Each term below says what it is, why it helps you judge a portfolio, and what to watch out for. No single number tells the whole story, so use several together.
No term matches that search.
CAGR
Compound annual growth rate
(ending value ÷ starting value) ^ (1 ÷ years) − 1
What it is. The steady yearly rate that would turn your starting value into your ending value.
Why it helps. One number that lets you compare results over different lengths of time.
Watch out. It hides the ride. Two portfolios with the same CAGR can feel completely different.
Drawdown
Fall from the last high
What it is. How far a portfolio has dropped from its previous peak, at any moment.
Why it helps. It shows what holding it actually felt like.
Watch out. A short test period can hide the big drops.
Max drawdown (MaxDD)
The worst fall
What it is. The largest peak-to-low drop in the period.
Why it helps. It tells you the worst pain you would have had to sit through. Losses are not symmetric: you need a bigger gain to climb back.
−10% needs +11% · −20% needs +25% · −34% needs +51% · −50% needs +100%
Watch out. It rests on one single event. The future worst case can be worse.
Ulcer Index
How deep and how long
What it is. A measure of drawdowns that counts both depth and duration. It squares each day's drawdown, averages the squares, and takes the square root. Lower is better.
Why it helps. A long, deep slump costs more peace of mind than a brief dip. This number reflects that.
Watch out. It depends on the window you pick.
UPI
Ulcer Performance Index
return ÷ Ulcer Index
What it is. Return per unit of drawdown pain. Some versions subtract a risk-free rate first. Higher is better.
Why it helps. It is our main ranking number, because it rewards a smooth ride and quick recoveries, not just a high return.
Watch out. Compare it over the same window for every Keel. A short window can flatter it.
Sortino ratio
Return per unit of downside risk
What it is. Return above a minimum (often the risk-free rate) divided by the volatility of losing periods only.
Why it helps. It does not punish big up days.
Watch out. It does not see how long a drop lasts, so we weigh it below UPI.
Sharpe ratio
Return per unit of total volatility
What it is. Return above the risk-free rate divided by the volatility of all returns, up and down.
Why it helps. It is the most widely quoted measure, so it makes comparison with other sources easy.
Watch out. It treats gains as risk, and can flatter strategies that are smooth until they break.
Calmar ratio
Return per unit of worst fall
CAGR ÷ |max drawdown|
What it is. Yearly growth divided by the worst drawdown, often over three years.
Why it helps. A simple answer to: how much did I earn for the worst fall I sat through?
Watch out. It rests on one event, like MaxDD does.
Volatility
How much returns bounce
What it is. The spread of returns around their average, usually shown per year.
Why it helps. It sets expectations for day-to-day swings and feeds the weighting of holdings.
Watch out. Not the same as the risk of losing money.
Benchmark (SPY)
The yardstick
What it is. A reference to compare against. SPY is an ETF that tracks the S&P 500.
Why it helps. It tells you if the effort beat doing nothing.
Watch out. A benchmark may hold different assets and carry different risk than your Keel.
ETF
Exchange-traded fund
What it is. A basket of securities that trades like a stock.
Why it helps. One ticker gives broad exposure, so rules can rotate between asset classes easily.
Watch out. Not all ETFs are alike. Leveraged and inverse ETFs carry much higher risk.
Rebalancing
Bringing holdings back to plan
What it is. Buying and selling so the portfolio matches the Keel's target weights, on a schedule such as daily, weekly or monthly.
Why it helps. It keeps the portfolio you hold the same as the one you tested.
Watch out. More trading means more costs, and more taxes in taxable accounts.
Turnover
How much you trade
What it is. The share of the portfolio that is traded in a period.
Why it helps. A quick way to estimate costs and tax impact.
Basis point and slippage
bp · cost of trading
What it is. One basis point is 0.01%. Slippage is the gap between the price you expected and the price you got.
Why it helps. Our tests include a cost per trade (2 bp in the examples) so results are not unrealistically clean.
Watch out. Real costs can be higher, especially in thin ETFs.
Moving average and trend filter
SMA · EMA · ZLEMA
What it is. The average price over the last N days. Price above it suggests an uptrend. An EMA gives recent days more weight. A zero-lag EMA (ZLEMA) tries to cut the delay.
Why it helps. A simple, testable way to say whether an ETF is in favor.
Watch out. In choppy markets, trend rules flip back and forth and lose money.
RSI
Relative Strength Index
What it is. A 0 to 100 gauge of recent gains versus losses over N days. High means strong recent gains and low means weak.
Why it helps. Spots stretched or weak moves.
Watch out. It can stay high or low for a long time in a strong trend.
ATR
Average True Range
What it is. The typical daily price range of an ETF.
Why it helps. A measure of how jumpy a holding is, used to size bands and filters.
Inverse-volatility weighting
More of the calm, less of the wild
What it is. A holding's weight is higher when its price moves less.
Why it helps. No single wild ETF dominates the portfolio's ups and downs.
Watch out. It tilts toward calm assets such as bonds, which can lag in strong stock rallies.
Crash switch
A fast defensive rule
What it is. A rule that moves to defensive holdings after a sharp fall. For example: the S&P 500 drops more than 6% in 10 days.
Why it helps. Limits the damage in sudden crashes that slower trend rules catch late.
Watch out. It can trigger late or by mistake, and it can miss the rebound.
Regime
The market's current mode
What it is. A description of the market now, such as uptrend, downtrend or high volatility.
Why it helps. Rules can be chosen to fit the regime.
Watch out. Regimes are clear only in hindsight.
Back-test and hypothetical performance
Simulation on past data
What it is. A replay of your rules on historical prices.
Why it helps. You can see how a plan would have behaved before risking money.
Watch out. It is made with hindsight and does not predict the future. See the Hypothetical Performance notice.
Overfitting
Rules that fit the past too well
What it is. Settings tuned so closely to old data that they describe noise, not a lasting pattern.
Why it helps to know. It is the most common reason a good back-test fails in real life.
Warning signs. Many tuned settings, a short history, and a curve that looks too perfect.
In-sample and out-of-sample
Data used to design vs data held back
What it is. In-sample data is what you used to build the rules. Out-of-sample data is what you kept aside, or what happens after.
Why it helps. If a Keel still works on data it never saw, it is more believable.
Paper trading
Practice with fake money
What it is. Running the plan in a practice account.
Why it helps. Finds mistakes in settings, data and timing before real money is involved.
Keel
Your written plan
What it is. A set of rules, written in advance, that decides which ETFs you hold and how much of each.
Why it helps. Decisions are made calmly beforehand, tested, and applied the same way every time.
Beta
Sensitivity to the market
covariance(asset, benchmark) ÷ variance(benchmark)
What it is. How much an asset tends to move when its benchmark moves. A beta of 1.0 moves with the benchmark, 0.5 moves half as much, 1.5 moves half again as much, and a negative beta moves the other way.
Why it helps. It shows how much market risk you are carrying, and how a holding changes the swings of your portfolio.
Watch out. It depends on the benchmark you choose and the period you measure. It changes over time.
Alpha
Return beyond market exposure
return − [risk-free + beta × (benchmark return − risk-free)]
What it is. The part of an asset's return that its beta does not explain, usually shown per year (Jensen's alpha).
Why it helps. A positive alpha means it earned more than its market exposure alone would predict. It is a fairer test of a Keel than raw return.
Watch out. Alpha is noisy, especially over short windows, and past alpha often disappears. Check R² and the length of the window before you trust it.
Correlation
How closely two things move together
What it is. A number from −1 to +1. At +1 they move in step, at 0 they are unrelated, at −1 they move in opposite directions.
Why it helps. Holdings with low correlation to each other smooth a portfolio. Ten funds that all move together are really one bet.
Watch out. Correlations rise in crashes, just when you want diversification most.
R-squared (R²)
How much the benchmark explains
What it is. The share of an asset's day-to-day movement that the benchmark accounts for, from 0 to 1.
Why it helps. It tells you how much to trust beta and alpha. At 0.9 they are reliable. At 0.1 they describe little.
Tracking error
How far from the benchmark
What it is. The yearly spread of the daily return differences between a portfolio and its benchmark.
Why it helps. A low number means it behaves like the benchmark. A high number means it can look very different, for better or worse.
Information ratio
Extra return per unit of straying
average return difference ÷ tracking error
What it is. How much return above the benchmark you earned for each unit of tracking error, per year.
Why it helps. It separates steady outperformance from a lucky streak. Higher is better.
Watch out. Needs a long window to mean much.
Up capture and down capture
How much of the ups and downs you get
What it is. Up capture is its average return on days the benchmark rose, as a percentage of the benchmark's. Down capture is the same for days the benchmark fell.
Why it helps. You want high up capture and low down capture. 70% up and 40% down means you took most of the gains for far less of the losses.
Watch out. They depend on the period. A strong rebound can change them a lot.
Downside deviation
Volatility of losses only
What it is. The spread of returns that fall below a target, such as zero or the risk-free rate. Up days do not count.
Why it helps. It is the risk measure behind the Sortino ratio, and matches how investors feel risk.
VaR and expected shortfall
VaR 95% · ES 95%
What it is. Value at risk (VaR) at 95% is the loss that was matched or exceeded on the worst 5% of days in the history. Expected shortfall (ES) is the average loss on those worst days.
Why it helps. A plain answer to "how bad can a bad day be?". ES is more cautious because it looks at the tail.
Watch out. Based on history, so it can understate a new kind of crisis. It does not say how bad the very worst day can be.
Time under water
Drawdown duration
What it is. The longest stretch of time spent below a previous high before getting back to it.
Why it helps. Two portfolios with the same maximum drawdown can differ by years in recovery time. This shows how long you may wait.
Rolling returns
Returns over every window
What it is. Returns measured over many overlapping windows of the same length, such as every 1-year period, instead of one fixed start date.
Why it helps. It shows consistency and how much your result depended on when you started.
Risk-free rate
The return of waiting in cash
What it is. The yield on a very safe holding, such as short-term US Treasury bills. Sharpe, Sortino and alpha measure return above it.
Why it helps. It makes the question fair: how much did you earn beyond what cash would have paid?
Total return and price return
With or without dividends
What it is. Price return counts only the change in price. Total return also counts dividends and distributions that are reinvested.
Why it helps. For income-focused ETFs, price return alone can understate results by a lot.
Expense ratio
What the fund charges each year
What it is. The yearly fee, as a percentage of your holding, taken out of an ETF's returns.
Why it helps. Fees compound against you. A 0.03% fund and a 0.90% fund can end up very far apart over decades.
Position weight and concentration
How much sits in each holding
What it is. Weight is a holding's share of the portfolio's value. Concentration measures how much sits in a few holdings. The effective number of holdings is 1 divided by the sum of squared weights.
Why it helps. Ten equal holdings count as 10. One position of 90% counts as about 1.2, however many tickers you own.
Risk contribution
Who creates the swings
What it is. Each holding's share of the portfolio's total volatility, which depends on its weight, its own volatility and how it moves with the rest.
Why it helps. A holding that is 15% of the money can be 40% of the risk. This shows where the swings really come from.
Contribution to return
Who earned the gains
What it is. The percentage points of the portfolio's total return that came from each holding over a window. They add up to the portfolio's total.
Why it helps. Shows which positions did the work, and which dragged.
Educational definitions only. They are not investment advice.