Risk Management

The system uses two-layer risk control:

1. Per-pair risk control (execution config): configured per pair when creating a backtest or when creating / editing a bot—see Execution Config—covering take-profit / stop-loss, profit protection, position, holding duration, add-on, capital management, and consecutive-loss management;
2. Bot-level safety-net risk control: set when creating / editing a bot—overall take-profit / stop-loss and default leverage—to provide an overall safety net.

Backtest, live, and paper runs all enforce these settings automatically. The two layers run in parallel: whichever trips first triggers the corresponding risk-control action. This page covers each layer's config entry points, scope, and usage tips. For concepts see Trading Basics.

---

Where risk control is configured

Risk control has two main configuration entry points:

We recommend completing the per-pair execution config first, then setting the bot-level safety-net thresholds.

---

Per-pair risk control (execution config)

Per-pair risk control is the core; you can think of it as a "six-dimension protection matrix" (field names follow the system UI):

"Consecutive-loss management" sits in the Capital Guard dimension of the execution config and is one of the core per-pair risk controls.

Execution-engine view (how it lands at runtime)

Per-pair risk control can be viewed as four execution layers:

• Entry capital allocation layer: single entry vs. scaled entry—decides how initial capital is used.
• Dual-core risk layer: fixed TP/SL + profit protection—decides close logic.
• Dynamic position layer: holding-duration management + dynamic add-on—decides how risk evolves while holding.
• Capital protection layer: capital circuit breaker + profit protection + consecutive-loss stop—decides whether the strategy keeps running.

The four layers work together so the bot can extend profits in trending phases while keeping drawdowns under control elsewhere.

---

Bot-level safety-net risk control

When creating / editing a bot, you can set bot-level safety-net parameters (bot overall take-profit, overall stop-loss, default leverage). This doesn't replace per-pair risk control—it's the last line of protection at the bot level. The bot's total capital usage is the sum of the per-pair allocations and has no separate "total cap" configuration.

You can think of it this way:

• Per-pair level: controls how each trade and the holding process behave for that pair.
• Bot level: caps how much risk the entire bot can take on.

How the two layers work together

The two layers act as parallel constraints:

• Whichever layer's condition trips first runs its risk-control action (e.g. close / pause).
• The per-pair level leans toward "controlling trading behavior"; the bot level leans toward "guarding overall thresholds".
• When the same strategy runs on different bots / pairs, each instance's execution config can be tuned independently, while bot-level thresholds can be set per account or risk preference.

---

Configuration tips

• Execution config first: when creating a backtest or bot, set the per-pair take-profit / stop-loss, position, capital management, and consecutive-loss thresholds first, then optimize the buy / long & sell / short conditions.
• Bot thresholds as a safety net: use bot-level take-profit / stop-loss for overall protection—try not to set them too far from the per-pair thresholds.
• Keep backtest / paper / live aligned: keep both layers consistent during validation to make attribution and comparison easier.
• Adjust by layer: if the problem is per-trade behavior, adjust the per-pair level first; if portfolio-level drawdowns exceed expectations, tighten the bot-level thresholds.

---

Monitoring and review

• Live monitoring: use the dashboard and Trade Statistics to watch current drawdown, positions, per-trade P&L, consecutive-loss state, and odd fills.
• Alerts: if the system supports drawdown-limit, consecutive-loss, or failed-trade alerts, enable them in Notification Config and act in time.

Auto Takeover manages manual positions you opened on the exchange—it is a separate feature from strategy / bot risk control.

---

Usage tips

• Per-pair first, then bot-level: finish the execution config first, then add the bot's safety-net thresholds to avoid parameter conflicts.
• Always set consecutive-loss limits: set sensible consecutive-loss caps per pair to keep losses from spreading.
• Small-step tuning: change only a few key parameters at a time (e.g. stop-loss or consecutive-loss threshold); changing too much at once makes attribution hard.
• Pre-live cross-check: before going live, confirm both the execution config and the bot safety-net parameters match your current risk tolerance.

---

Common questions

Is all risk control inside the execution config? No. Per-pair execution config is the core; when creating a bot there's also bot-level safety-net risk control (overall take-profit / stop-loss and default leverage).

Which layer does consecutive-loss management belong to? The capital-management module of the per-pair execution config—it stops trading once the consecutive-loss count is reached.

What if the two layers conflict? Think of it as "first to trip, first to act"; in general, leave detailed control to the per-pair level and let the bot level only act as an outer safety net.

Do I need both layers in backtest? At minimum, fully configure the per-pair risk control; if you plan to use the bot safety net later, configure and validate both before paper / live.

---

Next steps

• Execution Config — Per-pair risk control (TP / SL, position, consecutive-loss, etc.)
• Live Trading (workflow) — Create a bot and set the safety-net risk control
• Trade Statistics — Review and monitor risk-control triggers
• Monitor — Iterate parameters based on performance
• Trading Basics, Backtest Basics — Concepts and validation