Trading bot glossary: key terms defined plainly
Algorithmic trading is dense with jargon, and a single unfamiliar term can derail an otherwise clear guide. This glossary defines the terms that matter most when building, testing and running a trading bot — from backtest and slippage to Sharpe ratio, funding rate and cointegration — in plain English, with links to the deeper guide for each. Use it as a reference: skim it once to map the landscape, then come back whenever a term trips you up. Every definition is written to stand on its own.
Core bot terms
Trading bot — a program that automatically places trades by running a trading algorithm against a live exchange.
Algorithm — the exact, testable set of rules that maps market data to buy/sell/hold decisions. See what is a trading algorithm.
Signal — the long/short/flat output of a strategy at a point in time; the trigger to act. See trading signals.
Strategy — the trading idea (e.g. trend following) that an algorithm encodes.
ccxt — the open-source library that connects bots to 100+ exchanges with one API. See how to use ccxt.
Strategy terms
Trend following — buying strength and riding sustained moves; few big winners, many small losses.
Mean reversion — buying weakness expecting a snap back to average. See mean reversion.
Breakout — entering when price clears a key level. See breakout bots.
Grid trading — layered buy/sell orders harvesting a range. See grid bots.
DCA — dollar-cost averaging: accumulating at intervals regardless of price. See DCA bots.
Statistical arbitrage — trading the mean-reverting spread between related assets. See stat arb.
Pairs trading — the two-asset, market-neutral form of stat arb. See pairs trading.
Risk and metrics
Drawdown — the peak-to-trough drop in account equity; the key risk metric. See drawdown.
Sharpe ratio — return per unit of volatility; risk-adjusted performance. See Sharpe ratio.
Risk-reward ratio — the size of a target versus its stop, which sets your break-even win rate. See risk-reward.
Position sizing — how much to risk per trade, usually ~1% of the account. See position sizing.
Win rate — the percentage of trades that are profitable; meaningless without the risk-reward ratio.
Risk of ruin — the probability an account falls so far it cannot recover.
Kill switch — automatic logic that halts the bot and flattens positions when a limit is breached.
Execution and market terms
Slippage — the gap between your expected and actual fill price. See slippage.
Market order — fills immediately at the best available price; certainty of fill, not price.
Limit order — fills only at your price or better; certainty of price, not fill. See order types.
Maker / taker — a maker adds liquidity (resting order), a taker removes it (market order); fees differ.
Market maker — quotes both sides to earn the spread. See market makers.
TWAP — slicing an order evenly over time to reduce impact. See TWAP.
VWAP — volume-weighted average price, an execution benchmark and indicator. See VWAP.
Iceberg order — a large order showing only a small visible slice. See iceberg orders.
Funding rate — the periodic payment between perpetual-futures longs and shorts. See futures bots.
Leverage — borrowing to size up; multiplies gains and losses. See leverage.
Liquidation — forced closure of a leveraged position when margin runs out.
Technical and data terms
OHLCV — open, high, low, close, volume: the candle data a bot trades on. See data sources.
Backtest — running a strategy over historical data to estimate performance. See how to backtest.
Forward test / paper trade — running on live data with fake money. See paper trading.
Overfitting — tuning so tightly to history that the strategy memorises noise. See overfitting.
Look-ahead bias — using future data the bot could not have known at decision time.
Walk-forward analysis — rolling optimize-then-test validation. See walk-forward.
Monte Carlo simulation — reshuffling trades to estimate the outcome range. See Monte Carlo.
Cointegration — a stable, mean-reverting relationship between two prices; the basis of pairs trading.
API key — the credential a bot uses to trade; should be trade-only, no withdrawals. See key security.
VPS — a virtual private server that hosts the bot 24/7. See hosting.
Frequently asked questions
What is OHLCV in trading bots?
OHLCV stands for open, high, low, close and volume — the five values that summarise price activity for a candle over a fixed period. It is the standard market-data format a trading bot consumes: each row is one candle, and indicators, signals and backtests are all computed from a series of OHLCV rows fetched from the exchange.
What is the difference between a maker and a taker?
A maker places a resting limit order that adds liquidity to the order book and waits to be filled, while a taker places an order that immediately removes liquidity by matching against existing orders, such as a market order. Exchanges usually charge takers a higher fee than makers, which is why some bot strategies are designed to earn the maker rebate.
What does slippage mean for a trading bot?
Slippage is the difference between the price a bot expected to trade at and the price it actually filled at. It happens because the market moves between decision and execution and because a market order eats through the order book at worsening prices. Slippage is a real cost that can quietly turn a backtest-profitable strategy into a live loss if not modelled.
What is overfitting in one sentence?
Overfitting is tuning a trading strategy so tightly to historical data that it memorises random noise instead of a repeatable pattern, which makes it look excellent in a backtest but fail in live trading. Avoiding it requires keeping strategies simple, testing on out-of-sample data and paper trading before risking real money.