ICT Silver Bullet Strategy: The Time Windows, the Setup and What a Chart Can Show

The ICT Silver Bullet is a time-based entry model from the Inner Circle Trader. Inside one fixed hour on the New York clock, you wait for price to sweep a high or low, then displace the other way and leave a fair value gap. You enter on the retrace into that gap, aimed at the opposing liquidity.

I build a chart grader, so before writing this I read the guides that rank for it. Nearly every one sells a course, an indicator, a funded account or a broker signup. They mostly agree on the setup. Where they slip is the clock. Several label the windows EST all year, a few print hours that contradict their own text, and most of their GMT columns are right for only half the year. The time table below is checked against NIST and GOV.UK, and the last section says what SnapPChart can and can't see on a Silver Bullet chart.

Quick Answer: What Is the ICT Silver Bullet?

Outside those hours the same pattern isn't a Silver Bullet, just a fair value gap entry. Plenty of windows produce no setup at all, and sitting those out is part of the model.

It's an Inner Circle Trader model, part of the ICT teaching of Michael J. Huddleston. ICT taught it on his YouTube channel, including a May 2023 mentorship video called ICT Silver Bullet Time Based Trading Model. I haven't transcribed that video, so the rules below are the consensus of the ranking guides, with their disagreements flagged. Some sites run Huddleston's name as a byline without being his channel, so the video is the only place I point to for ICT himself.

What Are the ICT Silver Bullet Times?

Three one-hour windows, always read on New York local time. The guides call them the London, New York AM and New York PM Silver Bullets. Each sits inside one of the broader session windows ICT calls kill zones, so the Silver Bullet is a one-hour slice of a kill zone. The ICT concepts section of the TJR guide covers kill zones. I'm not adding a kill-zone clock here, because the guides don't agree on those hours either.

The New York times never move. In winter New York runs on Eastern Standard Time (UTC-5) and in summer on Eastern Daylight Time (UTC-4), so what changes twice a year is your local equivalent and the UTC columns. UTC and GMT are the same thing for this table. Per NIST's daylight saving time page, US daylight saving starts at 2:00 a.m. on the second Sunday of March and ends at 2:00 a.m. on the first Sunday of November. In 2026 that's March 8 to November 1.

The UK changes on different dates. GOV.UK's clock change page has the clocks going forward at 1am on the last Sunday in March and back at 2am on the last Sunday in October, which in 2026 is March 29 and October 25. London is normally five hours ahead of New York. For two stretches a year it's four: March 8 to 28 and October 25 to 31 in 2026. In those weeks the London Silver Bullet runs 07:00 to 08:00 London time instead of 08:00 to 09:00, and any table built on one fixed offset is an hour out.

That's also why the "EST" label in so many guides is a problem. It's only correct from early November to early March. The rest of the year New York is on EDT, and a reader who converts "10:00 AM EST" with a UTC-5 offset shows up an hour late.

Set your chart to New York, not a fixed offset

Most hour-off mistakes start in the chart settings. A chart set to a fixed offset like UTC-5 or UTC-4 is right for part of the year and wrong by an hour for the rest. Set the platform's time zone to New York (often listed as New York or America/New_York) and the window lines stay put all year. Some platforms print the current offset next to the city name. That's fine.

Which window do most traders use?

Most of the guides treat the New York AM window, 10:00 to 11:00 AM, as the main one. That's their preference. Nobody has measured it. It opens 30 minutes after the 9:30 AM cash open on the NYSE and Nasdaq, once the opening auction is done. The post on the best time of day to trade stocks covers that first hour from the stock side. The London window at 3:00 AM New York falls in the London morning (08:00 to 09:00 London time for most of the year), which the forex sessions guide walks through, and futures trading hours covers when NQ and ES trade around all three.

What Else Is Scheduled Inside the Silver Bullet Windows?

None of the guides I read put the windows next to the actual calendar, so here it is. Every time below is New York time, taken from the source pages.

Two of those land in the first minute of a window. The BLS JOLTS release schedule lists a 10:00 AM release time, the minute the AM window opens. The Fed's FOMC statements carry a 2:00 p.m. release time (the September 17, 2025 statement reads "For release at 2:00 p.m. EDT"), which is the first minute of the PM window on meeting days. The FOMC meeting calendar lists each one, and the 2026 dates include January 28, March 18, April 29 and June 17. The CPI release schedule has CPI at 8:30 AM, 90 minutes before the AM window.

A release in the first minute can print the sweep and the displacement in one candle. That's a news spike, and it isn't the sequence the model describes. A few guides say to skip or invalidate a window with high-impact news inside it or just before it. Whatever rule you pick, check the calendar before the window opens.

The other end of the day is quieter. CME's daily maintenance break for ES and NQ runs 5:00 to 6:00 PM, well clear of all three windows, so none of them falls while the futures are shut.

How to Trade the ICT Silver Bullet, Step by Step

This is the sequence most guides describe, in the order you'd check it. The first two steps happen before the window opens. Most steps have a guide or two that does it differently, and I've said where. If the vocabulary is new, the six ICT and SMC setup criteria are a decent primer first.

1. Set the higher-timeframe bias. On the daily, 4-hour or 1-hour, decide which way the market is drawing. The Silver Bullet runs with that bias. The intraday move reverses the sweep, and the trade runs with the bigger trend. Most guides treat a gap that points against the bias as a non-setup, or at best a weaker one.
2. Mark the liquidity before the window opens. The prior day high and low, the session highs and lows (Asia, London, the overnight range), the week high and low, and any equal highs or equal lows. Stops rest beyond those levels. One side is what gets swept, and the other is your target. Some guides add a premium and discount filter, taking longs only from a gap in the lower half of the current range and shorts only from the upper half.
3. Wait for the clock. Nothing that happens before the window counts, however clean it looks. Chart on New York time.
4. Wait for the sweep. Price runs one side's stops, for a long by trading through the pre-window low and snapping back. Most guides require this step, and a minority only prefer it. All of them agree a sweep on its own is not an entry. The liquidity sweep setup covers what a clean one looks like, and why stops get hunted covers why they cluster where they do. On NQ, ES holding its own low while NQ sweeps is an SMT divergence, an optional second check.
5. Wait for displacement and a structure break. A fast, large-bodied move the other way that breaks the last swing against it, which for a long is the last lower high. Guides call the break a market structure shift (MSS) or a break of structure (BOS). Different names, same step. No impulse means no gap, and no gap means no trade.
6. Find the fair value gap. A fair value gap is a three-candle pattern where the ranges of candle 1 and candle 3 don't overlap. The space between them is the gap, and the middle candle is the displacement. It has to be created inside the window. The fair value gap strategy page covers what makes one valid, and a few guides also accept an order block left by the same move as the entry zone.
7. Enter on the retrace into the gap.Not on the displacement candle. Chasing that candle buys near the top of the move with the stop a long way below. Inside the gap there are two common choices. The edge nearest price fills more often. The 50% midpoint (ICT's consequent encroachment) gives a better price and gets missed more. Pick one before the window opens.
8. Place the stop. Guides use three references, from tight to wide. Beyond the first candle of the gap, beyond the displacement swing, or beyond the swept extreme. The swept extreme is the usual invalidation, because if price gets back past the low it just swept, the read was wrong. The tighter the stop, the more often ordinary noise reaches it.
9. Target the opposing liquidity. The pool on the other side, so for a long the equal highs, the session high or the prior day high above. Some traders take a partial at a nearer internal pool and leave the rest for the external draw. Most guides want at least 1:2 reward to risk from entry to target, and some ask for 1:3. If the target is too close for that, skip the window. The room check in the next section puts a number on "too close".

The gap forms inside the hour. The trade can outlive it.

Most guides agree on one timing rule. The displacement and the gap it leaves have to be created after the window opens, so a gap from 9:50 doesn't count at 10:05. They split on the retrace. Some cancel the order if price hasn't come back into the gap by the end of the hour. Others allow a retrace slightly after, as long as the displacement printed inside it. Many let a filled trade run past the window to its target, while others prefer to close or trail tightly near the end of the hour. Decide which rule you're running before the window and log it.

Which timeframe?

Execution happens on 1- to 5-minute charts, and the levels get marked on the 15-minute or higher. Sources split on 1-minute vs 3-minute vs 5-minute for the entry, and none of them has data on which works better. A faster chart shows you more gaps, and more of them are tiny. Pick one and stay on it long enough to judge it. The multi-timeframe analysis guide covers splitting the bias chart from the entry chart without overthinking it.

Where Power of Three fits

Some guides describe the hour as ICT's Power of Three (accumulation, manipulation, distribution) squeezed into sixty minutes. A quiet range, the sweep, then the real move. It's a handy way to read the sequence, and it doesn't change any of the rules above.

Which Markets Does It Work On, and How Much Room Does It Need?

The guides mostly agree on the markets. Liquid ones only, because a one-hour model with a 15-pip room check falls apart when the spread eats a chunk of the move.

The 10-point rule in dollars

The most repeated filter is at least 10 index points (handles) or 15 forex pips between your entry and the opposing liquidity. Some guides use 5handles. Either way it's a room check, a way to skip windows where the target sits too close to be worth the risk. It isn't a promised move.

In dollars, the E-mini S&P 500 contract specs put ES at $50 times the index, so a 0.25-point tick is $12.50 and 10 points is $500 per contract. The E-mini Nasdaq-100 contract specs put NQ at $20 times the index, $5.00 a tick, so 10 points is $200 per contract. You'll also see the rule written as 40 ticks, which is the same 10 points. The 5-handle version asks for $250 of room on ES and $100 on NQ. That's just the arithmetic. If those figures are large for your account, micro futures trade the same index points at a fraction of the dollar value.

Gold shows up in enough guides to count as common, and they don't share a room number for it. Crypto is disputed. It never closes, and some guides say a New York hour means less there. Others say the liquid pairs work.

Does the ICT Silver Bullet Actually Work?

Nobody can show you. There's no audited win rate for the Silver Bullet. Some sites quote win rates anyway, with no sample size, test period or trading costs behind them. Other guides say plainly that no audited public statistic exists, and they're right. The description under ICT's own video is the standard CFTC Rule 4.41 disclaimer about hypothetical results. So this page makes no probability claims, and I'd be wary of any page that does. Nothing about the model is guaranteed.

What the model does give you is structure. One hour, a fixed sequence, a defined stop and a defined target. That makes it easy to test honestly. Run it on a demo first, then pick one window and one market and log every session, setups and blanks alike, for long enough to judge it on your own fills. That's the only win rate that applies to you. Testing a rule on history and checking one live setup are different jobs, covered in backtesting vs grading a setup before entry.

No setup, no trade

Many windows produce nothing. No sweep, or a sweep with no displacement, or a gap that forms at 11:05. Zero-trade sessions are normal, and they're what the time filter is for. Forcing a window is how a tight model turns into overtrading. A common rule is one attempt per window, and some traders allow a second only once the first is at breakeven. That's a rule traders set for themselves, not ICT law. The guides that print a number of setups per week disagree with each other and show no count, so I'm not printing one.

Where the Silver Bullet fails

• Sweep kept going
  Price takes the low and keeps falling. That was a breakdown, not a sweep, and the stop beyond the swept extreme is what gets you out.
• News made the gap
  A JOLTS print at 10:00 or an FOMC statement at 2:00 can sweep and displace in one candle. The shape looks right and the cause is a headline, which is why some guides skip news windows outright.
• Fought the trend
  A high gets swept in a strong uptrend and you short it because the window is open. The trend can just resume and run the short over.
• No room
  A clean sequence with the opposing pool closer than your room filter fails the trade before price does anything.
• Wrong clock
  A chart on a fixed UTC offset during the out-of-step weeks puts you in the wrong hour. What you traded was a fair value gap with a misleading name.

It also isn't a full system. It has no position sizing, no daily loss limit and no rule for a run of losing windows. You still need all of those on top of it.

What Can AI See on a Silver Bullet Chart?

Less than the Silver Bullet asks for, and I'd rather say that up front. SnapPChart grades one uploaded screenshot of one instrument. It has no Silver Bullet label. It doesn't know the clock unless the chart's time axis shows it, never checks that the gap formed inside the hour, doesn't time your entry and doesn't send alerts. The time part of this model, which is most of what makes it a Silver Bullet, stays with you.

From the screenshot, it reads the ICT pieces around the trade. The liquidity sweep (with the swept level and which side), fair value gaps, order blocks and swing structure. When the time axis is readable it may also name the broader ICT session the chart is in, which is a session label and not a Silver Bullet check. Those reads are display-only. They sit next to the grade and don't move it, so a clean sweep never lifts the grade and a missing one never drags it down. This piece on how AI reads smart money concepts goes further into that.

The grade scores the trend it can see and covers with-trend pullbacks only, the same continuation logic behind a momentum trading strategy. A Silver Bullet taken with the higher-timeframe trend, like the one in the diagram, is the version it lines up with: a sweep of a pullback low inside an uptrend, then the move back up. One taken against an established trend reads as that trend's pullback, not as a reversal. And it judges the trend only from the screenshot. Zoom out far enough that the higher highs and higher lows before the window are in the shot. Crop to the window and the pre-window pullback, and it mostly sees lower highs and lower lows, which reads as a downtrend.

So pairing the two looks like this.

1. Check the clock yourself.Confirm you're inside the window on a chart set to New York time, and that the calendar is clear.
2. Screenshot after the gap forms inside the window. Ideally with the time axis visible, and zoomed out to show the higher-timeframe trend.
3. Read what comes back. The sweep, gap, order block and structure sit next to the grade, and the grade scores the setup as a pullback in whichever trend the screenshot shows.
4. Keep the timing call yours.Whether the gap formed inside the hour is something the tool doesn't check, and it won't pretend to.

Two more things to know. Futures, forex, metals, index and crypto charts have their grade capped at B, so a clean NQ Silver Bullet earns a B at most. Your first grade is free and shows the setup's grade and the general read. Entry, stop, targets and the trade plan come with a paid plan.

For the wider picture, there's more on AI chart analysis from a single screenshot, and the AI trading guide's section on what AI can't see on a chart.