Trade Statistics
Multi-level trade statistics—from the home overview, into a bot's details, down to per-pair backtest results—with full trade records and K-line entry/exit replay for review and improvement.
Trade Statistics aggregate the trading data produced during live and paper runs and present P&L, win rate, drawdown, and other key metrics at multiple levels—home overview, bot detail, per-backtest-pair results, and individual fills. They also provide full trade records and K-line entry / exit playback, and are the core basis for reviewing performance and spotting overtrading or risk issues. The exact entry points, fields, and filters depend on the system; for concepts see Trading Basics.
Statistics levels
The system presents stats in four layers, "from the whole to the detail"—from a one-glance account-wide view on the home page down to per-trade P&L for a single pair in a backtest:
| Level | Scope | Main content | Use |
|---|---|---|---|
| Home overview | All bots on the account | Total assets, total P&L, today's P&L, number of running bots, overall return curve | Get a single view of the account's performance and run state |
| Bot stats | A single trading bot | Initial / current capital, cumulative P&L, win rate, return curve across all pairs on the bot, plus per-pair run state | Decide whether the bot is meeting expectations, and whether to adjust or stop |
| Per-backtest-pair stats | One pair under a backtest task | The pair's total trades, win rate, profit factor, max drawdown, and return curve during the backtest, plus full fills and K-line entry / exit markers | Judge how that pair behaves on historical data during Result Analysis |
| Trade records & entry / exit points | A single fill or K-line point | Time, price, size, P&L of each entry / exit, and the trigger condition; entry / exit markers can be replayed on the K-line chart | Bar-by-bar review of signals and execution; find problematic trades |
Paper and live use the same statistics structure and fields—only the capital and fills are virtual / real respectively; backtest stats use a similar structure, which makes cross-comparison easy.
Full trade records
Every fill generates a standalone trade record, viewable on the bot detail page, the backtest result page, or the trade statistics page. Common fields:
| Field | Description |
|---|---|
| Time | Entry / exit time, to the second |
| Pair and direction | e.g. BTC/USDT long / short, buy / sell |
| Entry / exit price, size, amount | Actual fill price, size, and notional amount |
| Fees / P&L | Fees on that trade, net P&L, and P&L ratio |
| Holding time | Time from entry to exit |
| Trigger condition | The strategy buy / long & sell / short condition that triggered the trade—useful for checking the trade executed as intended |
You can filter and sort by time range, pair, strategy, direction, P&L sign, etc., to zoom in on the trades you care about.
K-line entry / exit playback
On a pair's K-line chart you can overlay the entry / exit points for the selected time window or backtest task, and play them bar by bar to see:
- Entry timing: was the entry at a reasonable spot (e.g. trend start, breakout, pullback)?
- Exit timing: was the exit triggered by a strategy condition, take-profit, stop-loss, or another risk control?
- Market context: price action, volatility, and adjacent indicator state at the entry / exit moment.
Playback is available in backtest, paper, and live; in backtest you can also compare how different parameter sets exit the same market, helping you find more robust parameters.
Using statistics and records for improvement
- Return and drawdown: start with the equity curve and drawdown at the home and bot level to gauge whether live / paper performance is close to the backtest; if it keeps diverging, check whether the execution config, slippage, and fee assumptions match.
- Trade behavior: watch trade frequency, win rate, profit factor, and holding-time distribution to spot overtrading, stops set too tight, or targets taken too early; cross-check with the metric interpretation in Backtest Result Analysis.
- Per-trade review: for losing or odd fills, use K-line playback to inspect entry / exit context, and look for commonalities (a certain time window, direction, instrument) to inform changes to conditions or execution config.
- Backtest vs runtime comparison: compare live / paper stats with the matching backtest analysis (win rate, profit factor, max drawdown); when the gap is large, suspect parameter or cost-assumption issues first.
Improvement loop: review statistics and records → diagnose → adjust strategy or execution config → re-validate via backtest → update live / paper. See Monitor, Execution Config.
Usage tips
- Whole-to-detail: start from the home overview, then drill down to the bot and individual pairs, and finally to per-trade fills and entry / exit points; avoid getting lost in individual trades before understanding the big picture.
- Regular review: review statistics daily / weekly / monthly to catch issues early; statistics monitor both live and paper.
- Compare with backtest: if live / paper win rate, profit factor, or frequency keep diverging from backtest, check whether the execution config is consistent and the slippage / cost assumptions are reasonable.
- Pair with alerts and takeover: when Auto Takeover is configured, use the stats and per-trade records to diagnose what happened after a takeover triggers, and decide whether to adjust the strategy or execution config.
Common questions
How do I filter records? In the system, filter by time range, pair, strategy, direction, P&L, etc. (per system support); filtering helps focus on a specific period or strategy.
How do I use the stats to improve the strategy? Start with the equity curve and drawdown at the home / bot level to spot problematic windows; drill down to the per-trade records and K-line playback in that window to inspect entry / exit context, then adjust the buy / sell conditions or execution config accordingly; validate the change via backtest before updating live—see Monitor.
Are paper statistics the same as live? The structure and fields match—only the capital is virtual and the fills are simulated; use them to review and validate the paper run, then compare with live stats to estimate the gap.
Can backtest stats be compared directly with live stats? The core metrics (win rate, profit factor, max drawdown, equity curve) are computed in similar ways, so they can be compared; but live is also affected by slippage, liquidity, and API latency, so some gap from backtest is expected—focus on whether the trend and order of magnitude are close.
Next steps
- Monitor — Performance review and strategy optimization
- Execution Config — Adjust position and risk
- Risk Management — Execution config and bot-level safety net
- Backtest Result Analysis — Metric interpretation and comparison
- Live Trading, Paper Trading — View stats while the bot is running
- Auto Takeover — Review after takeover, with stats and records