Glossary

AI Trading Glossary

The AI, trading, and market terms that actually matter in 2026 \u2014 defined in plain English, with deeper-dive links where it helps.

AI

Pattern Recognition
A machine learning approach that identifies recurring formations in price or volume data. Quanta’s Time Machine uses pattern retrieval to compare a current setup with historical analogues and summarize what followed them. Learn more
Dynamic Time Warping (DTW)
A similarity algorithm that compares two time-series sequences while allowing for non-linear stretching. Quanta uses it within a versioned historical analogue index so patterns with different pacing can still be compared.
Calibration
A measure of how well a model’s predicted probabilities match observed outcomes. A well-calibrated model’s probability bands should align with their observed frequencies over a sufficiently large sample. Tracked via Expected Calibration Error (ECE). Learn more
Expected Calibration Error (ECE)
The weighted average gap between predicted probabilities and observed outcome rates. It is meaningful only when evaluated on a sufficiently large sample that was not used to fit the model.
Walk-Forward Validation
A validation method that trains a model on a rolling historical window and tests on the immediately following period, repeatedly. It can reduce in-sample bias, although it does not eliminate overfitting or market-regime risk. Learn more
Regime Detection
Classification of the current market environment, such as trending, mean-reverting, high-volatility, or low-volatility. Regime labels can provide context for interpreting a model output, but are uncertain and can change quickly. Learn more
Feature Vector
A numerical representation of an asset or moment in time, built from inputs such as returns, momentum, volatility, and volume. Similarity models compare these representations alongside price sequences.
AI Copilot
A Gemini-powered conversational research assistant that can explain Quanta outputs and help users explore stock and crypto context. It is a research interface, not a source of personalised trade recommendations. Learn more

Trading

Sharpe Ratio
Risk-adjusted return: average excess return divided by standard deviation of returns. It is a portfolio or strategy statistic, not a measure of directional signal accuracy. Historical simulation metrics depend on their assumptions and do not represent live capital. Learn more
Sortino Ratio
Like the Sharpe ratio, but it penalizes downside deviation rather than total volatility. It can be useful when upside and downside return distributions are asymmetric.
Drawdown
The percentage decline from a peak to a trough in portfolio value. Max drawdown is the worst peak-to-trough drop in the strategy’s history.
Position Sizing
The process of deciding how much capital to allocate to a position. Common inputs include risk tolerance, stop distance, volatility, liquidity, and exposure to correlated holdings. Learn more
Portfolio Heat
Total capital at risk across open positions, sometimes adjusted for correlation. Traders may use a portfolio-heat limit as a personal risk constraint; Quanta does not enforce one automatically.
ATR (Average True Range)
A measure of price volatility over a lookback window. Used to set stop distances that adapt to the current vol regime, instead of arbitrary fixed percentages.
Backtest
Simulating a trading strategy against historical data to estimate forward performance. Reliable backtests use walk-forward validation, realistic slippage, and survivor-bias-free data. Learn more
Slippage
The difference between an expected execution price and the actual fill price. Its impact varies with liquidity, order size, volatility, order type, and trading venue, so simulations should state their assumption explicitly.

Patterns

Cup and Handle
A bullish continuation pattern resembling a teacup: a rounded bottom (cup) followed by a slight downward consolidation (handle), then a breakout. AI ranks cup-and-handle setups by depth, duration, and volume profile.
Bull Flag
A short consolidation against the prior trend, forming a flag shape. Traders often interpret it as a possible continuation setup, but the pattern alone does not establish a probability of follow-through.
Head and Shoulders
A reversal pattern with three peaks: a higher middle (head) flanked by two lower peaks (shoulders). Inverse pattern signals a bottoming process.
Breakout
Price moving above a resistance area or below a support area, sometimes accompanied by expanded volume. Historical analogues can add context about how superficially similar breaks behaved afterward.
Mean Reversion
A class of strategies based on price or spreads moving back toward a reference level after a deviation. Results depend heavily on the chosen horizon, regime, costs, and whether the reference remains relevant.

Market Structure

Implied Volatility (IV)
The market’s forward-looking estimate of how much a stock will move, extracted from options prices. AI tools compare IV to historical realized volatility to find mispriced setups. Learn more
Realized Volatility
The observed variability of returns over a historical window. Comparing realized and implied volatility can provide context for options research, but does not establish that an option is mispriced.
Unusual Options Activity (UOA)
Options activity that is large or unusual relative to a contract’s normal volume or open interest. It may reflect speculation, hedging, market-making, or multi-leg activity, so intent cannot be inferred from size alone.
Order Flow
The sequence and imbalance of buy and sell interest reaching a market. Interpretation is difficult because visible activity can reflect hedging, market-making, execution algorithms, or directional positioning.
Market Breadth
A measure of how many securities participate in a market move, such as the share above a moving average. Breadth provides context but does not by itself determine whether a move will persist.

Crypto

On-Chain Data
Blockchain-derived metrics such as exchange flows, stablecoin supply, miner reserves, and large-holder concentration. These inputs can add context, but Quanta’s current public crypto workflow is based on market-data patterns rather than an on-chain signal product. Learn more
Funding Rate
The periodic payment between long and short holders of perpetual crypto futures, designed to anchor the contract to spot. Extreme funding rates flag positioning extremes.
Basis
The price difference between a futures contract and its underlying spot asset. Basis varies with funding conditions, demand for leverage, time to expiry, market access, and execution costs.
Liquidation
Forced closure of a leveraged position after margin requirements are breached. Liquidation estimates may provide market context, but venue coverage and assumptions can differ materially.

Quanta

Time Machine
Quanta’s pattern-retrieval workflow that compares a current setup with indexed historical analogues and shows the attached outcome distribution. Learn more
Proof Fund
A historical paper-portfolio simulation used to explore how a defined set of rules might have behaved. It is hypothetical, does not use live capital, and is separate from Quanta’s forward signal-results ledger. Learn more
Signal Results Ledger
Quanta’s public record of matured high-model-score directional signal observations. Read the current win rate with its sample size, confidence interval, as-of date, data-quality note, and methodology; it is not a portfolio return. Learn more
Daily Signals
A scheduled feed of ranked stock and crypto research setups. Each available row includes its direction, model score, context, and access state; availability depends on plan limits. Learn more

Ready to put these to work?

Quanta AI applies these concepts across its supported stock and crypto universe \u2014 free to start, no credit card required.

Start free Full learning hubCompare platforms