Dimanche, café froid, un bot qui tourne depuis mardi —
No Lambo screenshots. No “passive income” promise. Just my working notes on grid bots — what I set, what broke, what the fees really ate, and the maths I wish someone had shown me before my first bot bought all the way down.
je teste d'abord sur papier. Toujours. Ça m'a déjà évité au moins trois grosses bêtises.
fig. 1 — le prix fait des allers-retours, le bot ramasse les miettes. Quand ça casse, ça casse.
à toi de jouer — outil gratuit
Set a range, a grid count and a budget — then watch a simulated price wander inside your grid. Every time it crosses a level, the bot buys or sells and banks the spacing, minus fees. This is a random-walk illustration for learning, not a prediction of any real market.
Rule of thumb this teaches: if your spacing % is not comfortably above 2× your fee, the exchange earns more from your bot than you do.
mes notes, dans l'ordre où je les aurais aimées
Every guide is written from the mechanics up — how the bot actually places orders, what the numbers mean, and where beginners lose money. No signal groups, no “10x bot settings,” no sponsored rankings.
en vrai, comment ça marche ?
Pick a lower and upper price. Everything the bot does happens inside that box — nothing outside it.
The range is divided into equal levels (“grids”). Each level gets an equal slice of your capital.
Buy orders sit below the current price, sell orders above it. A filled buy instantly places a sell one level up — and the reverse.
Each completed round trip banks the grid spacing minus fees. Sideways chop is profit; a breakout ends the game.
sur mon bureau, en ce moment
The difference between a hobby and a money leak is written records. These are the exact sheets referenced in the guides — the starter sheet is free.
One page to fill in before you press “start” on any bot: range logic, spacing vs. fees check, exit plan for both breakouts.
Reference cards for the six decisions that matter — range width, grid count, arithmetic vs. geometric spacing, trailing, stop conditions — for BTC, ETH and large-cap alts.
A spreadsheet that separates the three numbers platforms blur together: grid profit, price appreciation, and fees — per bot, per month, export-ready for tax time.
où faire tourner tout ça ?
There are only three real choices: the grid bot built into your exchange, a third-party bot platform that connects by API, or a free built-in specialist. What matters is fees, custody and whether the bot keeps working if the company disappears.
| Type | Examples | Cost shape | Best for | Watch out for |
|---|---|---|---|---|
| Exchange built-in | Binance, Bybit, KuCoin, OKX spot & futures grids | No subscription — you pay normal trading fees per fill | Beginners: nothing to install, funds stay in one account | Fewer controls; settings and stats vary a lot by exchange |
| Free built-in specialist | Pionex (exchange with 16+ built-in bots) | No bot fee — revenue comes from trading fees | Running several small grids cheaply | Your funds live on a smaller exchange — size positions accordingly |
| Third-party platforms | 3Commas, Cryptohopper, Bitsgap | Monthly subscription plus your exchange's trading fees | Multi-exchange portfolios, backtesting, advanced triggers | API keys are an attack surface — use trade-only keys, never withdrawal |
| Self-hosted / open source | Freqtrade and similar frameworks | Free software, your server and your time | Developers who want full control and auditability | You are the support desk, the security team and the uptime guarantee |
Grid trading is not passive income and it is not safe. A grid bot in a falling market keeps buying all the way down and ends up holding a full bag of a depreciating asset; on futures, the same move can liquidate the account. Everything on TradingGrids.com is education about how these tools work — not financial advice, and not a recommendation to trade. Never run a bot with money you cannot afford to lose, and never use leverage while you are learning.
on me demande souvent…
They make money in one specific condition: a market that oscillates inside a range long enough for many buy-low/sell-high cycles to complete. In that condition the profit is real but modest — it comes from the spacing, minus fees. In a strong uptrend the bot underperforms simply holding; in a strong downtrend it accumulates losses. The bot harvests volatility; it does not predict anything.
Enough that each grid slice clears the exchange's minimum order size with room to spare. With 20–30 grids, most people find $500–$1,000 the practical floor for a spot grid on a major pair — below that, minimum order sizes force awkward settings. Start smaller than you think, on spot, and treat the first bot as tuition.
Spot for learning, almost always. A spot grid's worst case is holding an asset that fell — painful but recoverable. A futures grid adds leverage, funding rates and liquidation: the worst case is a zeroed account. The full comparison is in the guides above.
Because platforms headline “grid profit” — the sum of completed cycles — while your total result also includes the falling value of the inventory the bot is still holding. Grid profit up + inventory down more = net loss. The Journal in the toolkit exists precisely to keep those numbers separate.
Arithmetic grids space levels by equal dollar amounts — simple, and fine for narrow ranges. Geometric grids space by equal percentages, which suits wide ranges and volatile assets because the spacing grows with price. For ranges under ~15%, the difference is minor; above that, geometric is usually the saner default.