Terms, in plain English

ACM (Adrian–Crump–Moench)

The NY Fed's model that splits a yield into expectations and term premium.

A term-structure model from the Federal Reserve Bank of New York that decomposes a Treasury yield into an expected-rate (risk-neutral) component and a term premium. Widely used because the NY Fed publishes it free. Worth remembering that it is an estimate, not a measurement — its output carries its own model error.

See also: Term premium · Risk-neutral (expected-rate) component

Basis point (bp)

One hundredth of a percentage point.

One hundredth of a percentage point: 100 bp = 1%. Rates people use it because saying "rates rose 0.25%" is ambiguous — a quarter of a percentage point, or a quarter of one percent of the current rate? "25 bp" can only mean one thing.

Constant maturity (GS2, GS10)

A yield series held at a fixed maturity, so it's comparable over time.

An individual bond's remaining life shrinks every day, so tracking one bond doesn't give a clean time series. A constant-maturity series interpolates the curve to a fixed point — 2 years, 10 years — so today's 10-year yield is comparable to last year's. FRED publishes these as GS2, GS10, and so on.

Expectations hypothesis

A long rate ≈ the average short rate people expect, plus a premium.

The idea that a longer-term interest rate is roughly the average of the short-term rates the market expects over that period, plus a term premium for bearing the risk. It's why a 2-year Treasury is often read as the market's forecast of the average policy rate over two years — and why the Fed's grip is tight at the short end and loose at the long end.

See also: Term premium · Federal funds rate

Federal funds rate

The overnight rate the Fed targets — its main policy lever.

The interest rate banks charge each other for overnight loans of reserves. The Fed sets a target range for it, and it is the lever nearly everything people mean by "the Fed raised rates" refers to. It is an overnight rate — which is precisely why its grip on a 30-year loan is a question rather than an assumption.

See also: Expectations hypothesis

FOMC (Federal Open Market Committee)

The Fed committee that sets the federal funds rate target, meeting roughly eight times a year.

The Federal Open Market Committee — the body within the U.S. Federal Reserve that sets monetary policy. It meets about eight times a year to choose the target range for the federal funds rate and to decide on other tools such as the size of the Fed's bond holdings. When people say "the Fed raised rates," it is an FOMC decision they are describing.

See also: Federal funds rate · Zero lower bound (ZLB)

Greenspan's conundrum

2004–06: the Fed hiked 396 bp and long rates barely moved.

Between June 2004 and June 2006 the FOMC raised the federal funds rate by 396 basis points and long-term rates barely responded — the 30-year mortgage moved 39 bp. Greenspan called it a conundrum in 2005. It is the cleanest historical demonstration that the short end and the long end are not the same lever.

See also: Federal funds rate · Term premium

MBS spread

The gap between mortgage rates and Treasuries — its own moving part.

Mortgages are bundled into mortgage-backed securities, which yield more than Treasuries. That gap compensates for prepayment risk (borrowers refinance when it suits them, not you), plus bank funding conditions and origination economics. It moves for reasons of its own, which is part of why a mortgage rate isn't the policy rate plus a constant.

See also: Prepayment risk

Prepayment risk

Borrowers refinance when rates fall — so you get your money back at the worst time.

A US mortgage borrower can repay early, and does so exactly when rates drop and reinvesting is least attractive. That optionality is the borrower's and the lender bears it, so mortgage yields carry compensation for it — a component that has nothing to do with the Fed's target rate.

See also: MBS spread

Risk-neutral (expected-rate) component

The part of a yield reflecting expected future short rates, with the term premium stripped out.

In a term-structure decomposition, the portion of a bond's yield that corresponds to the average short-term interest rate the market expects over the bond's life — what the yield would be if investors demanded no extra compensation for risk. The remainder is the term premium. It is called "risk-neutral" because it prices the bond as if investors were indifferent to risk; like the term premium, it is a model estimate rather than an observed number.

See also: Term premium · ACM (Adrian–Crump–Moench) · Expectations hypothesis

Term premium

The extra yield demanded for locking money up longer.

The part of a long-term yield that isn't explained by expected future short rates — the compensation investors want for the risk of holding a bond for years rather than rolling overnight. It is not directly observable, so it has to be estimated by a model, which is why two people can disagree about its level while looking at the same yield.

See also: ACM (Adrian–Crump–Moench) · Expectations hypothesis

Yield curve

Yields plotted against maturity — the rate to lend for 3 months, 2 years, 10 years, and so on.

A plot of the yield on bonds of the same credit quality against their time to maturity — for Treasuries, the rate to lend the government money for three months, two years, ten years, thirty years. Its usual upward slope reflects the extra yield demanded for locking money up longer; it can flatten or invert when short rates sit high relative to long ones. Where a rate falls on the curve governs how much the central bank's overnight rate reaches it.

See also: Term premium · Expectations hypothesis · Constant maturity (GS2, GS10)

Zero lower bound (ZLB)

When the policy rate is pinned near zero and can't fall further.

The period when a central bank's policy rate sits at or near zero and conventional cuts are unavailable — in the US, roughly 2009–2015 and 2020–2021. It distorts any statistical relationship involving the policy rate, because a variable that cannot move cannot explain anything that does.

See also: Federal funds rate

ADF test (Augmented Dickey–Fuller)

The standard test for whether a series wanders or reverts to a mean.

A hypothesis test for non-stationarity. A low p-value says the series reverts to a stable level; a high one says you can't rule out that it wanders. It's the routine first check before regressing two time series on each other — and skipping it is how spurious regressions get published.

See also: Stationary / non-stationary · Unit root · Spurious regression

Autoregressive (AR)

A model that predicts a series from its own recent past values.

Describes a model in which a variable is explained by its own earlier values — this month's level regressed on last month's, and so on. Financial and macro series are often strongly autoregressive, which is both useful (recent history forecasts the near future) and hazardous (it produces the serial correlation that inflates apparent significance). An autoregressive model of a rate alone is a common benchmark to beat.

See also: Serial correlation (autocorrelation) · No-change benchmark (random walk)

Chow test

Tests whether a relationship broke at a specific date.

A test for whether a regression's coefficients are the same before and after a chosen break point. Useful for asking whether a relationship survived 2008, or the zero-lower-bound years, or 2020 — and if the answer is no, a full-sample estimate is an average of regimes rather than a description of any of them.

Cointegration

The test for whether two wandering series are genuinely tied together long-run.

Two series can each wander indefinitely and still be bound together, so the gap between them stays stable — like a dog and its owner on a lead. That's cointegration, and it's what separates a real long-run relationship from a spurious one. If two series are cointegrated, a regression of one on the other means something; if they aren't, it usually doesn't.

See also: Spurious regression · Stationary / non-stationary · Error-correction model (ECM)

Directional accuracy

How often a forecast gets the direction of the move right — 50% is a coin flip.

The share of forecasts that correctly call the sign of the next change — up versus down — regardless of size. A value of 50% is what a coin flip would achieve. It is a different and weaker claim than a positive out-of-sample R²: a model can get the direction right more often than not while still forecasting the magnitude poorly, so the two should not be conflated.

See also: Out-of-sample R² · No-change benchmark (random walk)

Durbin–Watson statistic

A quick read on serial correlation. ~2 is healthy; near 0 is a red flag.

A one-number check for serial correlation in a model's residuals. Around 2 means the errors look independent; near 0 means each error closely tracks the last one, and the reported precision of the model can't be believed as printed.

See also: Serial correlation (autocorrelation) · HAC / Newey–West standard errors

Error-correction model (ECM)

A model of two cointegrated series that pulls them back toward their long-run relationship.

A specification used when two series are cointegrated. It models short-run changes while including a term for how far the series have drifted from their long-run relationship, so each period part of that gap is "corrected." The speed of correction is itself an estimated coefficient. An error-correction model is only valid when cointegration holds; without it, the framework does not apply.

See also: Cointegration · Stationary / non-stationary

F-test (F-statistic)

Tests whether a set of variables jointly explains anything — a regression's overall significance.

A test of whether several coefficients are jointly zero, most often used for a regression's overall significance: does the model as a whole explain more than nothing? The F-statistic is the ratio of explained to unexplained variation, adjusted for degrees of freedom, and a large value with a small p-value rejects the "explains nothing" null. Like the t-statistic, it relies on the error assumptions holding.

See also: t-statistic · p-value · R² (R-squared)

HAC / Newey–West standard errors

A correction that fixes overstated precision without changing the estimate.

Heteroskedasticity- and autocorrelation-consistent standard errors — a correction for when errors are neither independent nor equally sized. It leaves the estimate alone and widens the uncertainty around it, which is often enough to cut a headline t-statistic in half. Named for Whitney Newey and Kenneth West.

See also: Serial correlation (autocorrelation) · Heteroskedasticity · t-statistic

Heteroskedasticity

Errors that are bigger in some periods than others.

When the size of a model's errors varies with conditions — small in calm periods, large in volatile ones. Common in anything financial. Like serial correlation, it doesn't bias the estimate but it breaks the standard errors, so significance tests can't be taken at face value.

See also: HAC / Newey–West standard errors

Mean reversion

The tendency of a series to drift back toward its long-run average after moving away.

The tendency of a variable that has moved far from its historical average to move back toward it over time. A mean-reverting series is stationary — it has a level to return to — whereas one that wanders indefinitely is not. The concept underlies both statistical tests of stationarity and the general observation that measures far from their historical norms have tended to move back toward them, though it says nothing about when.

See also: Stationary / non-stationary · Shiller CAPE (CAPE ratio)

No-change benchmark (random walk)

Predict that tomorrow equals today. Embarrassingly hard to beat.

The most naive forecast available: next period's value equals this period's. It sounds trivial and is notoriously difficult to beat for financial series. Any forecasting model that can't outperform it has not demonstrated anything, however good its in-sample fit looks.

See also: Out-of-sample R²

Ordinary least squares (OLS)

The standard way to fit a line: pick the slope that makes the squared errors as small as possible.

The most common method for fitting a regression line. It chooses the intercept and slope that minimise the sum of the squared vertical distances between the data points and the line. OLS is the workhorse of applied statistics, but its headline output — the coefficients and their significance — can only be read as printed when its assumptions hold, which is what the diagnostic tests around it check.

See also: R² (R-squared) · t-statistic · Serial correlation (autocorrelation)

Out-of-sample R²

Does the model beat a naive benchmark on data it never saw? Can be negative.

R² computed on data the model wasn't fitted to, measured against a benchmark. Unlike ordinary R² it can be negative — meaning you'd have done better with the naive guess. It's the honest test, because fitting history well is easy and predicting is not.

See also: No-change benchmark (random walk) · R² (R-squared)

p-value

The chance of a result this strong if there were really no effect. Below 0.05 is 'significant'.

The probability of seeing a result at least as extreme as the one in hand if the null hypothesis — usually "no relationship" — were true. A small p-value, conventionally below 0.05, is taken as evidence against the null. It is not the probability that the hypothesis is true, and it inherits the model's assumptions, so a p-value computed on autocorrelated data can look far more decisive than it is.

See also: t-statistic · F-test (F-statistic) · Serial correlation (autocorrelation)

R² (R-squared)

Share of a variable's movement the model accounts for. Easy to overread.

The fraction of the variation in the thing you're explaining that your model accounts for, from 0 to 1. A high R² feels like proof and often isn't: two series that both trend upward over 25 years will produce a high R² whether or not they have anything to do with each other. It measures fit, not truth.

See also: Spurious regression · Out-of-sample R²

RMSE (root mean squared error)

A forecast's typical miss, in the units of the thing predicted. Smaller is better.

The square root of the average squared forecast error — a single number for how far a model's predictions typically fall from what happened, in the same units as the variable. Squaring penalises large misses more than small ones. It is only meaningful in comparison: a model's RMSE is judged against a benchmark such as the no-change forecast, not read in isolation.

See also: Out-of-sample R² · No-change benchmark (random walk)

Serial correlation (autocorrelation)

Today's error resembles yesterday's — so you have less information than you think.

When a model's errors are correlated across time rather than independent. It doesn't bias the estimate itself, but it means your 300 monthly observations carry far less independent information than 300 truly separate ones — so the standard errors come out too small and everything looks more certain than it is.

See also: Durbin–Watson statistic · HAC / Newey–West standard errors

Spurious regression

Two drifting series look related because both drift, not because they're linked.

Granger and Newbold's 1974 result: regress one wandering series on another and you reliably get a high R² and a huge t-statistic even when the two are entirely unrelated. The regression is measuring shared trend, not a relationship. It is the single most common way a time-series result can be technically correct and completely meaningless.

See also: Cointegration · Stationary / non-stationary · R² (R-squared)

Stationary / non-stationary

Whether a series returns to a stable average or wanders indefinitely.

A stationary series has a stable mean and variance — knock it away and it comes back. A non-stationary one wanders with no particular level to return to. The distinction matters because most standard regression results assume stationarity, and applying them to wandering series produces confident-looking nonsense.

See also: ADF test (Augmented Dickey–Fuller) · Unit root · Spurious regression

t-statistic

How many standard errors an estimate sits from zero. Above ~2 is 'significant'.

An estimate divided by its standard error — roughly, how confident you can be that a relationship isn't zero. Above about 2 is conventionally "significant." It depends entirely on the standard error being right, so a t-statistic computed on autocorrelated data can be wildly overstated while looking authoritative.

See also: HAC / Newey–West standard errors · Serial correlation (autocorrelation)

Unit root

The statistical fingerprint of a series that wanders without returning to a fixed level.

A property that makes a time series non-stationary — informally, the series has no fixed level to return to and its shocks are permanent rather than fading away. Testing whether a unit root can be rejected is the standard way to decide whether a series is stationary before regressing it on another; if two series each carry a unit root and are not cointegrated, a regression between them is likely spurious.

See also: Stationary / non-stationary · ADF test (Augmented Dickey–Fuller) · Cointegration

Price/earnings ratio (P/E)

A stock's price divided by its earnings per share — the basic yardstick of how richly it's valued.

A company's share price divided by its earnings per share, or a whole market's price divided by its aggregate earnings. It expresses how many dollars an investor pays for each dollar of annual profit, and is the most common single gauge of how cheap or expensive a stock is. Because it uses one period's earnings, it can be distorted when profits are temporarily high or low — the problem the cyclically-adjusted version is built to address.

See also: Shiller CAPE (CAPE ratio)

Share buyback (repurchase)

A company buying back its own shares, which lifts per-share figures by shrinking the share count.

When a company uses cash to repurchase its own outstanding shares, reducing the number in circulation. Because earnings are then divided over fewer shares, buybacks raise earnings per share even if total profit is unchanged, and they return cash to shareholders as an alternative to dividends. Their growing use is one reason analysts debate whether historical valuation yardsticks are directly comparable to today's.

See also: Price/earnings ratio (P/E)

Shiller CAPE (CAPE ratio)

Price over ten years of inflation-adjusted earnings — valuation smoothed across a full cycle.

The cyclically-adjusted price-to-earnings ratio: a market's price divided by the average of the past ten years of company earnings, each adjusted for inflation. Averaging a decade of earnings smooths out the boom-and-bust swings that make an ordinary P/E lurch, so the CAPE is used to compare valuation across long stretches of history. It was popularised by economist Robert Shiller. Research has linked elevated readings to weaker average returns over long horizons, though it is a poor guide to the next month or quarter.

See also: Price/earnings ratio (P/E) · Mean reversion

Beta

How much an asset tends to move per 1% move in the market.

The slope from regressing an asset's returns on the market's: a beta of 1.2 means the asset has tended to move 1.2% for each 1% market move. It measures exposure to systematic risk, not quality — a low-beta stock can still go to zero on a company-specific failure. Betas are sample estimates: the number shifts with the period, the return frequency, and which index stands in for "the market."

See also: Systematic risk · Capital Asset Pricing Model (CAPM)

Capital Asset Pricing Model (CAPM)

The model tying an asset's expected return to its beta against the market portfolio.

Sharpe's equilibrium model in which an asset's expected excess return equals its beta times the market's expected excess return — only systematic risk is priced, because idiosyncratic risk can be diversified away for free. It anchors the vocabulary of beta and alpha, but it rests on strong assumptions, its market portfolio is unobservable, and decades of factor evidence show average returns varying in ways one market beta does not explain. A foundation to reason from, not a settled description of returns.

See also: Beta · Market portfolio · Factor investing

Correlation

How closely two assets move together, from −1 to +1. The engine of diversification.

A standardized measure of how two return series move together: +1 is lockstep, 0 is no linear relationship, −1 is mirror image. Diversification's benefit comes from correlations below +1 — the lower, the better the offset. Correlations are estimated from history, drift over time, and have a habit of rising toward +1 in a crisis, which is exactly when the offset is wanted most.

See also: Covariance · Diversification · Serial correlation (autocorrelation)

Covariance

Correlation with the units left in — the raw input portfolio risk math actually uses.

How two assets' returns move together in raw units, before standardizing into correlation. Portfolio variance is assembled from the covariances of every pair of holdings, which is why adding a low-covariance asset can lower total risk even when that asset is volatile on its own. Estimated covariances carry estimation error, and a large portfolio needs a great many of them — small errors compound when an optimizer takes the numbers literally.

See also: Correlation · Modern Portfolio Theory (MPT)

Diversification

Spreading capital across holdings so no single company's fate decides the outcome.

Holding many imperfectly correlated positions so that company-specific misfortunes partly offset. Diversification can shrink idiosyncratic risk — the part specific to individual companies — toward zero as holdings multiply, but it cannot remove systematic risk: when the whole market falls, a diversified portfolio falls with it. It reduces dependence on any single judgment; it is not a shield against market-wide losses.

See also: Idiosyncratic risk · Systematic risk · Correlation

Efficient frontier

The set of portfolios offering the most expected return for each level of risk.

In mean-variance analysis, the curve of portfolios that cannot be improved on — no more expected return without more volatility, no less volatility without giving up return. Portfolios below the curve are dominated by ones on it. The frontier is a construct of its inputs: computed from historical estimates it looks precise, but it moves when the estimates do, so the real frontier is known only approximately and only in hindsight.

See also: Modern Portfolio Theory (MPT) · Market portfolio

Idiosyncratic risk

Company-specific risk — the part diversification can actually eliminate.

Risk specific to one company or a narrow group — a failed product, a fraud, a lost contract — largely unrelated to the broad market. Because these shocks partly cancel across many holdings, idiosyncratic risk falls quickly as a portfolio broadens, and theory treats it as uncompensated: bearing it is a choice, not a necessity. Concentrated portfolios are dominated by it; broad indexes hold almost none.

See also: Systematic risk · Diversification

Market portfolio

Theory's portfolio of all investable assets, weighted by value. No real fund holds it.

In classical asset pricing, the value-weighted portfolio of every investable asset — all stocks, bonds, real estate, private businesses, even human capital. The CAPM's statements are made against this portfolio, and it is unobservable in practice: broad public-equity indexes such as a total-market fund or the S&P 500 are useful proxies, not the thing itself. Roll's critique follows — because the true market portfolio cannot be observed, the CAPM cannot be cleanly tested against it.

See also: Capital Asset Pricing Model (CAPM) · Efficient frontier

Modern Portfolio Theory (MPT)

Markowitz's framework: judge every holding by what it does to the whole portfolio.

Harry Markowitz's 1952 framework for choosing portfolios by expected return and risk together, where risk depends not just on each holding's volatility but on how holdings move with each other. Its core insight is that the portfolio, not the individual security, is the unit of decision. It supports diversification and covariance-aware construction; it does not prove that any particular index is optimal, and its outputs are only as good as the return, volatility, and correlation estimates fed in — which come from history and do not sit still.

See also: Diversification · Efficient frontier · Covariance

Systematic risk

Market-wide risk that diversification cannot remove — the risk that gets paid.

The component of risk tied to the whole market — recessions, rate shocks, crises — that hits nearly all equities at once, so adding more stocks does not diversify it away. In asset-pricing theory it is the risk investors are compensated for bearing, since idiosyncratic risk can be shed for free by diversifying. A broad index fund holds essentially pure systematic risk — which is why indexing reduces single-company risk yet fully participates in market-wide drawdowns.

See also: Idiosyncratic risk · Beta · Diversification

Efficient Market Hypothesis (EMH)

Prices reflect available information because investors compete to act on it; hard to beat doesn't mean always right.

Fama's proposition that market prices reflect available information quickly enough that no one should expect to earn persistent abnormal returns from that same information. It does not assume investors are omniscient or rational: efficiency is an outcome of competition, not an assumption about people. Analysts and traders seek information, form differing expectations, trade on them, and prices adjust — and visible mispricing attracts further research and capital, a process that is partly self-correcting rather than perfectly so. Prices can still be wrong, sometimes for years, because information is costly, beliefs differ, arbitrage is limited, and bearing risk is not free. Grossman and Stiglitz sharpen the point: prices could never be perfectly informative, because then no one would be paid to produce the research that makes them informative.

See also: Price discovery · Shiller CAPE (CAPE ratio)

Factor investing

Deliberately tilting a portfolio toward traits linked to differences in average returns.

Constructing portfolios to target characteristics that research has linked to differences in average returns — value, size, momentum, profitability, investment, low volatility. Distinct from broad indexing: a cap-weighted index contains companies displaying every characteristic, which is not the same as deliberately targeting one. The factor record is genuinely contested — premiums shrink after publication, replication is uneven, costs and crowding eat into implementation, and every factor has endured long stretches of underperformance.

See also: Capital Asset Pricing Model (CAPM) · Market-capitalization weighting

Float-adjusted market capitalization

Market value counting only the shares the public can actually trade.

A company's share price times only its freely tradable shares — excluding insider stakes, strategic holders, and other locked-up blocks. Major index providers weight by float rather than total market cap so that index weights reflect what investors can actually buy, and so index-tracking money is not chasing shares that never trade. Adjustment rules differ by provider, which is one reason the "same" index concept can carry slightly different weights in different hands.

See also: Market-capitalization weighting · Index reconstitution

Index reconstitution

How index membership is re-formed under published rules — scheduled in some families, as-needed in others.

The re-forming of an index's membership under its methodology — eligibility screens, size and liquidity thresholds, float updates, and, for some indexes, committee judgment. Some index families reconstitute on a fixed calendar; others, including the S&P U.S. indices, make changes as needed with no scheduled reconstitution date. Companies leave through acquisition, spin-off, or restructuring as well as decline, and additions are not endorsements: reconstitution is maintenance of a rule set, not a verdict on business quality. It is also where index funds must trade, which makes the rules themselves economically consequential.

See also: Market-capitalization weighting · Float-adjusted market capitalization

Market-capitalization weighting

Weighting holdings by company value, so prices — not a manager — set the portfolio.

Weighting each holding in proportion to its market value, usually float-adjusted. Weights then track prices on their own: winners grow into larger positions and decliners shrink without a discretionary rebalance, which keeps turnover low and lets the portfolio mirror the market's aggregate judgment. The same mechanism concentrates the portfolio in whatever has already risen — it allows winners to run, but it is not a formal momentum strategy, and it is not valuation-neutral.

See also: Float-adjusted market capitalization · Price discovery · Index reconstitution

Momentum factor

Buying recent winners and selling recent losers — a deliberate sort, not a side effect.

The pattern documented by Jegadeesh and Titman (1993): stocks that outperformed over roughly the past three to twelve months tended, on average, to keep outperforming over the following months. As a strategy it is a deliberate construction — sort securities by prior returns, buy the winners, short or avoid the losers, rebalance on a schedule. That is different in kind from market-cap weighting, which sorts nothing and merely lets winning positions grow. Momentum's historical premium has come with sharp crashes, high turnover, and real trading costs.

See also: Factor investing · Market-capitalization weighting

Price discovery

The market process that turns competing judgments and trades into a price.

The process by which trading among investors with different information and views produces a market price. Every order — fundamental, quantitative, technical, forced, or indifferent — pushes on the negotiation, and the price that clears is the market's working estimate of value, continuously revised. It is the output of disagreement, not a declaration of truth: prices can stay wrong for long stretches, but they carry the aggregated judgment a cap-weighted index inherits.

See also: Efficient Market Hypothesis (EMH) · Market-capitalization weighting

Survivorship bias

The distortion from studying only what survived — winners stay visible, failures vanish.

The error of measuring history using only the entities still alive at the end. Funds that were merged or liquidated, and companies that were delisted, drop out of naive samples, flattering the average that remains. It is why a backtest built on today's index members overstates what an investor could actually have earned, and why serious scorecards measure the full opportunity set as it stood at the start of each period — not just the survivors.

See also: Index reconstitution

Bearish divergence

Price makes a higher high while momentum makes a lower high — a fading advance.

When price makes a higher high while a momentum indicator such as RSI makes a lower high, the indicator fails to confirm the new price high. This shows weaker momentum relative to the earlier peak, but it does not date a top: divergence can persist while price continues rising. On its own, RSI divergence does not prove falling market participation, institutional distribution, or large-holder selling.

See also: Relative Strength Index (RSI) · Simple moving average (SMA)

Distribution

Large holders selling into a rising market — supply handed off near a top.

In technical analysis, the process of large or informed holders selling their positions into strength, typically while prices are still rising and the wider public is buying. It is the mirror image of accumulation and, in the Wyckoff framework, the phase that precedes a decline. "Selling into strength" describes the same behaviour. It is inferred from price and volume rather than observed directly.

See also: Wyckoff phases · Bearish divergence

Elliott Wave

R. N. Elliott's theory that markets move in repeating five-up, three-down wave patterns of crowd psychology.

A framework set out by Ralph Nelson Elliott in the 1930s: market prices unfold in repetitive wave patterns driven by swings in crowd psychology — a five-wave 'impulse' in the direction of the trend (labelled 1–5), then a three-wave 'correction' against it (A–B–C). The pattern is held to be fractal, the same shape recurring across timeframes. It describes the rhythm of advances and declines; it is a lens on market structure, not a precise forecasting rule.

See also: Bearish divergence · Wyckoff phases

Relative strength

Comparative performance across assets or versus a benchmark — distinct from the single-asset RSI oscillator.

Relative strength, also called comparative relative strength or price relative, compares the performance of one security, sector, asset class, or portfolio with another — usually a benchmark. It is commonly plotted as a price ratio: a rising ratio means the first asset is outperforming, while a falling ratio means it is underperforming. Repeating the comparison against a common benchmark helps rank leaders, laggards, and rotation across a group. Do not confuse it with the Relative Strength Index (RSI), a 0–100 oscillator computed from one asset's own recent average gains and average losses. In Wilder's RSI formula, "relative strength" means average gain divided by average loss — not performance versus another asset. Both can inform momentum analysis, but they measure different relationships.

See also: Relative Strength Index (RSI) · Sector rotation

Relative Strength Index (RSI)

A 0–100 momentum gauge of how one-sided recent gains have been.

A momentum oscillator that scores recent price action from 0 to 100 by weighing the size of gains against losses over a lookback window — here 14 and 30 days, using Wilder's smoothing. It uses only the instrument's own price history; it does not compare that instrument with a benchmark or another asset. Readings above 70 are conventionally called overbought and readings below 30 oversold, although neither is a standalone trading signal. A common application is divergence analysis: when price makes a higher high while RSI makes a lower high, the new price high has not been confirmed by equally strong internal price momentum. That is a warning of weakening momentum, not proof of an imminent reversal.

See also: Bearish divergence · Relative strength

Retracement

A partial reversal within a larger trend — a dip inside an advance, or a bounce inside a decline.

A temporary move against the prevailing trend that gives back only part of the preceding move before the trend resumes. An uptrend that rises 100 points and then falls back 30 has retraced 30% of the advance. Retracements are distinguished from full reversals, which give back the whole move and turn the trend; telling the two apart in real time is one of the central difficulties of trend-following.

See also: Simple moving average (SMA)

Simple moving average (SMA)

The average closing price over the last N days, redrawn each day — a smoothed trend line.

The average of the closing price over a fixed window of the last N trading days, recomputed every day so the line slides forward with the market. It smooths out day-to-day noise to show the underlying trend. Different periods answer different questions: a 50-day average tracks the intermediate (roughly two-month) trend, while a 200-day average smooths about a year and is the common yardstick for the long-term trend — which is why a chart usually plots more than one. Price holding above a rising average is read as an uptrend.

See also: Relative Strength Index (RSI)

Technical analysis

Reading a market's own price and volume — supply and demand on the chart — to judge probable direction.

The study of a market's own price and volume history — the running record of supply and demand — to gauge trend, momentum and probable direction, rather than the worth of the underlying business (that is fundamental analysis). Its premise is that price already reflects what is known, and that the patterns crowds leave on a chart tend to repeat. It is in fact the older discipline: charting traces to 18th-century Japanese rice traders and to Charles Dow's writings around 1900, decades before Graham and Dodd codified fundamental valuation in the 1930s.

See also: Relative Strength Index (RSI) · Simple moving average (SMA) · Bearish divergence

Wyckoff phases

Richard Wyckoff's four-stage cycle: accumulation, markup, distribution, markdown.

A framework from Richard Wyckoff describing a market cycle in four phases: accumulation (large buyers building positions quietly at a base), markup (the rising trend), distribution (those buyers selling into strength near the top), and markdown (the decline). It is a lens on the structure of a cycle and the hand-off of holdings from more-informed to less-informed participants, not a precise timing rule.

See also: Distribution · Elliott Wave

Bear market

A decline of 20% or more from a peak.

A market decline conventionally defined as a fall of 20% or more from a recent peak. Bear markets tend to be shorter and sharper than the bull markets around them, and historically have given back only a fraction of the preceding advance. The term also describes a pessimistic stance ("bearish").

See also: Bull market

Bull market

A sustained rise in prices, conventionally a gain of 20% or more from a low.

A prolonged period of rising prices in a market, conventionally marked from a low once prices have risen 20% or more. Bull markets have historically lasted longer and travelled further than the declines that separate them. The term also describes a broadly optimistic stance ("bullish").

See also: Bear market

Market breadth

How many stocks are taking part in a move, not just where the index is.

A measure of how widely a market move is shared across its individual stocks, rather than how the headline index is doing. Breadth is strong when most stocks are rising with the index and weak when a rising index is carried by only a few names — a divergence often read as a sign of a fragile advance. It is one component of what is meant by "market internals."

See also: Bearish divergence

Sector rotation

Money shifting between industry sectors as the economic cycle turns.

The movement of investment capital between industry sectors — technology, energy, financials, utilities and so on — as the economic cycle progresses and different parts of the market are rewarded at different stages. Tracking which sectors lead and which lag is a way of reading where the market thinks the cycle is, and a principal use of relative-strength analysis.

See also: Relative strength · Top-down analysis

Top-down analysis

Starting from the macro picture, then sector, then the individual security.

An approach that works from the broad picture inward: the macroeconomic backdrop sets the opportunity set, sector and factor trends narrow it to where returns are being rewarded, and individual security selection happens within that frame. It contrasts with bottom-up analysis, which starts from the individual company. The two are often combined.

See also: Sector rotation