Trading Basics

A trading bot is structured in two layers—bot-level + per-pair-level. Strategy, K-line timeframe, and execution config sit at the per-pair level; execution mode, exchange, and the bot-level safety net (overall TP/SL, default leverage) sit at the bot level. Covers core elements, relation to backtest, the flow, and what to watch out for.

Trading execution is "strategy + execution config" running on a trading bot—it produces buy / sell signals from real-time market data and, based on the configured position, stop-loss, take-profit, max drawdown, and other execution config items, actually places orders, manages positions, and applies risk control either on the exchange or in a paper environment. The bot itself uses a bot-level + per-pair-level two-layer structure: the bot level decides exchange, execution mode, and global risk; the per-pair level decides the bound strategy, K-line timeframe, execution config, and the capital allocated to that pair—the bot's total capital usage is the sum of per-pair allocations. This page covers the core concepts of trading execution, how it relates to strategy and backtest, an overview of this module, and the recommended flow; for specific features see Trading Modes, Live Trading, Paper Trading, etc.

Core elements

ElementLayerDescription
Execution modeBot levelLive trading (real capital) / paper trading (virtual capital); see Trading Modes
Exchange and APIBot levelLive requires connection and API authorization in Exchange Config
Bot-level safety netBot levelOverall take-profit, overall stop-loss, and default leverage—act as a safety net alongside per-pair execution config; see Risk Management
Strategy (signals)Per pairProvides buy / long & sell / short conditions (indicators, thresholds, combinations); see Strategy Basics
K-line timeframePer pairThis pair's execution main timeframe—the engine evaluates signals at this period
Execution config (risk)Per pairPosition, stop-loss, take-profit, max drawdown, scale-in, and other runtime parameters; see Execution Config
Per-pair capitalPer pairInitial capital allocated to this pair (part of the execution config)—the bot's actual capital usage = the sum of per-pair allocations

Relation to strategy and backtest

  • Strategy: only describes "when to buy, when to sell"; the trading bot is its "running form"—the bot uses its capital and execution config to actually place orders.
  • Backtest: replays the same "strategy + execution config" on historical data to validate performance; backtest and bots share the same execution-config concept, so reusing the configuration tuned in backtest on the bot makes attribution easier.
  • Recommended flow: Create Strategy → Backtest validation → Paper Trading (1–2 weeks) → Small live → Scale up based on performance.

This module at a glance

PageContent
Trading ModesThe difference between execution mode (live / paper) and trading mode (Strategy / AI / News-Event), and how to pick
Live TradingPreparation, ways to start, and runtime monitoring with real capital
Paper TradingWhat Paper Trading is and how to use it
Auto TakeoverTake over positions you opened manually on the exchange and manage exits by preset rules
Trade StatisticsFill records, statistics, and review
Risk ManagementTwo-layer risk control (bot level + per-pair level) configuration and monitoring / alerts

For the entry point and full steps to create and start a trading bot, see Live Trading (workflow), where you can pick Live or Paper.

Execution flow overview

Strategy and backtest ready

Create Strategy and Backtest are done; the target pair behaves stably on historical data, and the backtest produced a sensible execution config.

Create a bot (bot-level configuration)

In Live Trading, create a new bot. Set name, exchange, and bot-level safety-net risk controls, and pick Live Trading or Paper Trading; for live, the exchange API must already be configured.

Configure each trading pair (per-pair-level)

Add pairs to the bot one by one; for each, bind a strategy, pick a K-line timeframe, and fill in the execution config (initial capital, position, stops, max drawdown, scale-in, etc.). The pair's initial capital is what this pair occupies on the bot, so the bot's total capital usage = the sum across pairs. One bot can manage multiple pairs independently.

Run and monitor

The bot runs each pair on its own signals and execution config. Watch fills and P&L through Trade Statistics and the dashboard; the two-layer controls in Risk Management cover both per-trade and overall risk. If you also have manual positions on the exchange, Auto Takeover can manage their exits.

Adjust and iterate

Use Monitor & Adjust to review performance; tune the strategy or per-pair execution config as needed—saved changes take effect per the system's prompt (e.g. reload or restart), or go back to backtest for re-validation before redeploying.

Things to watch out for

  • Paper before live: once the backtest passes, run Paper Trading for 1–2 weeks or more before switching to Live to keep trial-and-error costs down.
  • Keep configurations consistent for attribution: keep paper and live using the same pair, K-line timeframe, and execution config you used in backtest; reassess risk and document the reason for any change.
  • One-to-one management: under one exchange, a trading pair can only be managed by one bot; a bot can manage multiple pairs.
  • Capital and mindset: only put live capital you can afford to lose; start small; be prepared for short-term drawdowns.

Common questions

How does trading execution differ from a backtest? A backtest simulates on historical data and doesn't involve real capital or real orders; execution depends on real-time market data and the exchange API, with real or simulated orders and risk control.

Do I have to backtest before trading? Strongly recommended: validate via backtest → run paper for 1–2 weeks → small live, then scale. Skipping any of these steps before going live noticeably raises the risk.

Where is the execution config set? In the per-pair config under each trading bot (position, stops, max drawdown, scale-in, etc.); see Execution Config. The same strategy can run with different execution configs on different bots / pairs—they don't interfere with each other.

When do my changes take effect? Strategy and per-pair execution config changes take effect per the system's prompt (e.g. reload or restart the bot) once saved; historical fills are unaffected. The exact prompt is whatever the system shows.

Next steps