The platform uses a small number of nouns consistently. Learn these nine and every screen becomes readable.
Strategy#
One market, one timeframe, one set of rules. Everything else hangs off it. A strategy holds indicators, signals, an optional bias, and risk management, and you edit each of those on its own tab.
Changing a strategy does not change past backtests — re-run the backtest to see the effect of an edit. Use Commit to snapshot a version you like before experimenting, so you can restore it.
Ticker, Data Feed and Timeframe#
The three things you choose when you create a strategy and rarely change after.
- Ticker — the market, written as
BASE/QUOTE:BTC/USDT,ETH/USDT. - Data Feed — where the price history comes from. Crypto Datafeed is sourced from Binance.
- Timeframe — the candle size your rules are evaluated on: 1m, 5m, 15m, 30m, 1h, 4h or 1d. One rule behaves completely differently on 1m and 1d.
Indicator#
A calculation your rules read — RSI, a moving average, Bollinger Bands, news
sentiment. Each indicator you add publishes one or more named lines, and
your rules compare those lines. A line is written IndicatorName.line, using
the name you gave the indicator when you added it.
Raw price is always available without adding anything: open, high, low,
close, volume.
Signal#
The rules that open and close positions, organised into named signal groups. Each rule has an Action (enter or exit position), a Direction (long or short), and one or more Conditions.
A condition compares a Line against either a Threshold line or a
Fixed value, using a Condition operator — goes above, goes below,
is greater than, is less than, equals, and so on. Multiple conditions in
a rule are joined with AND: all must be true.
Bias#
An optional directional state — long, short or neutral — produced by its own rules. Bias exists so you can say "only take longs while the bigger picture agrees" once, rather than writing that condition into every rule.
A bias does not filter anything on its own. You have to reference it as a condition in your signal rules for it to have any effect.
Risk management#
The stop loss and take profit attached to a strategy. The stop caps what a losing trade costs; the take profit banks a winner, optionally in stages. Set the stop before you size the position, not after.
Venue#
Where a strategy actually runs. A venue pairs a strategy with an account and a mode — paper (simulated capital on the platform's shared paper exchange) or live (real orders on an exchange you have connected).
The same strategy can have several venues at once. Running it on paper and live simultaneously is normal and often useful: the paper venue keeps generating an honest track record while the live one trades.
Each venue holds its own allocated capital, leverage and profit and loss.
Position#
One trade, from the order that opened it to the order that closed it. A position belongs to a venue — which is what makes it paper or live — and carries its size, entry and exit prices, fees, leverage and profit and loss.
Positions in the sidebar shows every position you own with fund-grade statistics across whatever you filter to.
Basket#
A group of strategies you allocate capital to and manage together, rather than one at a time. Useful when you want several uncorrelated strategies to share a pot.
Committee#
A team of AI analysts you create, which reviews trades and gives a verdict. A Squad argues each position from four angles — technical, fundamentals, news and market sentiment — with a bull and a bear on each side, then two managers decide. Squads are called committees in the API, the CLI, the MCP tools and in URLs — that naming is stable for automation.
A Squad is not tied to one strategy. It covers the markets you give it and publishes a decision for each on its own schedule; any strategy on those markets can read it. That is the point: you build a view you trust once and apply it consistently.
A Squad never trades. Every decision is recorded with the price at the time and graded against what the market did next, so it accumulates a score you can check before deciding to follow it.
Grid#
A way to test many variations of one strategy at once. You set ranges for the parameters you are unsure about and the platform backtests every combination, so you can look for a region of settings that all work rather than a single lucky one.