Black Friday: all 12 tools. Until 1 December.Black Friday $606.98 one by one

$606.98 one by one $49 once Save $557.98Get the SuiteGet it
ProEA Lab · MT5, TradingView and trading research · Evidence standards
Systems

Which ICT Killzone Holds the Day's High? Gold and Bitcoin, Counted.

Where the day's high and low formed, by ICT killzone, on 163 days of gold and 1,476 each of Bitcoin and Ether, beside what a random walk would give.

PL · 7 min read

Short answer: of the four ICT killzones, New York AM (08:30–11:00 New York) is the only one that held the day's high and low more often than both its share of the day's time and a random walk's share on gold, Bitcoin and Ether: on 15.3% and 17.2% of 163 gold days, and 14.6% and 15.0% of 1,476 Bitcoin days. It is 10.9% of the 18:00–17:00 trading day, and a random walk would put the high there on 7.5% of days. More than half of gold's daily highs formed outside every killzone.

Killzones are windows on the New York clock that ICT traders watch. There is no official table, and the sources disagree: some list no New York PM killzone and add a London close window instead. We counted the four that ICT Levels Desk draws by default (Asia, London, New York AM and New York PM), which is also the set the killzone times tool shows first. For each window, we counted where the day's high and low formed, on the same gold, Bitcoin and Ether candles as our initial balance counts. The other sets the tool offers aren't counted here.

Gold, 163 days (1 October 2025 to 29 May 2026), New York time, 18:00 (the evening before) to 17:00.

Each roll draws 250 random-walk days of five-minute steps and marks where each day's high landed.
Gold: where the day's high and low formed, trading day
WindowTime shareRandom walkDay's high hereDay's low here
Asia (20:00–00:00)17.4%15.1%13.5% (22)14.7% (24)
London (02:00–05:00)13.0%8.5%6.1% (10)6.7% (11)
New York AM (08:30–11:00)10.9%7.5%15.3% (25)17.2% (28)
New York PM (13:30–16:00)10.9%12.1%8.0% (13)8.0% (13)
Outside every killzone47.8%56.9%57.1% (93)53.4% (87)
Counts over days, not a tested strategy. A day's high or low belongs to the window holding the first five-minute candle that reached it. 10 weekdays were left out: holidays, early closes and gaps in the feed, any day missing a candle. The random-walk share assumes the same size of move in every candle, and was added after the first count (the plan's amendment 1). Bars run from 0 to 60%.

Pick a market and a day above, then roll some random days. A random walk puts its high near the start or the end of the day more often than in the middle, so a window's share of the day's time isn't what a random walk gives it. The dashed tick shows what it does give. That walk assumes the same size of move in every five-minute candle; real markets get busier at some hours (the 08:30 data, the 09:30 open), which is one reason real days differ from it. We added the random walk's share after the first count.

What we counted

  • The day: 18:00 New York (the evening before) to 17:00, the futures day and gold's whole session in our data: 23 hours, 276 five-minute candles. A day takes the date it ends on.

  • Days counted: days ending Monday to Friday with all 276 candles.

    • Gold: 163 days, 1 October 2025 to 29 May 2026. Ten weekdays were left out for holidays, early closes and one gap in the feed.
    • Bitcoin and Ether: 1,476 days each, January 2021 to August 2026. Two weekdays were left out, the file's first and last.
    • Weekends aren't counted, because the killzones run only inside the forex and futures week.
  • The windows:

    • Asia 20:00–24:00;
    • London 02:00–05:00;
    • New York AM 08:30–11:00;
    • New York PM 13:30–16:00.

    Everything else is "outside every killzone". A candle belongs to a window when it opens at or after the window's start and before its end, so the 11:00 candle is outside New York AM.

  • Where the high formed: the window holding the first five-minute candle that reached the day's high. The same goes for the low.

  • Data:

    • Gold: Dukascopy's XAUUSD bid.
    • Bitcoin and Ether: Binance's BTCUSDT and ETHUSDT perpetuals.

    These are the candles the IB and ORB counts used, each file checked by its SHA-256.

We wrote the definitions down and committed them before the first count. After the count we added one thing, the random walk's share. It came from a formula, not from the data, and no count changed.

Where the day's high and low formed

The trading day, 18:00 to 17:00 New York. Each cell gives the share of days whose high / low formed in that window.

Window (New York)Share of the dayRandom walkGoldBitcoinEther
Asia, 20:00–24:0017.4%15.1%13.5% / 14.7%14.2% / 16.4%14.7% / 18.0%
London, 02:00–05:0013.0%8.5%6.1% / 6.7%7.3% / 5.4%7.3% / 6.2%
New York AM, 08:30–11:0010.9%7.5%15.3% / 17.2%14.6% / 15.0%14.2% / 13.3%
New York PM, 13:30–16:0010.9%12.1%8.0% / 8.0%14.6% / 14.4%12.3% / 12.2%
Outside every killzone47.8%56.9%57.1% / 53.4%49.3% / 48.8%51.5% / 50.2%

Three readings hold on all three markets:

  1. New York AM held the high and low about twice as often as a random walk would (1.8 to 2.3 times), and 1.2 to 1.6 times its share of the day's time. It is the only window above both on all three markets; on Bitcoin and Ether, New York PM is above both too. On Bitcoin and Ether it held more highs and lows than the random walk's 7.5% in every calendar year from 2021 to 2026 (2026 runs to August). Bitcoin's lowest year was 2021, with highs on 10.0% of days.
  2. London held fewer than its time share, and fewer than the random walk's. London is 13% of the day but sits in its middle, where a random walk would put the high on 8.5% of days. It held 6.1%–7.3% of highs. That is about half its time share, and 14%–36% under the random walk's share across highs and lows on the three markets.
  3. About half of the days' highs and lows formed outside every killzone. A random walk would put 56.9% there. Gold came close (57.1% of highs, 53.4% of lows); Bitcoin and Ether came lower (48.8%–51.5%). On gold, the day's high and low both formed inside killzones (not always the same one) on 18.4% of days (30 of 163), and both outside on 28.8% (47).

New York AM, year by year, on the 18:00 day (high / low; a random walk gives 7.5%, the window's time is 10.9%):

YearBitcoinEther
202110.0% / 10.8%10.8% / 10.0%
202214.6% / 19.2%17.3% / 15.4%
202314.2% / 15.8%16.2% / 13.1%
202416.4% / 11.1%13.7% / 10.7%
202517.6% / 13.4%13.4% / 10.7%
2026 (to August)15.0% / 22.5%13.3% / 23.7%

New York PM is where the markets part. On Bitcoin it held the high on 14.6% of days, above the random walk's 12.1%. On gold it held 8.0%, below it. In raw counts, ignoring each window's length, the killzones are closer than the shares make them look: on Bitcoin, New York PM held as many highs as New York AM (216 days each), and on Ether, Asia held 217 highs to New York AM's 209.

From the midnight open

ICT reads the day from New York midnight, which ICT Levels Desk calls the true day open. From 00:00 to 17:00, Asia falls before the day starts, so only three killzones are left.

Window (New York)Share of the dayRandom walkGoldBitcoinEther
London, 02:00–05:0017.6%14.2%12.3% / 11.7%14.1% / 11.6%14.3% / 13.0%
New York AM, 08:30–11:0014.7%9.5%22.7% / 23.9%19.0% / 19.4%17.3% / 18.2%
New York PM, 13:30–16:0014.7%14.4%8.6% / 10.4%16.5% / 16.5%14.6% / 14.7%
Outside every killzone52.9%61.9%56.4% / 54.0%50.4% / 52.5%53.8% / 54.1%

From the midnight open, New York AM held gold's high on nearly a quarter of days (37 of 163), and its low on nearly a quarter (39).

The silver bullet hours

The three one-hour windows every silver bullet guide we checked gives, counted on the 18:00 day. Each hour is 4.3% of it.

Hour (New York)Random walkGoldBitcoinEther
London, 03:00–04:002.8%2.5% / 2.5%2.6% / 1.7%2.1% / 2.2%
New York AM, 10:00–11:003.1%9.2% / 8.6%5.8% / 6.6%5.8% / 5.7%
New York PM, 14:00–15:004.5%2.5% / 2.5%5.3% / 6.0%4.7% / 5.2%

The New York AM hour held gold's high on 15 of 163 days, about three times what a random walk puts in that hour, and Bitcoin's on 86 of 1,476. On no gold day did the high and the low both form inside silver bullet hours.

What this doesn't say

  • It isn't a strategy. It counts where the day's high and low formed, and not what a trade entered in a killzone would have made. Nothing here is after costs.
  • Gold's sample is short: 163 days, all in one eight-month stretch. A share of 6.1% is 10 days. Bitcoin and Ether have nine times as many days; their New York AM counts by year are below.
  • New York AM is a crowded window. It holds the 08:30 US data releases and the 09:30 US stock open. These counts don't separate the killzone from those events.
  • A tie goes to the earlier candle. When two candles reach the same high, the first one counts.

Put the killzones on your chart

ICT Levels Desk draws the four killzones these counts use as session boxes on New York time. When each box closes, its high and low become lines. A candle that trades through a line and closes back marks it swept; the first close beyond it ends the line. It also draws the midnight open and the prior day's high and low. Its full Pine v6 source is in Pine Pro Library, so you can read every rule.

ProEA's ICT Levels Desk: $49 once in Pine Pro Library (every desk), full Pine v6 source for TradingView; also in the $49 Complete Suite. Facts checked 15 November 2026.

For the windows on your own clock, ICT killzone times shows each one live, in New York time and in yours.

  • The counts:
    • definitions in docs/research/2026-11-16-killzone-stats-plan.md, committed before the first count, with one amendment after it (the random walk's share; no count changed);
    • the script, scripts/research/killzone-stats.mjs;
    • the results, data/research/killzone-stats/results.json.

Frequently asked questions

Which ICT killzone holds the day's high most often?

For its length, the New York AM killzone (08:30–11:00 New York). It is 10.9% of the 18:00–17:00 trading day, and it held the day's high on 15.3% of 163 gold days and 14.6% of 1,476 Bitcoin days, about twice the 7.5% a random walk would put there. In raw counts it is close: on Bitcoin, New York PM held as many highs (216 days each), and on Ether, Asia, which is four hours long, held a few more (217 against 209).

Do most daily highs form inside a killzone?

About half do. On gold, 57.1% of daily highs and 53.4% of lows formed outside all four killzones (Asia, London, New York AM, New York PM). On Bitcoin it is close to even, with 49.3% of highs and 48.8% of lows outside, so just over half inside. Together the four windows are about half of the 23-hour day.

How often does the London killzone hold the day's high on gold?

On 6.1% of 163 gold days (10 days), and the low on 6.7% (11 days). London (02:00–05:00 New York) is 13% of the trading day, but it sits in the middle of it, where a random walk would put the high on about 8.5% of days.

How often does the silver bullet hour hold the day's high?

The New York AM silver bullet (10:00–11:00 New York) held gold's daily high on 9.2% of days (15 of 163) and Bitcoin's on 5.8% (86 of 1,476). The hour is 4.3% of the trading day; a random walk would put the high there on about 3.1% of days.

Does this mean I should trade the New York AM killzone?

These are counts of where the day's high and low formed, not a tested strategy. They don't say what happens after a killzone, and nothing here is after costs. The New York AM window also holds the 08:30 US data releases and the 09:30 US stock open.

Continue reading

False Breakout in Trading: Two Closes to Confirm, a Wick Back Inside to Fade

A false breakout is a move beyond a level that doesn't hold: price trades through the high and the candle closes back inside. While the candle is open, nobody can tell it from a real one; the close decides. Two written rules from one TradingView indicator answer it: two closes in a row beyond the prior hour's range to confirm a break, or a wick through one side and a close back inside to fade it. Worked by hand, a film of each, and a drill where you call every close.

Oct 30 · ProEA Lab

Cover titled The false breakout, called on the close: illustrative 15-minute candles. The prior hour from 07:00 to 08:00 New York is boxed from 100.00 to 102.00. In the next hour a candle wicks through both sides (Wait), the next closes above 102.00 (1 of 2), the next trades to 102.90 and closes back inside at 101.60 (Fade), with a stop at 102.90 and a 2R objective at 99.00, and the last closes inside (Wait).

1 Hour Candle Strategy: The Previous Hour's High and Low, Read on 15-Minute Closes

The 1 hour candle strategy as a rule you can check: the previous hour's high and low, two 15-minute closes beyond one side or a wick back inside, a stop of at least 1 ATR and a 2R objective. Plus the question most guides skip, which hour counts: TradingView's 1-hour candle starts at the session's open, a UTC clock hour doesn't, and the New York window moves an hour against UTC on 1 November 2026.

Oct 30 · ProEA Lab

Cover titled 1 hour candle strategy, by the close: four illustrative 15-minute candles from 09:00 to 10:00 New York build an hour whose high is 4,116.00 and low 4,107.50, shaded as the range; beside them the same hour as one 1-hour candle. The next two 15-minute candles close above 4,116.00, marked 1 and 2, and a long plan is drawn from the second close with its stop at 4,116.00 and its 2R objective at 4,128.00.

Initial Balance vs Opening Range: The First Hour, Counted.

The initial balance (the first hour) against the 15-minute opening range, counted on 164 days of gold and 1,477 each of Bitcoin and Ether: how often each broke, on which side, how far, and where the day closed.

Oct 27 · ProEA Lab

Cover titled Initial balance vs opening range, counted: bars for gold, Bitcoin and Ether showing the share of days the 15-minute opening range was broken on both sides, 56 to 62 percent, against the first hour, 18 to 28 percent.