What NuxGrid’s Bitcoin Grid Trading Backtest Actually Simulates

The NuxGrid backtest replays a Bitcoin grid trading strategy on historical 5-minute candles. It lets you observe buys, sells, fees, open positions and capital changes, with limits compared with live trading.

BTC / USDT (5 min)Illustration · Demonstration data
Interpreted TP sell
Simulated buy
70,00068,00066,00064,000 09:0009:3010:0010:3010:55
Simulated buyInterpreted TP sellLink between events
01 · REPLAY HISTORICAL CANDLES

How does the backtest work?

Historical closed 5-minute candles, focused on BTC_USDT and Binance-style USDT markets.

  • Existing sells: high ≥ target.
  • Buys: low ≤ level, with sufficient cash.
Model and metric details

The database backtest uses 5-minute OHLC historical candles. The public web flow is centred on BTC_USDT and Binance-style USDT markets.

For each candle, the engine first checks existing positions. An existing sell is considered triggered when the candle high reaches its target. The engine then calculates a new buy candidate. A buy is considered triggered when the candle low reaches that level and sufficient simulated cash is available.

The model does not assume the exact sequence inside a candle. A position opened during the current candle is not immediately sold during that same candle.

One 5-minute candle, step by stepIllustration · Demonstration data
Candle being analysed 09:0009:3010:0010:3010:55
  1. 1
    Read the candleOHLC · closed candle
  2. 2
    Check the levelsExisting sells, then buys
  3. 3
    Simulate executionAccording to engine rules
  4. 4
    Update the portfolioCash, positions and equity

Multiple strategies: all sells → BTC conversions → buys, in strategy order 1 → 4. A strategy that just sold does not buy again on that candle. With one strategy: conversion after each sell, then calculation of a buy. A new buy is never sold on the same candle.

02 · WHAT IS INCLUDED

What the simulation includes

Input settings, simulated buy and sell fees, and up to four strategies sharing one simulated wallet.

Model and metric details

A run can include the initial bankroll, purchase amount, percentage grid spacing, take-profit percentage, fee percentage, optional profit-to-BTC conversion, and up to four strategies sharing one simulated wallet.

NuxGrid uses closed 5-minute candles and records unavailable ranges when history is missing. Simulated buy and sell fees are applied. Conversion contributions can accumulate until the simulated exchange minimum is reached; amounts below that minimum may remain pending.

Initial capital

Starting amount

Purchase amount

Order size

Spacing

Percentage based

Take-profit

Percentage based

Fees

Buy and sell

Up to 4 strategies

Shared wallet
  • 5-minute candles
  • Optional BTC conversion
  • Missing-data detection
03 · MEASURING RESULTS

Example backtest results

Read completed cycles alongside positions still open. Realized profit is calculated after simulated trading fees.

Model and metric details

Realized profit comes from completed sell cycles after the simulated trading fees used by the engine. It does not include the market value of positions that remain open.

Final equity combines remaining cash, open positions marked at the last candle close, and BTC reserve value when applicable. Other metrics include ROI, maximum drawdown, fees paid, completed cycles, open positions, minimum cash, capital engaged, concurrent positions, unrealized loss, and blocked buy attempts. A buy-and-hold BTC comparison is also calculated.

Backtest results (example)Illustration · Demonstration data

Fictional example, no expected return. Initial capital: 10,000 USDT; no BTC conversion in this example.

Final equity
10,742 USDTCash: 9,486 USDT
Realized profit
+486 USDTCompleted cycles
ROI
+7.4%Static illustration
Maximum drawdown
−8.1%Drawdown
Total fees
−86 USDTBuy + sell
Open positions
33 positions · value: 1,256 USDT

Final equity = cash + positions marked at the last close + any BTC reserve. Realized profit covers completed cycles after fees; it excludes the value of open positions.

04 · WHAT IS NOT SIMULATED

What is not simulated

The model applies rules to OHLC candles. It does not reproduce every condition of a live exchange.

  • Order-book depth
  • Slippage
  • Network latency
  • API failures
  • Live partial fills
  • Intracandle order
  • Cancellation / execution races
05 · INTERPRETING A RESULT

How to interpret a result

A report describes a specific period and configuration. Check its assumptions before comparing results.

Checklist for interpreting a report

  • Period / market
  • Resolution (5 minutes)
  • Strategy settings
  • Fee assumption (buy and sell)
  • Data gaps
  • Open positions
  • Realized profit
  • Final equity
  • ROI / maximum drawdown
  • Buy / sell cycles