NY AM, week commencing Monday 21 September 2026

At a glance

Three scores, because the indicator makes three separable claims.

The four share columns are the indicator's own figures, read off the status table at the session open, and they sum to 100 on every row. Gold marks the branch the tables rated highest. The London column is a whole number because the study rounds it, half up. The exact ranges are 166.50, 184.25, 145.75, 261.00 and 157.25 points, so the two opening distances quoted in each session below sum to the exact range and can sit up to half a point away from the figure in this column: Monday's 47.25 under and 119.25 above make 166.50, shown here as 167. London printed above its own average range on Tuesday and Thursday and below it on the other three days. Wednesday's prior sweep reads Swept both, the first time that label has appeared in the forty published sessions. No red folder USD release fell at any time during the week, so no result on this page sits inside a news window. Times are the 5-minute bar in which the level was crossed, and every time on this page is read off the 5-minute bar data rather than off the pictures.

Four modal branches from five, and why that is the number to distrust

The branch the tables rated highest was the one that happened on Monday, Tuesday, Wednesday and Thursday. No published week had managed more than three, and the record on this claim now stands at twenty-two from forty. Four things are worth saying before anyone reads a skill estimate into it.

The four-way branch is the broadest claim the indicator makes and the one with the least behind each call. Its favourite this week was rated 53, 37, 52, 37 and 52 percent. Two of the four that landed were 37% calls, which is to say the table gave the outcome a little over one chance in three and it happened, twice, on the same four numbers. If each of the five branches landed at exactly the rate printed beside it, four or more from five has a probability of about 14%, one week in seven. Five consecutive mornings on one instrument are not independent draws, and dependence between them makes a run like this more likely than that figure says, so 14% is a floor rather than an estimate. It is a good week. It is one week.

Wednesday's first sweep was decided by three and a half points. The table put 85% on the low going first. Price opened 3.50 points above the London low and the opening bar traded seventeen points through it, so the call was correct before the open snip was taken. The 85% was not tested by that: every 07:30 bar this week spanned at least 12 points, and a level three and a half points from the open sits inside such a bar. What the table did say that was tested is the branch: low only at 52%, which needed the London high, 142.25 points above the open, to be left alone. It was, by 81.75 points, at the 09:30 high, and from there price fell 372.75 points in eleven bars, 308.75 through the level against a median penetration of 97.

Tuesday and Thursday drew the same table and answered it the same way, and that is not evidence the table is right. Both sessions read 37 / 35 / 18 / 10 on 215 sessions with the first sweep at 53 / 47, the thinnest call the record has seen. Both opened inside the range, one above EQ and one below. Both went down first, Tuesday to 76.75 points above the London low and Thursday to 49.75, and both then took the high, at 09:30 and 10:25. Seven draws of this bin now read four high onlys, two low onlys and one both sides against a stated 37 / 35 / 18. Two of the seven, on 25 and 27 August, went to the low on the same 53% call. The table did not know on Tuesday which of those it would be, and neither did anyone reading it.

Five first sweeps from five, for the third time. The calls were rated 74, 53, 85, 53 and 88, a mean of 71%, and all five landed. Under the same independence assumption that is about a 16% event, one week in six, and it has now happened in three of eight weeks, which is either a run or a sign the table's thin calls are better than they print. Forty sessions cannot say which. The running figure is thirty-one from forty, and the indicator's own hit-rate row sat at 51 to 52% against a 56% expectation all week, on 263 to 278 setups loaded, which is where it sat last week too.

The five NY AM sessions

Left, the session opening at 07:30. Right, the same chart at 11:30 with the outcome latched.

Monday 21 September

Through the high at 08:05, back inside for an hour and a half, then 316 points through it

First sweep ✓ · high only, the 53% favourite ✓

The fourth draw of 53 / 20 / 24 / 3 from 479 sessions, the table that drew on Friday 4, Friday 11 and Thursday 17 September. London had run 167 points, below its own average, and price opened at 30225.75, 47.25 points under the London high and 119.25 above the low, above EQ. First sweep 74% to the high.

The high went at 08:05, in the eighth bar, and only just: that bar reached 30277.25, 4.25 points through, and the next bar was back inside the range. Price then spent an hour and a half inside it, down to a session low of 30202.25 on the 09:20 bar, which is 23.50 points under the open and 95.75 points above the London low. From that low it rallied 386.75 points in under two hours, to 30589.00 on the 11:15 bar, 316.00 points through the London high against a median up penetration of 82, and closed at 30584.50.

First sweep correct at 74% and the modal branch landed, high only at 53%, with the 77% high taken and the 44% low not. The take itself was worth four and a quarter points at 08:05; the three hundred that followed came two hours later and the table makes no claim about them. This bin has now drawn four times: both sides twice, on 4 and 11 September, and high only twice, on 17 and 21 September.

• London range167 pt, below average
• Prior sweepSwept high
• Open vs London EQAbove EQ
• Four-way53 / 20 / 24 / 3
• First sweep▲ 74% · ▼ 26%
• P(high taken)77% → taken 08:05
• P(low taken)44% → not taken, by 95.75 pt
• Session high30589.00, 316.00 pt through
• Session low30202.25, 95.75 pt short
• Samplen 479, High

Tuesday 22 September

Down 31.75 points first, then a 105-point bar through the high at 09:30

First sweep ✓ · high only, the 37% favourite ✓

The sixth draw of 37 / 35 / 18 / 10 from 215 sessions, one of two Medium bins in the record, the other n 173, and the one Thursday would draw again two days later. London had run 184 points, above its own average, and price opened at 30779.50, 75.75 points under the London high and 108.50 above the low, just above EQ. First sweep 53% to the high, the thinnest call the record has seen.

Price went down first. The session low printed on the 08:05 bar at 30747.75, 31.75 points under the open and 76.75 points above the London low, and for the next hour and a half nothing traded outside the range. The 09:30 bar then ran from 30763.25 to 30868.25, 105.00 points in five minutes, and crossed the London high on the way. The session high came on the 10:30 bar at 30997.75, 142.50 points through against a median up penetration of 88, and the close at 30928.25 was 73.00 points above the level.

First sweep correct at 53% and the modal branch landed, high only at 37%, the 55% high taken and the 53% low not. Six draws of this bin now read three high onlys, two low onlys, on 25 and 27 August, and one both sides; its 10% inside branch has not appeared. On the numbers alone a 53% call is a coin with a thumb on it, and this page should not pretend the coin knew.

• London range184 pt, above average
• Prior sweepSwept low
• Open vs London EQAbove EQ
• Four-way37 / 35 / 18 / 10
• First sweep▲ 53% · ▼ 47%
• P(high taken)55% → taken 09:30
• P(low taken)53% → not taken, by 76.75 pt
• Session high30997.75, 142.50 pt through
• Session low30747.75, 76.75 pt short
• Samplen 215, Medium

Wednesday 23 September

Opened 3.50 points above the low and took it on the opening bar, climbed 81 points from the 07:30 bar low, then 372.75 points down in 55 minutes from the 09:30 high

First sweep ✓ · low only, the 52% favourite ✓

The eighth draw of 10 / 52 / 38 / 0 from 840 sessions, the largest bin in the record. London had run 146 points and its sweep column read Swept both, the first time that label has appeared in forty published sessions. Price opened at 30952.50, below EQ and 3.50 points above the London low, with 85% on the low going first.

It went on the opening bar. The 07:30 bar traded down to 30932.00, seventeen points through the level, and the open snip already carries the take and both ticks. Price then climbed back inside the range for two hours, to a session high of 31013.00 on the 09:30 bar, which is 81.75 points under the London high and as close as the 38% both-sides branch came. That same 09:30 bar traded 87.00 points top to bottom, and over the next eleven bars price fell 372.75 points to 30640.25 on the 10:25 bar, 308.75 points through the London low against a median down penetration of 97. The last hour recovered to 30827.75 and the session closed at 30773.00, 176.00 points below the low.

First sweep correct at 85%, low only at 52% landed, the 90% low taken and the 48% high not. A take on the opening bar is not new, it happened on 4 August to the high and on 20 August to the low, but a 3.50-point opening distance leaves the first sweep with almost nothing to call. Eight draws of this bin have produced six low onlys, one both sides and one inside, against a stated 52 / 38 / 0, and its 10% high-only branch has still not appeared.

• London range146 pt, below average
• Prior sweepSwept both
• Open vs London EQBelow EQ
• Four-way10 / 52 / 38 / 0
• First sweep▲ 15% · ▼ 85%
• P(high taken)48% → not taken, by 81.75 pt
• P(low taken)90% → taken 07:30, opening bar
• Session high31013.00, 81.75 pt short
• Session low30640.25, 308.75 pt through
• Samplen 840, High

Thursday 24 September

Tuesday's table again: down 50 points first, nine short of the high at 09:45, through it at 10:25

First sweep ✓ · high only, the 37% favourite ✓

The seventh draw of 37 / 35 / 18 / 10 from 215 sessions, the same table as Tuesday, served this time to a below-EQ open after a London of 261 points, the widest of the week. Price opened at 30470.50, 161.25 points under the London high and 99.75 above the low. First sweep 53% to the high, as on Tuesday.

As on Tuesday, price went down first: the session low printed on the 08:35 bar at 30420.50, 50.00 points under the open and 49.75 points above the London low. The climb from there came 9.00 points short of the London high on the 09:45 bar, gave 91.50 points back to the 09:55 bar, and crossed the level on the 10:25 bar, nearly three hours into the window. The session high of 30644.25 on the 10:35 bar was 12.50 points through against a median up penetration of 89, and the last fifty minutes gave 140.25 of it back: the 11:10 bar traded to 30504.00 and the session closed at 30514.00, 117.75 points back inside the range.

First sweep correct at 53%, high only at 37% landed, the 55% high taken and the 53% low not. Tuesday and Thursday are the cleanest pair the record has: the same four numbers, the same 53 / 47 call, opens on opposite sides of EQ, and the same outcome by the same route. That is two draws, and this bin's seven now read four high onlys, two low onlys and one both sides against a stated 37 / 35 / 18.

• London range261 pt, above average
• Prior sweepSwept low
• Open vs London EQBelow EQ
• Four-way37 / 35 / 18 / 10
• First sweep▲ 53% · ▼ 47%
• P(high taken)55% → taken 10:25
• P(low taken)53% → not taken, by 49.75 pt
• Session high30644.25, 12.50 pt through
• Session low30420.50, 49.75 pt short
• Samplen 215, Medium

Friday 25 September

The high in the second bar, 3.75 points through, then 314.50 points down to the low

First sweep ✓ · both sides, the 38% branch

The fourth draw of 52 / 8 / 38 / 2 from 766 sessions. Price opened at 30977.00, 17.75 points under the London high and 139.50 above the low, with 88% on the high going first, the strongest call of the week.

The high went at 07:35, in the second bar, and that bar was also the session high: 30998.50, 3.75 points through against a median up penetration of 92. Price then fell for the rest of the morning, crossing the London low on the 09:25 bar and running to 30684.00 on the 10:15 bar, 153.50 points through against a median down penetration of 100, a fall of 314.50 points from the session high. The recovery reached 30871.25 by 11:05 and the session closed at 30813.75, 23.75 points below the London low.

First sweep correct at 88% and both marginals taken, the 90% high and the 46% low. The branch was both sides at 38% against a 52% high-only favourite, the week's only modal miss. Four draws of this bin have now produced one high only, on 4 August, and three both sides, against a table that says 52 and 38. Four draws cannot tell you which of those numbers is off, or whether either is.

• London range157 pt, below average
• Prior sweepSwept high
• Open vs London EQAbove EQ
• Four-way52 / 8 / 38 / 2
• First sweep▲ 88% · ▼ 12%
• P(high taken)90% → taken 07:35
• P(low taken)46% → taken 09:25
• Session high30998.50, 3.75 pt through
• Session low30684.00, 153.50 pt through
• Samplen 766, High

The same forecast, more than once

Across the forty sessions published so far, eight context bins have each come up more than once, and thirty-nine of the forty sessions sit in one of them. The only session still on its own is Monday 10 August. Same context bin, same lookup, same numbers to the percentage point, on days weeks apart with different London ranges and different price levels. Every one of this week's five forecasts was a repeat, for the third week running, and two of them were the same forecast twice.

Two things follow, and they matter more than this week's score.

The forecast is a lookup, not a fit. Nothing about Thursday was used to produce Thursday's numbers. They are the numbers Tuesday got two days earlier, and the numbers five sessions before that got, because all seven landed in the same measured bin, on London ranges from 171 to 261 points and at opens more than eighteen hundred points apart (28972.00 on 14 September, 30779.50 on Tuesday). Four of the seven went high only, two low only and one both sides.

Identical forecasts resolving differently is the system working, not failing. Every one of the eight repeated bins has now produced more than one outcome. The largest, 10 / 52 / 38 / 0 from 840 sessions, has drawn eight times: six low onlys, one both sides and one inside, against a stated 52 / 38 / 0, and its 10% high-only branch has appeared in none of them. The 766 bin has drawn four times and gone both sides on three of them against a 38% share for that branch and a 52% favourite of high only. Eight draws cannot tell you whether 52 is the right number, and neither can forty sessions; what they can show is whether the table keeps returning the same answer for the same context bin, which it does.

What this week actually shows

Five things, from a week that scored well on two of its three claims and no better than usual on the third.

1. Four modal branches from five, the most of any published week, and the record moves to twenty-two from forty. The eight weeks now read three, two, two, three, three, two, three and four from five. Two of this week's four were 37% calls. Two 37% branches both landing is about a one-in-seven event for a calibrated table, and it is not evidence that the number should be higher.
2. Five first sweeps from five, the third such week, and the record moves to thirty-one from forty. The mean stated confidence across the five calls was 71%, from Tuesday's and Thursday's 53% to Friday's 88%. Wednesday's call was settled by the opening bar from 3.50 points away and says almost nothing about the table.
3. The high was taken on four of the five days, the low on two. Three sessions finished high only, one low only and one both sides. The four levels not taken were rated 44, 53, 48 and 53 percent, and the six taken were rated 77, 55, 90, 55, 90 and 46. Every level rated above 53% was taken and every level rated at or below it was not, except Friday's 46% low, which went.
4. Six of ten levels taken, against a mean stated 61%. A realised 60%, the same score as the two weeks before it. Across the record fifty-five of eighty levels have been taken against a mean stated 66%, a realised 69%. Eighty level events cannot close that gap in either direction.
5. Forty sessions settle nothing, and the indicator's own row is still behind. The realised versus expected row on the status table read 51 to 52% against 56% expected across the week, on 263 to 278 setups loaded, which is where it sat the week before. The record has a shape, not a verdict.