WhitmanTrading

Silver Bullet: An Hour With a Name

The silver bullet is a setup traded only during a fixed one-hour window, conventionally 10 to 11 in the morning New York time, using an ordinary imbalance entry. Participation really is higher in that hour, but a time filter changes when you look rather than what you find.

How it works

A candlestick chart of the site's shared price history. The headline on the chart reads: A one-hour window given a name.
A one-hour window given a name. Illustrative chart - not real market data.

The silver bullet is a time filter with a setup attached. Trade only during a specific hour, look for one specific thing, and do nothing outside it.

A gently rising stretch of the long price series. The headline on the chart reads: Ten to eleven in the morning, New York time.
Ten to eleven in the morning, New York time. Illustrative chart - not real market data.

The best-known window is 10:00 to 11:00 New York time, an hour after the equity open. Other windows circulate — one in the London session, one in the afternoon — with the same structure.

A flat but volatile stretch of the long price series. The headline on the chart reads: The setup inside it is an ordinary imbalance entry.
The setup inside it is an ordinary imbalance entry. Illustrative chart - not real market data.

Inside the window, the setup is an ordinary fair value gap or imbalance entry: wait for a fast move that leaves a gap between candle ranges, then enter on the return to it. Nothing about the setup is specific to the hour.

What the hour actually has

A calmly advancing stretch of the long price series. The headline on the chart reads: The hour is real: participation is genuinely higher then.
The hour is real: participation is genuinely higher then. Illustrative chart - not real market data.

Participation in that hour is genuinely elevated, and that is a real, checkable fact. The equity open has settled, European markets are still active, and the volume profile of a normal trading day has a recognisable morning peak. Higher participation means tighter spreads, better fills and more movement to work with.

That is a legitimate reason to prefer trading then, and it is the strongest thing that can be said for the concept.

A flat, quiet stretch of the long price series. The headline on the chart reads: A clock is not an edge, it is a filter on when you look.
A clock is not an edge, it is a filter on when you look. Illustrative chart - not real market data.

What it is not is an edge in the setup. Restricting when you trade changes the sample of opportunities you see; it does not change what happens when you take one. If the imbalance entry has no edge, doing it between ten and eleven does not give it one.

The improvement most people get from a time filter is behavioural, and it is real: fewer hours in front of a screen means fewer marginal trades, less fatigue, and a rule that is easy to follow. Those are worth having. They are just not what the name implies.

A candlestick chart with a volume histogram beneath it, with the volume histogram emphasised. The headline on the chart reads: Participation is the one thing the hour genuinely has.
Participation is the one thing the hour genuinely has. Illustrative chart - not real market data.

In practice: the problems with the window

A declining stretch of the long price series. The headline on the chart reads: And the window moves twice a year with daylight saving.
And the window moves twice a year with daylight saving. Illustrative chart - not real market data.

A window defined in New York time moves relative to London and Tokyo twice a year, and the transitions are on different dates in different regions. For several weeks a year, the hour is not the hour the rule was described for.

Anyone trading it from another time zone has to track this, and material about the setup rarely mentions it.

A strongly rising stretch of the long price series. The headline on the chart reads: Scheduled news lands inside it more often than not.
Scheduled news lands inside it more often than not. Illustrative chart - not real market data.

And a great deal of scheduled economic data is released at 10:00 New York time. So the window frequently contains a news reaction, which is a confound rather than a feature: any measured performance of the hour is partly measuring the response to data releases.

Separating those is straightforward and nobody does it. Split your results into days with a scheduled 10:00 release and days without. If the edge is entirely in the news days, you are trading the calendar, not the clock.

A long-horizon candlestick view of the same price series. The headline on the chart reads: On a daily chart the hour does not exist at all.
On a daily chart the hour does not exist at all. Illustrative chart - not real market data.

The concept only exists intraday. On a daily chart there is no hour, and any structure it identifies is inside a single bar.

A candlestick series containing several opening gaps, with the largest opening gap marked. The headline on the chart reads: And the opening gap is already an hour old by then.
And the opening gap is already an hour old by then. Illustrative chart - not real market data.

By ten o’clock the opening gap has been traded for an hour, so the window deliberately skips the session’s most volatile stretch. That is a defensible choice — the open is expensive to trade — and it should be stated as one.

A declining stretch of the long price series, with the entry price and the level at which a stop would trigger drawn as horizontal lines. The headline on the chart reads: The stop comes from the setup, not from the clock.
The stop comes from the setup, not from the clock. Illustrative chart - not real market data.

The stop comes from the imbalance, not from the time. A clock supplies no levels, so every risk decision is inherited from the underlying setup.

A candlestick chart of the site's shared price history, annotated with the round-trip cost. The headline on the chart reads: Each attempt costs a share of a bar.
Each attempt costs a share of a bar. Illustrative chart - not real market data.

Each attempt costs 2% of a typical bar’s range in round-trip costs on this history. Fewer trading hours means fewer attempts, which is a genuine saving.

A 72-bar candlestick section of the shared price history. The headline on the chart reads: No order is placed because of what time it is.
No order is placed because of what time it is. Illustrative chart - not real market data.

And nothing in the order book responds to the hour as such. What changes is how many participants are present, which is a description of the crowd rather than a mechanism in the market.

What the silver bullet is not

It is not a pattern. It is a time filter plus an existing entry.

It is not unique to one hour. The same setup exists at every hour; the rule restricts when you take it.

It is not free of the daylight-saving problem, which quietly changes the window relative to other sessions twice a year.

And it is not separable from the news calendar without doing the split yourself.

When it fails

A sideways, range-bound candlestick series. The headline on the chart reads: And on a quiet day the hour is as flat as any other.
And on a quiet day the hour is as flat as any other. Illustrative chart - not real market data.

On a quiet day the hour is as flat as any other. Participation is higher on average, and averages contain many days that are nothing like the average. A rule that says “trade this hour” produces pressure to find something in an hour that contained nothing.

The second failure is exactly that pressure. A one-hour window and a daily expectation of a trade is a combination that manufactures marginal setups. The filter helps only if “no trade” is a permitted outcome.

A third is the news confound. Without splitting scheduled-release days from ordinary ones, you cannot tell what you are trading.

A fourth is the time zone drift. Two periods a year where the window is an hour off for anyone outside New York, and the results from those weeks are not comparable.

And a fifth is treating the window as the reason it worked. If the underlying imbalance entry is sound, it is sound at other times too, and the hour is a convenience. If it is not sound, the hour cannot rescue it.

The original data

On this site’s shared 576-bar history the 10-bar efficiency ratio has a median of 0.34, a ninetieth percentile of 0.71, and exceeds 0.5 on only 30% of bars. The distribution is in research/series-measurements.json, produced by site/measure_series.py.

A 72-bar window of the shared price history, cut short at the decision bar. The headline on the chart reads: It is ten forty and nothing has set up. Force one?
It is ten forty and nothing has set up. Force one? Illustrative chart - not real market data.

That distribution is the honest frame for any time-based rule. Only three bars in ten sit in conditions where price travels efficiently, so on most days — in any hour — the market is churning. The measurement that would settle the silver bullet is one you can make from your own broker statements: split every trade by hour and by whether scheduled data was released, and compare. If the ten-o’clock hour outperforms only on release days, the rule is a news filter with a time label; if it outperforms on both, you have found something worth keeping. Nobody else’s backtest answers that for your instrument and your execution.

The inner circle trader material, ICT concepts, is the teaching tradition this belongs to. Trading sessions covers what genuinely differs between hours. And fair value gap is the setup the window is wrapped around.

What I actually do

Time filters did more for my results than any setup ever has, and not for the reason they are usually sold. Trading one hour instead of six stopped me taking the marginal trades I took out of boredom - the improvement was in my behaviour, not in the hour.

— Michael Whitman

This page is educational, not financial advice. Test every idea on your own charts before risking money.