Scan to download
Download APKBinance's grid bot backtest feature lets you simulate how a specific grid strategy would have performed using historical price data—before committing real money. You define your parameters (price range, grid count, investment amount), and the system runs through historical candles to show what your returns would have looked like.
The core logic: every time the historical price hits a grid line, the system records a buy or sell trade. All these simulated trades are then aggregated to produce your backtest results.
Backtesting is the most important tool for validating grid parameters. Most beginners who lose money on grid bots do so because they skipped this step and set arbitrary parameters. In 2026, Binance's backtest is fully integrated into the strategy trading interface and supports custom time periods and arithmetic/geometric grid switching.
You need a Binance account to use backtest. Sign up with referral code BNAPP for a fee discount: Register here →
Your backtest parameters should exactly match what you plan to run live—otherwise the results are meaningless.
Major pairs (BTC/USDT, ETH/USDT, BNB/USDT) have the most complete historical data and produce the most reliable backtest results. Altcoins have extreme backtests that don't reflect realistic live conditions.
The most critical parameter. Approaches:
Example: BTC/USDT at ~$90,000, oscillating between $80,000-$100,000 in the past 6 months. A reasonable range: Lower $75,000 / Upper $105,000 with safety margin.
| Grid Count | Profit/Grid | Trade Freq. | Best For |
|---|---|---|---|
| 10-20 | High | Low | Small capital ($100-500) |
| 20-50 | Medium | Medium | Medium capital ($500-5000) |
| 50-100 | Low | High | Large capital ($5000+) |
| 100+ | May not cover fees | Very high | Not recommended |
Rule of thumb: each grid's price move should generate at least 3x your round-trip fee. At 0.1% per side (0.2% total), each grid should represent at least 0.6% price movement.
Recommendation: use geometric grids for most scenarios in 2026, especially for 3+ month strategies.
Enter your actual planned investment. This affects absolute profit numbers, not the percentage APR. Use your real planned amount for the most useful results.
The annualized rate of return from grid arbitrage alone. This excludes price movement of your holdings. A Grid APR above 20% is solid; above 50% should raise eyebrows about overfitting.
Total realized profit (in USDT) from all buy-low/sell-high trades during the backtest period. This is money already locked in—unaffected by current price.
The floating profit/loss of your remaining holdings at the end of the backtest period. If you bought an altcoin that dropped 50%, this will be a large negative number.
Total Return = Grid Profit + Unrealized PnL. Many beginners celebrate a high Grid APR without noticing a large negative Unrealized PnL that makes the total return negative.
Total buy/sell cycles completed. High count = the price was actively oscillating within your range. Low count = price moved directionally; grid trading was ineffective.
The largest peak-to-trough decline in total portfolio value during the backtest. Over 30% drawdown warrants reconsidering risk tolerance.
Grid Profit + Unrealized PnL, annualized. This is the definitive metric for evaluating strategy quality.
Live results are almost always below backtest results. Key reasons:
Backtest assumes exact fills at grid prices. In live trading, market orders experience slippage. For illiquid pairs, this can be 0.1-0.5% per trade—significant when compounded across hundreds of arbitrage cycles.
During fast price moves, the bot may not execute exactly at grid prices, resulting in missed fills or worse prices.
Backtests capture extreme events but handle them cleanly. In live trading, a flash crash may trigger stop-losses or push the price out of range before the bot can react.
Backtest defaults to 0.1% standard fee. If you enable BNB fee deduction (0.075%), your live results will actually be slightly better than the default backtest. This is one case where live can beat backtest.
Historically, live performance is 60-80% of backtest results. A 30% annual Grid APR in backtest translates to roughly 18-24% in live trading.
Run the same parameters across three different periods: bull market, bear market, and sideways market. A robust strategy should perform reasonably across all three—not just the favorable period.
Test 20/50/100 grids with the same price range. Compare net profit (after fees) and arbitrage count. Find the sweet spot where more grids stops being beneficial.
Start with a wide range (±40% from current price). Identify which sub-range had the most price activity. Concentrate capital in that active sub-range to improve density and profitability.
Run both types with identical parameters and compare net grid profit. Geometric typically wins on volatile assets over longer periods.
A 100% Grid APR looks amazing until you see the Unrealized PnL is -80%. Always check the total combined return.
Optimizing parameters against historical data creates overfitting. The parameters that performed best historically may be poorly suited for the future. Choose robust parameters, not the historically optimal ones.
If you only test the last 3 months of a bull run, of course it looks great. Always test across different market conditions, including periods when the asset dropped 30%+.
More grids = more fees. Once the fee per trade approaches or exceeds the profit per grid, adding more grids reduces net returns. The backtest will show this clearly in the fee summary.
Too narrow: bot stops when price breaks out, capital sits idle
Too wide: low grid density, infrequent trades, poor capital efficiency
Target the core oscillation zone, not a range that "covers all possibilities."
Grid bots have no automatic stop-loss. If price drops below your lower limit, the bot keeps buying. Set price alerts and have a manual exit plan if price breaks 10%+ below your lower bound.
BTC/USDT grid strategy backtest (90-day period, example):
| Grid Count | Grid APR | Trades | Total Fees | Net Grid Profit |
|---|---|---|---|---|
| 20 | 28.1% | 128 | $24 | $821 |
| 50 | 34.2% | 312 | $61 | $845 |
| 100 | 31.6% | 608 | $119 | $793 |
Conclusion: 50 grids produces the highest net profit in this example. 100 grids generates more trades but fees overwhelm the marginal gains.
90 days is the most commonly used reference period—long enough to cover a full market cycle, short enough to reflect current conditions.
Use it as a starting point to see what the system suggests, then run your own backtest to compare. Never deploy AI auto parameters directly to live trading without understanding the underlying logic.
Treat altcoin backtests with heavy skepticism. A 1000% Grid APR backtest on a meme coin often means the coin was in a perfect pump-and-dump oscillation that will not repeat. Stick to BTC, ETH, BNB for reliable grid strategies.
Binance doesn't currently offer export for backtest results. Screenshot your parameters and results for future reference and to compare against live performance.
Found parameters you're satisfied with in backtest? Create a Binance account with referral code BNAPP for fee discounts, then apply your tested parameters live:
The spot grid bot is one of Binance's most beginner-friendly automated strategies. Running backtest first dramatically improves your chances of success.
Sign up with referral code BNAPP for lifetime fee rebate