Trading and statistics, explained for humans.

Anchoring

Judging a number by comparing it with another number you saw first, even when that first number is meaningless.

ATR (Average True Range)

How much a stock typically moves in a day, in dollars, averaged over recent days (usually 14).

Backtest

Testing a trading rule on historical data to see how it would have done.

Base rate

How common something is in the whole group, before you look at any particular story. Without it, a striking example can't tell you whether something is rare or typical.

Break even

The point where you have made back everything you lost, so your total profit is exactly zero.

Breakout

When a price rises above a level it couldn't get past before, such as its highest point of the last 20 or 60 days.

Bull and bear markets

A bull market is a long stretch of rising prices. A bear market is a long fall, often defined as 20% or more from the top.

Calibration

A forecaster is well calibrated when things they call '30% likely' happen about 30% of the time, and things they call '90% likely' happen about 90% of the time.

Cherry-picking

Choosing only the data, examples or time period that support your story, and quietly leaving out the rest.

Clustering illusion

The tendency to see meaning in clusters that randomness produces on its own. Random events don't arrive evenly spaced; they bunch up and leave gaps.

Confidence interval

A range around an estimate showing how far off it could plausibly be. A 95% interval is roughly the estimate plus or minus two standard errors.

Confounding variable

A hidden third thing that affects two other things at once, making them look connected when neither causes the other.

Control group

A comparison group that gets everything except the thing you are testing, so you can see what that thing actually adds.

Correlation

A number from -1 to 1 that shows how closely two things move together. Zero means no link at all.

Crossover

The moment one line on a chart passes through another, for example a fast moving average rising above a slow one.

Death cross

The opposite of a golden cross: a short-term moving average falls below a long-term one. Often treated as a sell signal.

Excess return

A return minus what the market did over the same period. It shows what a strategy added beyond just being invested.

Expected value

The average result you would get if you repeated something many times, weighting each outcome by how likely it is.

Fibonacci retracement

Price levels drawn at fixed fractions (such as 38.2%, 50% and 61.8%) of a previous move, where traders expect the price to pause or turn.

Flaw of averages

The mistake of treating the average case as if it will actually happen. The average outcome is often rare, and the result of average inputs is not the average result.

Friendship paradox

On average, your friends have more friends than you do, because very social people appear in many people's friend lists.

Golden cross

When a short-term moving average (usually 50 days) rises above a long-term one (usually 200 days). Often treated as a buy signal.

Goodhart's law

When a measure becomes a target, it stops being a good measure, because people find ways to move the number without improving the real thing.

Hit rate

The share of trades that made money.

Illusory correlation

Seeing a link between two things that aren't related, usually because the times they happen together are memorable and the other times are forgotten.

Independent events

Events where one result doesn't change the chances of the next, like coin flips. Believing a win is 'due' after a losing streak is the gambler's fallacy.

Index fund

A fund that buys every stock in an index, such as the S&P 500, and simply holds them. No stock picking, very low costs.

Intraday

Within a single trading day, using minute-by-minute or hour-by-hour prices instead of one price per day.

Large language model

An AI system trained on huge amounts of text to understand and write language, like the chatbots people use every day.

Law of large numbers

Over many repeats, the share of each outcome settles close to its true probability. Not because luck evens out, but because early streaks get drowned out by everything that comes after.

Law of truly large numbers

With enough people, days and chances, even very unlikely coincidences are certain to happen to someone, somewhere, regularly.

Leverage

Trading with borrowed money, so gains and losses are bigger than your own cash would allow.

Margin of error

The 'plus or minus' around a result from a sample: how far it could easily be from the true answer just because of who happened to be picked.

Market drift

The general tendency of stock prices to rise over long periods. Anything that holds stocks benefits from it, whatever the reason for buying.

Median

The middle value when everything is lined up in order. Unlike the mean (add up and divide), it isn't pulled around by a few extreme values.

Moving average

The average price over the last few days, recalculated every day, so the line moves along with the price.

Multiple testing

If you test many ideas, some will look good by chance alone. A multiple-testing correction raises the bar to account for all the tries.

Non-response bias

The error that appears when the people who don't answer a survey are different from the people who do.

Normal distribution

The bell-shaped pattern you get when many small, random influences add up: most values near the average, fewer and fewer further away, both sides alike.

Null hypothesis

The boring starting assumption in a statistical test: no effect, no difference, just luck. A test asks whether the data is surprising enough to reject it.

Odds

A way of stating a probability. '1 in 292 million' means one winning outcome among 292 million equally likely ones. Betting odds such as '10 to 1' compare losing outcomes with winning ones instead.

Outlier

A value far away from the rest of the data. Outliers can drag the mean a long way while barely moving the median.

P-value

If nothing interesting were going on (the null hypothesis), the chance of getting a result at least as extreme as the one you got.

Percent change

A change measured relative to the starting value. Because each change is a percentage of a different starting point, percentages don't simply add up.

Percentage points

The plain difference between two percentages. Going from 2% to 3% is 1 percentage point, even though it is a 50% increase.

Post-earnings drift

The idea that after a company reports profits better (or worse) than expected, its stock keeps drifting in the same direction for weeks, not just on the day.

Priced in

When the market has already moved to reflect a piece of news, so acting on it afterwards gains nothing.

Probability

How likely something is, from 0 (impossible) to 1 (certain), often written as a percentage.

Pullback

A temporary drop in a stock that is otherwise rising. Buying the dip means buying during or right after that drop.

R-multiple

A trade's result measured in units of the risk you took. Risk $100 and make $300, and that trade is +3R.

Regression to the mean

After an unusually extreme result, the next one tends to be closer to average, purely by chance.

Relative risk

How many times bigger one risk is than another. 'Doubles your risk' is relative; the absolute risk might only go from 1 in 10,000 to 2 in 10,000.

Relative strength

How well a stock has done compared with other stocks over a recent period, often the last three to twelve months.

Relative volume

Today's trading volume divided by the stock's normal volume, usually the average of the last 20 days. 2.0 means twice as busy as usual.

Replication

Repeating a test on fresh data to check that the first result wasn't a fluke.

Round trip

Opening and then closing one position, for example buying a stock and later selling it. Costs are often quoted per round trip.

S&P 500

An index of about 500 of the largest US companies. When people say 'the market', they usually mean this.

Sample size

How many observations a result is based on. Small samples swing wildly; large ones settle down.

Sampling bias

When the people or things you measured aren't a fair picture of the whole group, so your results lean one way.

Selling climax

A sharp fall on very heavy volume that is supposed to mark the bottom, because everyone who wanted to sell has finally sold.

Sentiment analysis

Using software to judge whether a piece of text sounds positive or negative.

Short selling

Betting that a price will fall: you borrow shares, sell them now and buy them back later, hopefully cheaper.

Standard deviation

Roughly how far a typical value sits from the average. Small means the values are consistent; big means they are all over the place.

Statistically significant

A test of whether a result is unlikely to be pure luck. Significant means 'probably not luck', not 'big' or 'good'.

Stop-loss

An order to sell automatically if the price falls to a level you chose in advance, to limit the loss.

Sunk cost fallacy

Letting money, time or effort that's already spent, and can't be recovered, steer a decision about the future.

Support level

A price where a stock has tended to stop falling, often because buyers stepped in there before. Moving averages are often treated as support.

Survivorship bias

Looking only at the winners that are still around and forgetting the losers that disappeared.

Trading costs

Everything you pay to trade: commissions, the spread between the buying and selling price, and for shorts, borrowing fees.

Trading signal

A rule that says when to buy or sell, such as 'buy when the 50-day average crosses above the 200-day'.

Trading volume

The number of shares that changed hands in a given period, usually one day.

Trailing stop

A stop-loss that moves up as the price rises, but never moves back down.

Truncated axis

A chart axis that starts above zero, so only the top of the values is shown and small differences look huge.

Uptrend

A period when a price keeps making higher highs and higher lows, often defined as the price sitting above its moving averages.

Volatility

How much and how fast a price jumps around.

Z-score

How many standard deviations a value is above or below the average. It lets you compare results measured in completely different units.