GridCheck

Method

How we test grid bots

Every number on GridCheck comes from replaying grid settings on real Binance prices with our own simulator. Here is exactly what that means, how close it gets to Binance's own figures, and where it falls short.

1. The data

Binance spot candles from Binance's public market-data API: 15-minute candles for bot checks and the weekly top-bots page, 1-hour candles for the pair pages. Only fully closed candles are used.

2. The simulator

A grid with N levels is N bands. Each band always has one resting order: a buy at its lower edge, or a sell at its upper edge once it holds coin. At launch, bands above the current price buy coin at market so they have something to sell. We charge 0.1% per trade (Binance's standard spot fee) and 5 basis points of slippage on market buys. To stay conservative, a band can fill at most once per candle.

3. Replaying the listed run

For a Marketplace bot we take its runtime and the time the Marketplace numbers were last refreshed, and replay exactly that window. This checks that our simulator and Binance agree before we say anything else.

4. The 60-day launch test

Then we launch the same settings every 12 hours over the previous 60 days, each run lasting as long as the listed bot. A listed bot shows one lucky or unlucky start; this shows the whole spread: the typical result, the worst 5%, and how often the price was even inside the range.

5. Why we compare with holding the same coin

A grid starts by buying coin. If its range sits above the price, it buys almost 100% coin at launch and then simply holds it. Compared with buy-and-hold or a 50/50 split, that looks like skill in a rising market, but it's just the coin going up. So we compare every run with holding the same share of coin the grid bought at launch. The difference is what the grid itself added.

How close are we to Binance?

17 / 17listed ROIs reproduced within about 2 points; median gap 0.6 (8 Oct 2026 capture)
14 + 4of 18 earlier snapshots matched or came close, none far off (30 Sep)
+12.17% vs +12.19%ZEC/USDT: Binance's figure vs our replay of the same run

Small gaps remain because the Marketplace refreshes hourly (we can't know the exact minute), and very busy grids fill more often within a candle than 15-minute data can show.

Limits, plainly

What we never do

GridCheck doesn't connect to exchange accounts, never asks for API keys, never holds money and never places trades. It analyses settings you choose and past prices. It isn't financial advice, and it isn't affiliated with Binance.