Every definition below has two parts: what the term means, and — where needed — the warning sellers omit. Italic links point to the in-depth guides.
A–C
Ask — The price at which you can buy. Always higher than the bid: the difference is the spread, your first cost.
ATR (Average True Range) — A measure of recent average volatility. The correct unit for stops, targets and thresholds: an ATR-based system adapts to the market, a fixed-point one breaks when the market changes scale. Technical analysis.
Backtest — Simulation of a strategy on historical data. An upper bound of reality, never a promise: without realistic costs and out-of-sample data it's just formatted self-deception. Overfitting.
Basis — Difference between a future/perpetual's price and spot. The basis of carry strategies.
Bid — The price at which you can sell.
Breakout — A break of a price level. False breakouts are so frequent that "buying the break" without context had, in my BTC tests, a gross edge ≈ zero.
Bps (basis point) — One hundredth of a percentage point (0.01%). The unit for measuring edges and costs in quant work.
CFD (Contract for Difference) — A contract with a broker replicating an asset's price without owning it. Convenient and regulated in Europe; inefficient for the long term (daily swaps). Brokers.
Circuit breaker — A loss limit that shuts down the system (or the trader). With a "latch": once triggered it stays triggered. The difference between a limit and a suggestion.
Commission — Fixed cost to execute an order. Always add it to the spread: 93% of one of my bots' loss was commissions.
Cooldown — Mandatory pause between trades. Anti-overtrading and anti-revenge constraint. Psychology.
D–F
Drawdown — Distance between the capital's peak and the subsequent low. The most important number in any track record: it measures how close you came to ruin. A −50% requires +100% to break even. Risk management.
Edge — Positive expectancy with an identifiable cause: structural, behavioral, informational or risk premium. If you don't know who pays it and why they should continue, you don't have an edge: you have a lucky streak.
Equity — Account value including open positions (unlike balance, which excludes them). Risk limits are computed on equity: live risk sits in open positions.
Expectancy — Expected average profit per trade: (win rate × avg win) − (loss rate × avg loss) − costs. The only equation that matters. Calculator.
Expert Advisor (EA) — A program that trades automatically on MetaTrader. It executes rules: it doesn't create an edge the rules don't have. EA guide.
Funding rate — Periodic payment between longs and shorts in perpetual futures, keeping price anchored to spot. Crypto.
Futures — A standardized contract to buy/sell at a future date, traded on regulated venues.
G–L
Gap — A price jump between one close and the next open. Why a stop is a request, not a price guarantee.
Grid trading — A strategy opening positions at regular intervals against the move. Typical profile: many small wins, rare devastating losses. Showcase win rate, nightmare tail. Anatomy of a deception.
Headroom — The equity's distance from the invalidating limit (e.g. a prop floor). Professional sizing contracts as headroom shrinks.
Leverage — Ratio between position controlled and margin committed. Multiplies profits and losses identically: it accelerates variance, creates no edge. Forex.
Liquidation — Forced closure of a leveraged position when margin runs out. In crypto perpetuals, liquidation cascades amplify moves.
Long / Short — A position betting up / down. Markets aren't symmetric: a system can have edge in one direction only — my gold EA proved it with embarrassing precision.
Look-ahead bias — Using information in a backtest that wasn't available at decision time. The industry's most insidious error: marvelous backtests, live systems dead on day one.
Lot — Forex volume unit: standard 100,000 units, mini 10,000, micro 1,000.
Overbought / oversold — Extreme oscillator readings. They don't mean "about to reverse": in strong trends the market stays extended for weeks. Technical analysis.
M–P
MAE / MFE — Maximum adverse/favorable excursion of a trade: how far it went against/for you before closing. Precious data for calibrating stops and targets.
Magic number — The identifier an EA stamps on its own orders so it never touches anyone else's.
Margin — Capital locked as collateral for a leveraged position.
Market maker — An operator (or broker) quoting both bid and ask, earning the spread.
Martingale — Doubling the stake after every loss. Mathematically: it moves the probability of ruin from "possible" to "a matter of time". The hidden engine of half the "infallible" EAs.
MQL5 / MQL4 — The programming languages of MetaTrader 5 and 4. MQL5 done right.
Out-of-sample (OOS) — Data never used during development, reserved for the final verdict. Without OOS, every backtest is in-sample: memory, not prediction.
Overfitting — Mistaking memory for understanding: optimizing a system on the past's noise. The number one cause of death of retail systems. Full guide.
Overtrading — Trading too much: more certain costs on an uncertain edge. Psychology.
Pip — The conventional minimum change of an exchange rate (fourth decimal; second on yen pairs). Forex.
Position sizing — Computing how much to buy/sell given the chosen risk: (capital × risk %) / stop distance in value. The most important decision of every trade. Guide · FastAutolot.
Profit factor (PF) — Gross profits / gross losses. Below 1 = losing. Beware: a high PF on few trades means nothing, and a 1.2 PF can evaporate with slightly worse costs.
Prop firm — A company granting capital to those who pass a "challenge" with drawdown constraints. You buy an option, not capital. The math.
R–S
R (risk multiple) — P&L expressed in multiples of the trade's initial risk: +2R = you made twice what you risked. The unit that makes differently-sized trades comparable.
Regime — A persistent market state (trend/range, high/low volatility). Systems almost always die of regime change, not bad luck.
Requote — The broker rejects your price and offers another. Measure it in demo before choosing a broker.
RSI — A momentum oscillator (0–100). It measures the extension of recent movement; it doesn't predict reversal. Technical analysis.
Scalping — Trading on horizons of seconds/minutes. The territory where costs matter most: avoid until you can compute your signal-to-cost ratio.
Slippage — Difference between expected and executed price. Invisible in backtests, real on your account.
Spread — The bid/ask difference: the cost you pay at every entry. It widens in thin hours and on news — exactly when it hurts most.
Stop loss — The order closing a position at a predefined loss. Decided before entry, in volatility units, and never widened. Guide.
Survivorship bias — Seeing only the winners because the losers don't publish. True of gurus, of EAs for sale, and of your own memories.
Swap (rollover) — Cost/credit for holding a position overnight. Tripled once a week. Structural on multi-day systems.
T–Z
Take profit (TP) — The order closing a position at a predefined profit. It must belong to the physics of the traded phenomenon: a price TP on a time-convergence phenomenon is a category error. Real case.
Timeframe — The time unit of the bars (M1, M5, H1, D1…). The minimum sustainable timeframe is decided by your costs, not your impatience.
Track record — A verifiable history of results. The 5 checks: real/demo, avg W/L vs win rate, pip expectancy, Z-score, drawdown with its date. Tools.
Trailing stop — A stop that follows price into profit. Useful for management; it doesn't create an edge that isn't there: with zero drift, no dynamic management turns a losing system into a winner.
VWAP — Volume-weighted average price: where the money went, not just the price. The benchmark of institutional execution. Full guide · MT5 indicator.
Walk-forward — Optimizing on a rolling window and validating on the next, repeatedly. It tests the procedure, not a single parameter set.
Win rate — Percentage of winning trades. Alone it means NOTHING: it must be compared with the break-even win rate derived from the avg win/loss ratio. An 81% can be mathematically losing. Proof.
Z-score (of a track record) — A measure of dependence between consecutive outcomes. Strongly negative = wins in clusters = the signature of grids and martingales.