Strategy builder glossary

Timeframe

The timeframe is how long one candle lasts: 1m, 5m, 15m, 1h, 4h or 1d. The engine checks every condition once per candle, when that candle closes, and acts on the close. So a 4h strategy looks at the market six times a day, and a 1d strategy once.

Shorter timeframes trade more often, which makes fees and slippage weigh more. Indicators count in candles of the run’s timeframe: RSI(14) on 1h covers 14 hours, on 1d two weeks. A condition can also read an indicator on a longer timeframe than the run’s — “I trade 1h, but only while the price is above EMA(200) on 1d”.

Conditions: ALL, ANY, NONE

A condition row compares an indicator with a number (RSI(14) < 30) or with another indicator (EMA(9) crosses above EMA(21)). Rows are combined by the group’s switch:

• ALL (AND) — every row must hold on the same candle;
• ANY (OR) — one row is enough;
• NONE (NOT) — the rows inside must not hold.

Groups nest, which is how “A and (B or C)” is written: a group set to ALL with a row A and a nested group set to ANY holding B and C. “Crosses above” is true only on the candle where the line passes from below to above — one candle, not every candle while it stays above.

Bar window

By default a row is checked on the current candle only. A bar window stretches it over several candles:

• Held for N — the row is true on each of the last N bars. It filters out one-candle blips: “volume above its average for 3 bars in a row”.
• Within N — the row was true on at least one of the last N bars. It lets two events a few bars apart count together.

Entry layers and the entry mode

An entry layer is one entry rule: its conditions, a side (long or short — spot is long only) and a size. Most strategies have one layer. Several layers let one strategy enter in stages — a first buy on a dip and a second, larger one if the dip deepens — or trade both directions, a long layer and a short one.

Each layer that holds a position counts against Max open positions (see Position size), so two layers that should each hold a trade need it at 2 or more; the editor refuses to save otherwise.

The entry mode switches between writing layers yourself and a DCA grid, where one start rule opens a deal and safety orders do the rest (see DCA grid).

Setup → trigger

A two-step entry for ideas that wait for a situation and then for a moment. The setup is the situation: when it is true, the layer is armed for a number of bars. The trigger is the moment to act, and it counts only while the setup is armed. An invalidation rule can disarm the setup early.

The editor edits the trigger like ordinary entry conditions; the setup itself is written in the Config tab, and the layer shows a «setup → trigger» chip when it has one.

Position size and max open positions

The size says how much one entry uses. Five modes:

• % cap. — a share of the free capital (on futures, the margin);
• USDT — a fixed amount;
• % risk and $ risk — the size at which hitting the stop loses that share of capital or that many USDT; they need a stop-loss;
• vol. — a size set by volatility, so a calm market gets a larger position than a wild one.

Max open positions caps how many positions the strategy holds at once, across all layers. At the default of 1 a new signal is ignored while a trade is open.

Cooldown and re-arm

Cooldown makes a layer enter at most once every N bars, counted from its previous entry. It keeps a condition that stays true for a while from firing on candle after candle.

Re-arm says when a layer may enter again: ALL_LAYERS_CLOSED (default) — once every position of the strategy is closed; THIS_LAYER_CLOSED — once this layer’s own position is closed; ALWAYS — even while its earlier positions are open, up to Max open positions. With ALWAYS a step in percent can require the price to move against the open positions before the next entry.

Exit scope: per position or aggregate

The take-profit and stop-loss can be set in percent, in multiples of ATR, in multiples of the risk (R), at an indicator’s level, or in parts (a ladder). The scope decides what they are measured from:

• Per position — each position has its own levels, measured from its own entry price;
• Aggregate — all open positions on a side close together, with the levels measured from their average entry price. This is the usual choice when you add to a position, and the only one a DCA grid uses.

Exit signal

An exit signal closes the trade when a condition becomes true, whatever the profit or loss — for instance when the trend that justified the entry ends. It works alongside the take-profit and stop-loss: whichever comes first closes the trade.

A strategy needs at least one way out: a take-profit, a stop-loss or an exit signal. Trend-following strategies often have no take-profit at all — an exit signal and a stop.

Pyramiding

Pyramiding adds to a position that is already open, when the layer’s entry condition holds again — usually only while the position is in profit, so a winning trade grows. An add skips the re-arm rule and the position cap: it is the same position, larger. Each add can be smaller than the one before, and the stop can follow the new average price.

It is the opposite of a DCA grid, which adds to a position that is losing.

DCA grid

A DCA grid is an averaging-down bot. A deal starts with a base order when the start rule holds. If the price then moves against it, safety orders add at set steps — each step and each order can grow — which pulls the average entry price toward the market. The whole deal closes at a take-profit measured from that average, or at an optional stop.

A grid needs room: Max open positions must be at least 1 + the number of safety orders. And a grid without a stop can sit in a deep loss for a long time — the backtest shows how long and how deep.

Leverage, margin and liquidation

On futures you put up margin and the exchange lends the rest. Leverage is the ratio: at 5x, 10% of capital as margin opens a position worth 50% of it. Profit and loss grow by the same factor.

Margin mode. Isolated: a position can lose only the margin assigned to it. Cross: all free capital backs every position — liquidation is further away, but a bad trade can take more. When losses eat the margin the position is liquidated: closed at a loss of almost all its margin. The liquidation buffer moves the liquidation price this many percent closer to the entry, as a safety margin.

Funding

Perpetual futures have no expiry; instead, longs and shorts pay each other every few hours (usually every 8) a small rate that keeps the futures price near spot. When most traders are long, longs pay shorts, and the other way round. With funding on, the backtest charges and credits these payments as the exchange would.

Fees and slippage

Fees are the exchange’s commission, charged on the entry and again on the exit as a percentage of the trade. Slippage is the gap between the price a signal sees and the price an order really fills at; it is charged on every fill too.

Both are small per trade and large over hundreds of trades. A backtest with zero costs flatters any strategy that trades often; keep the defaults or set the rates of your exchange.

Frequently asked questions