What happens after a pattern fires.

A pattern either predicts something or it does not. An event study is the plainest way to find out, before you add a single stop or target.

What an event study measures

An event study asks one question: after this condition happens, what does the market actually do next? Not what a strategy would do with a stop and a target in the way, just the raw forward return. Strip away trade management and you get the cleanest possible read on whether a pattern predicts anything at all.

How it works

You define a condition: a level, a crossover, a combination of a few things being true at once. The system finds every historical bar where that condition held, and measures the return over several forward horizons from each one. Those returns get compared to what the market does on an average day, the baseline, so a positive number that just tracks the market's normal drift does not get mistaken for an edge.

Testing several horizons matters because the same pattern can behave differently over different windows. A pattern might show something at a short horizon that fades by a longer one, or the reverse. One horizon gives you a single data point. Several give you a shape.

Comparing across calm and volatile markets

A move that looks large in a quiet market can be unremarkable during a volatile stretch, and the reverse is also true. ATR normalization expresses each forward return in units of the market's own recent range at the time, rather than a raw percentage. That makes results comparable across different periods in the same market, and across markets that move at completely different scales.

Event study versus backtest, different questions

An event study and a backtest are not competing methods. They ask different things. The event study asks whether the pattern predicts returns at all, with nothing else in the way. A backtest asks something narrower: can you actually capture that return with a real stop loss, take profit, and holding period. Run the event study first. If there is nothing there, no amount of stop placement will invent an edge that does not exist.

The four verdicts

Validation checks every result against two things: is the return different from zero, and is it different from what the market does anyway. Those two checks produce four outcomes.

• Strong edge. The return clears both checks: real, and better than the market.
• Marginal. The return clears the first check but not the second: real, but close to what the market does on its own.
• Pinned. Flat on its own, but different from the market baseline. Often useful for hedging rather than for a directional trade.
• No edge. Neither check clears. This is a complete result, not a failure. It tells you not to spend more time on the idea.

Why a spreadsheet cannot do this

You can calculate a forward return in a spreadsheet in one line. What a spreadsheet will not do on its own is handle overlapping occurrences correctly, hold a baseline steady for comparison, or apply a correction as the number of things you have tried grows. Those three pieces, the counting, the baseline, and the correction, are what turn a number into an answer you can trust. Skip any one of them and the average you get back is decoration, not evidence.

How Validation runs one

You describe the condition in plain language. The system builds it, shows you where it fired on a chart, and only runs the study once you approve what you see. From there it measures forward returns at several horizons, applies the baseline comparison and the correction for how many things you have tested on this idea, and returns one of the four verdicts. If the verdict is strong, the natural next step is a backtest to see if the edge survives real trade management. Every occurrence, every horizon, and every comparison is recorded, so the result comes with a trail you can check.

Questions people ask

What is an event study in trading?

An event study measures what happens to returns after a specific condition fires, with no stop loss or take profit involved. It answers whether a pattern predicts anything at all, before you add any trade management.

How is an event study different from a backtest?

An event study measures raw forward returns after a signal, while a backtest simulates actual trading with a stop loss and take profit. Run the event study first: if there is no signal, no amount of stop placement will create one.

What is ATR normalization?

It expresses each forward return in units of the market's own recent range at the time, rather than a raw percentage. That makes results comparable across calm and volatile periods, and across markets that move at very different scales.

What are the four edge verdicts?

Strong edge means the return is real and beats the market. Marginal means it is real but close to what the market does anyway. Pinned means flat returns that still differ from the market baseline. No edge means neither check passed, which is a complete result on its own.

Can I test the same pattern on more than one market?

Yes. The same formula can be tested across 2 to 10 related markets at once, and a pattern that holds across several of them is harder to dismiss as a coincidence in one dataset.