glossary.title

Alpha measures the excess return of a strategy above its benchmark after adjusting for market risk (beta). A positive alpha indicates the strategy generated value beyond what the overall market movement would explain. It is a core metric when evaluating the skill of a systematic or discretionary trader.

ATR quantifies market volatility by averaging the true range — the largest of the current high minus low, the absolute high minus prior close, and the absolute low minus prior close — over a lookback period. Traders use it to size positions, place stops, and assess whether a market is trending or consolidating. Higher ATR signals wider price swings and therefore more risk per trade.

Backtesting is the process of applying a trading strategy to historical market data to evaluate how it would have performed. It provides a statistical foundation for assessing viability before live deployment. Common pitfalls include look-ahead bias, survivorship bias, and overfitting to the test period.

A bear market is conventionally defined as a sustained decline of 20% or more from a recent peak in a broad index or asset price. Prolonged selling pressure, deteriorating fundamentals, or systemic deleveraging typically drive bear markets. Systematic strategies often require separate regime logic to handle reduced momentum and higher correlations during these periods.

Beta measures the sensitivity of an asset or strategy to movements in a benchmark index. A beta of 1 implies the strategy moves in lockstep with the market; a beta above 1 amplifies market moves; below 1 dampens them. Quant traders seek to decompose returns into beta (market exposure) and alpha (skill-based return).

Bollinger Bands envelope price with a middle simple moving average flanked by upper and lower bands set at a configurable number of standard deviations. When price touches or pierces the outer bands, it can indicate overextension, providing mean-reversion or breakout entry signals depending on the strategy context. Band width also functions as a volatility gauge.

A breakout strategy enters a position when price moves beyond a defined support or resistance level, on the premise that the breach signals the start of a new directional move. False breakouts are a significant challenge, often addressed with volume confirmation or volatility filters. Breakout systems tend to perform best in trending, news-driven markets.

A bull market describes a sustained period of rising prices, typically characterised by investor optimism, strong economic data, and trend-following strategies achieving high win rates. Defining when a bull market transitions to a bear market is itself a modelling challenge; many systems use moving average crossovers or drawdown thresholds as regime signals.

The Calmar Ratio divides the annualised compound return of a strategy by the absolute value of its maximum drawdown over the same period. It rewards strategies that generate strong returns without deep capital erosion, making it particularly relevant for managed futures and hedge fund evaluation. Higher values indicate a more favourable return-to-drawdown trade-off.

Consecutive losses (or the maximum losing streak) records the longest uninterrupted sequence of losing trades in a backtest or live record. Even high win-rate strategies can experience painful streaks due to clustering of correlated market conditions. Knowing the worst historical streak is essential for sizing positions so that a recurrence does not breach drawdown limits.

Correlation quantifies the linear co-movement between two time series on a scale from -1 (perfect inverse) to +1 (perfect positive). In portfolio construction, combining low- or negative-correlation strategies reduces aggregate volatility and drawdown. However, correlations can spike toward +1 during market stress, a phenomenon known as correlation breakdown.

Drawdown measures the decline in portfolio value from a peak to a subsequent trough, expressed as a percentage of the peak value. The maximum drawdown across all peak-to-trough intervals is the most commonly reported variant. It is used to stress-test whether a strategy's capital erosion is psychologically and contractually tolerable.

The equity curve plots the cumulative profit and loss of a strategy over time, providing an intuitive visual summary of performance quality. A smooth, upward-sloping curve with shallow drawdowns suggests consistent edge; a volatile or declining curve warrants investigation. Many traders also apply technical or statistical filters to the equity curve itself to pause a system during adverse regimes.

Expectancy quantifies the average dollar amount a strategy expects to gain or lose per unit of risk per trade, combining both win rate and the average win-to-loss ratio. A positive expectancy is a necessary condition for a viable strategy; the magnitude determines how much edge exists per trade. It is calculated as (Win Rate × Average Win) − (Loss Rate × Average Loss).

In-sample data is the historical period used to develop and optimise a strategy's parameters, while out-of-sample data is a separate, unseen period used to validate those parameters. Genuine predictive power should persist on out-of-sample data; performance that collapses outside the optimisation window is a strong signal of overfitting. Best practice allocates a significant fraction of history (often 30–40%) for out-of-sample testing.

Liquidity describes how quickly and cheaply an asset can be bought or sold without materially moving its price. High liquidity (tight bid-ask spreads, deep order books) reduces transaction costs and slippage, which is critical for high-frequency or large-scale strategies. Systematic traders must model liquidity constraints to produce realistic backtest results.

MACD is calculated as the difference between a fast and slow exponential moving average, with a signal line (EMA of MACD) overlaid for crossover signals. A histogram version plots the distance between the MACD and its signal line, helping visualise momentum acceleration or deceleration. It is one of the most widely used trend and momentum indicators in systematic strategies.

A market maker continuously quotes both buy and sell prices for a security, profiting from the bid-ask spread while providing liquidity to the market. In electronic markets, designated market makers compete with high-frequency proprietary traders who perform a similar function algorithmically. Systematic strategies that trade against market makers must account for spread costs as a direct drag on performance.

Maximum drawdown (MDD) is the single largest peak-to-trough percentage decline in portfolio value observed over a specified history. It is the most widely cited risk metric because it captures the worst-case capital loss an investor would have experienced. Strategies with low MDD relative to their returns command a higher Calmar Ratio and are generally more deployable in institutional portfolios.

Mean reversion strategies exploit the empirical tendency of asset prices or spreads to return toward a long-run equilibrium after deviating significantly. The strategy buys when a metric (price, z-score, RSI) falls below its historical mean and sells when it rises above it. Statistical confirmation of mean-reverting behaviour (via tests such as the Augmented Dickey-Fuller test) is a prerequisite for robust implementation.

Momentum strategies capitalise on the empirical finding that assets that have outperformed recently tend to continue doing so over intermediate horizons. The effect has been documented across equities, commodities, currencies, and fixed income, and is one of the most replicated factors in academic finance. Momentum strategies typically suffer during sharp market reversals and require careful drawdown management.

A moving average smooths price data over a rolling window to reduce noise and reveal the underlying trend direction. Simple moving averages weight all observations equally, while exponential moving averages assign greater weight to recent prices. Crossover signals (short MA crossing above or below long MA) are foundational triggers in many trend-following systems.

Overfitting occurs when a strategy's parameters are tuned so specifically to historical noise that the model loses its ability to generalise to new data. Signs include dramatically better in-sample than out-of-sample performance, or strategies with an excessive number of rules relative to the amount of training data. Regularisation techniques, cross-validation, and keeping parameter counts low are key defences.

Pairs trading identifies two historically co-integrated assets and trades their price spread: buying the underperformer and selling the outperformer when the spread deviates beyond a threshold, with the expectation of reversion. It is a market-neutral strategy that insulates exposure from broad market moves while remaining exposed to spread-specific risk. Cointegration must be monitored continuously as relationships can break down.

Profit Factor is the ratio of gross profit to gross loss across all trades in a period; a value above 1.0 means the strategy earns more than it loses in aggregate. A profit factor of 1.5 or higher is often cited as the minimum bar for a viable systematic strategy, though this depends on win rate and trade frequency. It complements the Sharpe Ratio by focusing on raw P&L rather than risk-adjusted return.

The Recovery Factor divides the net profit of a strategy by its maximum drawdown, indicating how many times the worst-case loss was recovered in returns. A recovery factor above 3 suggests the strategy earns meaningfully more than it has historically risked. It is a quick screening metric for evaluating risk-adjusted performance alongside the Calmar and Sharpe ratios.

The risk-reward ratio compares the expected profit on a trade to the expected loss if the stop is hit, expressed as a multiple (e.g. 1:2 means risking 1 to make 2). Higher ratios allow a strategy to be profitable even with a win rate below 50%. The ratio should be set based on the market's actual volatility and the historical distribution of favourable moves, not arbitrary multiples.

RSI is a momentum oscillator bounded between 0 and 100, measuring the magnitude of recent price gains versus losses over a configurable lookback period (typically 14 periods). Readings above 70 are conventionally overbought; readings below 30 are oversold — though in strong trends these thresholds often fail. Many systematic strategies use RSI divergence or extreme readings in conjunction with trend filters rather than as standalone signals.

The Sharpe Ratio divides the excess return of a strategy above the risk-free rate by the standard deviation of those returns, providing a standardised measure of risk-adjusted performance. A Sharpe above 1.0 is considered acceptable, above 2.0 is strong, and above 3.0 is exceptional for a live trading strategy. It penalises both upside and downside volatility equally, which is why the Sortino Ratio is often preferred for skewed return distributions.

The Sortino Ratio modifies the Sharpe Ratio by using downside deviation (volatility of negative returns only) in the denominator rather than total standard deviation. This makes it a more appropriate risk-adjusted metric for strategies with positive skew, where high upside volatility would otherwise unfairly penalise performance. A Sortino above 2.0 is generally considered strong.

The spread is the difference between the best available ask (buy) price and the best available bid (sell) price for an asset at any given moment. It represents an immediate cost that any market-order trader incurs on entry and exit. Tight spreads reduce transaction costs and are essential for short-term or high-frequency strategies where spread costs can dwarf other expenses.

Statistical arbitrage (stat arb) exploits pricing inefficiencies between related instruments using quantitative models, typically relying on mean reversion of a constructed spread or factor portfolio. Unlike pure arbitrage, stat arb carries model risk and is not risk-free; the spread may widen before reverting, creating substantial drawdowns. It is commonly executed as a high-turnover, market-neutral portfolio of dozens to thousands of simultaneous positions.

Trend-following strategies take long positions in rising markets and short positions in falling markets, profiting from sustained directional price moves across asset classes. The strategy accepts many small losses (during choppy, ranging markets) in exchange for large, skewed winning trades when trends extend. Managed futures programmes are the canonical institutional implementation of systematic trend following.

Volatility measures the dispersion of returns for an asset or portfolio over a given period, most commonly expressed as annualised standard deviation. Realised volatility is calculated from historical price data; implied volatility is derived from options prices and represents the market's forward-looking expectation. Volatility is both a risk measure and a tradeable asset class (via VIX products and variance swaps).

Walk-forward analysis is a rolling validation methodology that repeatedly re-optimises strategy parameters on a moving in-sample window, then evaluates performance on the immediately following out-of-sample window. Stringing together many consecutive out-of-sample windows produces a realistic simulation of live trading performance. Strategies that degrade significantly across walk-forward steps are likely overfit to historical data.

Win rate is the percentage of trades that close at a profit, and is one of the most intuitive but potentially misleading metrics in strategy evaluation. A strategy with a 40% win rate can still be highly profitable if its average winner is substantially larger than its average loser. Win rate must always be evaluated alongside the average win/loss ratio and overall expectancy to draw meaningful conclusions.