Learn › Event analysis

Event Analysis: What Happens After a Market Event

Event analysis takes a market condition — an RSI reading below 30, a MACD crossover, a candlestick pattern, a breakout above a channel — finds every time it occurred in the past, and measures what happened to price, volatility and volume around those moments. The question it answers is what has tended to follow this condition, and how widely did the outcomes vary?

Define an event, find it, measure around it

• Define the event as an exact condition, the same way as a trading rule: RSI(14) < 30, MACD crosses above its signal line, Close above the 20-day Donchian upper band.
• Find every occurrence across the chosen instruments and period.
• Measure a window before and after each one — the price path, the size of the move, how quickly it happened, volatility and volume.
• Compare the result with the ordinary bars that were not events, so the event is judged against what the market was doing anyway.

Why a single example is not enough

A chart annotated with one well-chosen occurrence shows what can follow a condition. Event analysis shows what did follow it across every occurrence, including the ones that went nowhere. The useful outputs are distributions: the typical path, and how far the outcomes spread around it.

How to read the results

• The median path — the middle outcome at each bar after the event. Medians are used because a few extreme outcomes can drag an average a long way.
• Percentile bands — the range that held the middle 75% and 95% of outcomes. A wide band around a small median move means the condition says little about direction.
• The scatter of individual outcomes — every event as its own point, so a result carried by a handful of cases is visible.
• Volatility and volume — how turbulent the market was around the event, measured against that episode's own normal level, and how much trading there was.
• The number of events — every figure rests on a count. Twelve events and twelve hundred are different kinds of evidence.

Two design choices that change the answer

An event lasts as long as its condition. "RSI below 30" is often true for several bars in a row. Counting each bar as a separate event would count one episode many times. Treating the whole stretch as one event — starting on its first bar and ending on its last — keeps episodes independent, and the window after it is measured from where the condition ended.

Conditions depend on the market around them. The same signal can behave differently in a rising and a falling market, or in a calm and a volatile one. Splitting the events by the trend or volatility at the time shows whether a result belongs to one market state. The Test Bench's MACD studies apply this: what followed 15,698 crossovers across 197 US large caps, then the same comparison conditioned on trend and volatility. Our RSI studies test both moments of an oversold spell: the day RSI crosses below its line and the day it crosses back.

Event studies in finance, and in technical analysis

In finance research, an event study measures how a stock's price behaved around a dated event — an earnings release, a merger announcement, an index inclusion — usually as the return beyond what the market as a whole did. The method is the same one described here: define the event, align every occurrence on it, and measure the window around it.

Event analysis in GU Analyser applies that method to market conditions defined from price, volume and indicators, rather than to corporate announcements. It suits questions such as "what followed a close below the lower Bollinger Band?" or "did volume rise after a squeeze released?".

What event analysis can and cannot establish

It describes what followed a condition in the data tested. It does not turn the condition into a trading strategy — there are no entries, exits, sizing or costs — and a clear historical pattern is evidence about the past, not a forecast. To test a full set of rules with trades and costs, see how to backtest a trading strategy without code.

Doing this in GU Analyser

• The condition is built on the same visual canvas as a strategy, from any of the 60+ indicators and price, with AND, OR and NOT. Premium adds candlestick patterns as conditions.
• Choose the timeframe (daily or hourly; weekly and five-minute on Premium with your own Tiingo key), how far back to look, and how many bars after each event to measure.
• Preconditions restrict the study to one trend or volatility state, for the events and the comparison bars alike. Compare side by side reports the two halves of a market state together.
• Events can be separated by whether the condition retriggered soon after, so clustered signals are not counted as independent evidence.
• The free plan studies the S&P 500 ETF (SPY). Premium pools up to ten instruments into one sample, across 400+ US stocks, ETFs and indices.
• With your own Tiingo key, up to 50 Tiingo instruments (100 on a Tiingo Power key) combine into one study. They run ten at a time, and the result is the same as one run over all of them.
• Premium saves up to 30 instrument lists in the List builder. With your Tiingo key on, a list holds Tiingo instruments, up to 100; without it, up to ten.
• A run that spans more than five missing trading sessions in an instrument's data is refused, with an option to start that instrument after the gap.

For what the conditions themselves measure, see the indicator reference — for example RSI, MACD and Bollinger Bands.