futures Grid and Infinity Grid, and Binance's own posts describe Spot Grid and Futures Grid bots.
Where grid profit and loss really come from
Platforms usually show "grid profit" in big green numbers. That figure counts only completed round trips. Your real result has three parts:
- Grid profit: step size minus fees, times the number of round trips. This is the part the bot controls.
- Unrealised P&L: the gain or loss on coins the bot is holding right now. After a drop this can dwarf grid profit.
- Fees: two per round trip. They are already inside grid profit, but on tight grids they can eat most of it.
The math for one grid is: net profit per round trip ≈ (upper level − lower level) ÷ lower level − 2 × fee rate. With a 1% step and a 0.1% fee you keep about 0.8%. With a 0.3% step you keep 0.1%, and two thirds of your gross profit goes to the exchange. That is why "more grids" is not always better.
Where grid bots shine
- Sideways markets with frequent swings of a few percent
- Liquid pairs with low fees and tight spreads
- Fully mechanical, so no emotional decisions
- Easy to audit: every order is visible in your exchange history
Where they fail
- Strong uptrends: the bot sells early and ends in stablecoins
- Crashes below the range: you hold coins at a loss
- Tight grids on high-fee venues: fees take most of the profit
- Leveraged futures grids: liquidation if price runs against you
Try your own range in the grid simulator
The simulator below uses your range, grid count, spacing and fee to calculate net profit per grid, then runs the bot over a random hourly price path. It is a simulation, not a backtest on real history, so use it to understand behaviour rather than to predict returns. Press "Simulate another path" a few times and watch how often the price escapes the range.
Grid bot simulator
Simulated, not a backtestFor formulas, fee break-even and more worked examples, open the full grid bot profit calculator.
Best crypto grid trading bot platforms compared
There are two ways to run a grid bot. Exchange-native bots (Pionex, KuCoin, Binance) run inside the exchange, so there are no API keys to leak, but your funds and the bot live in one place. Third-party platforms (Bitsgap, 3Commas, WunderTrading, TradeSanta) connect to your exchange by API and let you run grids on several venues from one dashboard. Freqtrade is the do-it-yourself route.
| Platform | Grid types | Cost | Where it runs | Watch out for |
|---|---|---|---|---|
| Pionexpionex.com ↗ | Spot, infinity, reverse, leveraged, futures | Free bots; 0.05% spot fee | Is the exchange | Custodial; licensing hard to verify |
| Bitsgapbitsgap.com ↗ | GRID and COMBO bots | Free plan; Basic $29/mo | 16 exchanges via API | Return claims are vendor marketing |
| 3Commas v23commas.io ↗ | Grid bots, spot and futures | Starter $20/mo (2 grid bots) | 10 exchanges via API | v1 bots did not migrate in Sep 2026 |
| WunderTradingwundertrading.com ↗ | GRID, Multi-Pair GRID | Free plan; Basic $19/mo | 20 exchanges via API | Prices appear geo-localised |
| KuCoinkucoin.com ↗ | Spot, futures, infinity grid | Built in; 0.1% spot fee | Is the exchange | Not open to new US users |
| Binancebinance.com ↗ | Spot grid, futures grid | Built in; 0.10% spot fee | Is the exchange | EU services restricted since Jul 2026 |
| TradeSantatradesanta.com ↗ | Grid and DCA, spot and futures | From $25/mo | Via API | Full exchange list not published clearly |
| Freqtradefreqtrade.io ↗ | No built-in grid; you code it | Free (GPL-3.0) + server | Your own server | Needs Python skills |
Pionex: the cheapest place to start
Bots are free and spot fees are 0.05% per side, half of what many exchanges charge, which matters a lot for grid strategies. The catch is that you deposit onto Pionex itself. Read the Pionex review for the custody and licensing questions.
Bitsgap: the most polished multi-exchange option
Bitsgap runs GRID and COMBO bots on 16 exchanges, offers a free plan and a 7-day Pro trial without a card, and rejects API keys that have withdrawals enabled. Its "11% average 30-day grid return" is a vendor claim we could not verify. Details in our Bitsgap review.
3Commas v2 and WunderTrading
3Commas v2 was built with WunderTrading after v1 shut down on 11 September 2026; Starter includes two spot grid bots. WunderTrading adds a Multi-Pair GRID across several pairs. See the 3Commas review.
Keep your long-term coins somewhere boring
Grid bots need a working balance on an exchange, but your long-term holdings do not have to sit next to an experimental bot. CEX.IO has operated since 2013, is a FinCEN-registered MSB with money transmitter licences in 39 US states plus DC, and is FCA-registered in the UK. It has no grid bots of its own, which makes it a simple place to buy and hold while you test strategies elsewhere.
Registrations checked October 2026 on the company's legal pages. Availability depends on your country; crypto is high-risk.
- Since 2013One of the longest-running crypto exchanges still operating
source - 39 US states + DCFinCEN-registered MSB with state money transmitter licences (NMLS 1804170)
source - FCA-registeredUK cryptoasset business under the MLRs, FRN 1007192
source - PCI DSS Level 1Card-payment security standard, 2026 certificate published
source
How to choose pairs and ranges (and why "best pairs" lists go stale)
Searches for the best crypto pairs for a grid trading bot return lists of coins with impressive past returns. Treat them with suspicion. A pair is good for grids only while it behaves a certain way, and that behaviour changes. A coin that chopped sideways for two months can start trending the week after the list is published. Any list based on last month's grid profit is a snapshot of luck, not a property of the coin. We deliberately do not publish one.
What you can do is check the conditions that make a grid work, every time you launch a bot:
| Criterion | What to check | Weight (our view, 0–5) |
|---|---|---|
| Liquidity | Deep order book and tight spread, so limit orders fill at your levels | |
| Fee tier | Round-trip fee well below the grid step; maker rebates or low-fee venues help | |
| Range behaviour | Price has repeatedly bounced between similar highs and lows recently | |
| Volatility | Daily swings bigger than your step, but not so wild price leaves the band | |
| Trend strength | No strong one-way trend; news-driven coins break ranges often |
Setting the range
Look at the last few weeks on a daily chart and mark where price repeatedly turned. Set the range a little wider than those turns so normal noise does not push price out. A wider range keeps the bot trading for longer but spreads capital thinner and lowers profit per grid. A narrow range earns more per trade until the first breakout.
Mean reversion is the real test
Large, liquid pairs often fit better than small caps, not because they are "safer" but because their swings tend to be more two-sided. Stablecoin pairs barely move, so the grid never fills. Freshly listed tokens can fall sharply within hours and leave you holding the whole position. For more strategy ideas, see our AI crypto trading strategies guide.
Grid trading bot strategy: settings checklist
Before you press start, walk through these settings. Most grid losses we read about in user reviews trace back to skipping one of them.
Pick spot first
Run your first grids on spot with no leverage. Move to futures only once you understand how a breakout feels on your P&L.
Check the fee and step
Look up your real fee tier, double it for a round trip and make sure the grid step is at least three to five times that number.
Choose spacing
Geometric for ranges wider than about 30%, arithmetic for narrow, stable ranges.
Set grid count from the step, not the other way round
Decide the step you need, then divide the range by it. Also check each order meets the exchange minimum order size.
Add a stop and a take-profit
Use the platform's stop-loss below the range and decide what happens on a breakout above: stop, trail the range, or hold.
Protect the account
For API-based platforms, create a trade-only key with withdrawals disabled, IP whitelisting on, and keep the bot in a sub-account.
Paper trade, then size small
Use demo mode for a week or two, then start with an amount you could watch fall 30% without changing your life.
Judge on total profit and loss
Compare the bot with simply holding the same coins. If it consistently lags, the market is not a grid market.
If grid trading is not a fit for your market view, compare it with recurring buys or safety-order strategies in our crypto DCA bots guide, or start with the low-cost options in free AI crypto trading bots. Nothing here is financial advice; grid bots can and do lose money.
Frequently asked questions
Are crypto grid trading bots profitable?
They can be in a market that moves sideways inside your range with enough swings to fill orders, and when the gap between grid lines is well above your round-trip fee. They lose money, or badly lag simply holding, when price breaks out of the range and stays out. Vendor return figures are marketing and cannot be independently verified, so test your own settings in demo mode first.
Can you lose money with a grid bot?
Yes. If price falls below your lower limit, the bot ends up holding coins bought on the way down, and that open position can be worth far less than you put in. On a futures grid, leverage adds liquidation risk on top. Grid profit is often shown separately from the unrealised loss, so always look at total P&L, not just "grid profit".
How many grids should I use?
Work backwards from fees. Divide your range by the number of grids to get the spacing, then make sure each step is at least three to five times the round-trip fee. With a 0.1% fee per order (0.2% round trip), a step under roughly 0.6% leaves too little after costs. More grids mean more trades but thinner profit per trade and less capital per order.
What is the best pair for grid trading?
There is no permanent answer. A good grid pair is liquid, swings often within a band and tends to return toward a middle price. Those traits change with the market, so any "best pairs" list goes stale within weeks. Check recent volatility and range behaviour yourself before each new bot, and favour pairs with tight spreads and low fees.
Is a grid bot better than a DCA bot?
They suit different markets. A grid bot profits from repeated swings inside a range. A DCA bot either buys on a schedule for long-term accumulation or averages down into a dip and sells on a bounce. In a strong trend neither is magic. See our DCA bots comparison for how the two DCA styles work.
Is a Pionex grid bot safe to use?
Pionex bots run on the exchange itself, so you do not hand API keys to a third party. The trade-off is custody: your funds sit on Pionex, its licensing is hard to verify and its Trustpilot score was 2.6/5 when we checked in October 2026. Read our Pionex review before depositing more than you can afford to have tied up.
Should I use a spot grid or a futures grid?
Start with spot. A spot grid can lose value but cannot be liquidated. A futures grid lets you run long, short or neutral grids with leverage, which multiplies both the grid profit and the damage from a breakout. If you do try futures, keep leverage low and check where the liquidation price sits relative to your range.
Keep reading
Sources and fact-check dates
We link to primary sources wherever possible. Dates show when the page was published or when we last checked it.
- Pionex fee schedule · checked 2026-10-21
- Pionex blog: grid trading bot types · updated May 2026, checked 2026-10-21
- Bitsgap pricing and supported exchanges · checked 2026-10-21
- 3Commas pricing (v2 plan limits) · checked 2026-10-21
- 3Commas v2: everything you need to know · dated 2026-09-11, checked 2026-10-21
- WunderTrading pricing (GRID and Multi-Pair GRID) · checked 2026-10-21
- KuCoin review: bots (Spot Grid, Futures Grid, Infinity Grid) · dated 2026-03-09, checked 2026-10-21
- Binance Feed: Binance trading bot types · checked 2026-10-21
- TradeSanta pricing · checked 2026-10-21
- Freqtrade documentation · checked 2026-10-21
