The lesson of the week · Sunday 18 Oct 2026

The theme. US data hotter than expected, technology down, oil above 100 because of Hormuz; and the Fed taking the urgency off the October hike.

The method. 2 long orders with a limit entry, dollar/yen at 157.680 and Nasdaq at 28,800, valid until 23 Oct 2026.

Confidence LOW for both (2 out of 6: a score from fixed rules, not a probability). It is not a forecast: it is the risk premium.

The past week (12-16 Oct). Technology down. S&P 500 −1.9%, Nasdaq −4.5%, Dow +0.7%, Nikkei −5.4%. The SOX, the semiconductor index, closed on Friday its eighth losing session in a row, the longest streak since 2008; CoreWeave −23.7% on debt fears.

• The why: chips supply the race to artificial intelligence, which is partly financed with debt. If the market fears for that debt, the most exposed stocks fall first, then the indices that weigh them most. The Dow, where technology weighs less, rose.
• The example: the broker's Nasdaq 100 from its 6 Oct high, 31,365.74, to Friday's 29,180.10: -6.6% in eight sessions.

The hot US data, meaning above expectations: CPI (consumer inflation) 3.7% versus 3.6%; PPI (producer prices) +1.3% on the month versus 0.5-0.6%; retail sales +0.8% versus 0.3%; industrial production +0.7% versus 0.4%.

• The why: a tenth of CPI looks small, but the market was pricing the Fed on the expected number. Every extra tenth makes a rate hike more likely: usually government bond yields and the dollar rise, and gold falls, since it pays no interest.

The Fed. On 16 Oct, in Bangkok, Chair Warsh said “I don't get ahead of meetings”: the priced-in October hike fell from about 28% to 16%. “Priced in” means already inside prices, as a probability in rate futures. At least one hike by 9 Dec, though, is priced at 88%: the market sees it in December more than in October.

Oil. WTI +10.4% to 101.41, Brent +4.9% to 109.86. 11.6 million barrels a day are passing through Hormuz against a normal 17.2, about a third less: less supply, a higher price, and more inflation for the Fed to fight.

Rates, dollar, metals. US ten-year at 5.09% (−14 basis points). The dollar down for the first week since early September. Gold flat at 4,165, silver −5.5/−6%.

The weekend, the briefing writes, was not quiet: attacks on ships in Hormuz at a record level, missiles and drones intercepted over Riyadh, US plans for a strike with the MOAB bomb according to CBS. And an opening: Iran says it would reopen Hormuz if the US accepts its conditions. On Monday markets may open with a jump (gap): what moves them may be oil.

The week ahead (19-23 Oct). The Fed is silent until 29 Oct (blackout period): data and quarterly results move markets.

• Monday at 04:00 Italian time, the GDP of China (expected 4.4%), the largest purchaser of many raw materials.
• Wednesday 21 Oct the British CPI at 08:00 (expected 3.4%), which weighs on the BoE, and above all Meta, Tesla and IBM: within the validity of the limit on the Nasdaq.
• Friday the CPI of Japan at 01:30 (core expected 2.1%, previous 1.7%), which weighs on the BoJ; then at 15:45 the US flash PMI (above 50 activity is growing).

What is already priced in. Fed (3.75-4.00%) on 28 Oct: hold 84%, hike 16%. BoJ (1.25%) on 30 Oct: hike 10-12%, 88-90% by December. BoE (3.75%) on 5 Nov: hike at 81%. A priced-in event moves little, the surprise moves: if on 28 Oct the Fed holds, it is the 84% coming true; if it hikes, it is the 16% surprise.

If… then…

• Dollar/yen: if on Friday the Japanese CPI comes out above 2.1%, the BoJ hike gets closer, usually the yen strengthens and the price falls toward the limit at 157.680; if it comes out below, the carry stays as it is. Further ahead, a Fed hike on 28 Oct widens the rate gap, a surprise BoJ hike on 30 Oct narrows it.
• Nasdaq: if on 21 Oct the quarterly results disappoint, the decline may reach 28,800 and fill the limit; if they reassure, the limit may expire on 23 Oct without being filled: no loss, an opportunity not taken.

Why the weekly exists

• The average move grows with time, the noise with the square root of time: the first is worth 2% of the second at 3 hours, 5% at 2 sessions, 11% at 10, 15% at 20, 26% at 60. From 2 to 10 sessions: time times five, a ratio a little more than double, that is the square root of five.
• With the stop at 2 ATR (two average days) the long beats the cost at every horizon on 8 instruments out of 14; the short always loses, because the drift (the rise over the years) is against it and the cost adds up: that is why the weekly is long only.
• The example: the DAX over 20 sessions rises on average by +0.282 ATR, but spread and swap cost 0.325. Edge -0.042: out, even though over the last 3 years its average is +1.065. Rising is not enough: it has to rise more than it costs to hold.
• It is the risk premium, plus the carry where the broker pays the swap. It is not a forecast: it works while markets rise, and in a three-year decline it loses.

The field and the swap (reading of 16 Oct 2026). 8 out of 14 admitted (positive edge after spread and swap, and positive move over the last 3 years): S&P 500, Nasdaq, Dow Jones, Nikkei, gold, silver, WTI oil, dollar/yen. On the long side the swap, the cost or income of each night, is paid almost everywhere (Nasdaq -0.0109 ATR a night) and is earned only on WTI oil (+0.0066) and dollar/yen (+0.0083). Since 2 Oct only oil has changed (long swap changed by 53%): 50% of its edge is today's swap, which can turn. On dollar/yen the carry is worth +0.224 of the +0.515 edge.

Already in play. The gold order of 2 Oct (limit 4,105, stop 3,940, target 4,375, confidence LOW 0 out of 6) touched the entry on 6 Oct (MetaTrader bars); low afterwards 4,066.31 on 7 Oct; on Friday 4,163.21, $58 in favour with the stop $165 from the entry. Time exit on 2 Nov 2026. The record had written it: in 2026 the thesis is in the red.

Brake and track record. 5 weekly proposals (3 in progress, 2 not filled), sum of net R +0.00 over 2 measured; the brake triggers at -12 R and has not triggered.

The orders, recorded on Friday 16 Oct at 22:59 (server time). These are the method's orders, recorded beforehand and judged afterwards in the public record: they are not financial advice.

Summary

• Dollar/yen · long · limit 157.680 · stop 155.550 · target 160.450 · until 23 Oct 2026
• Nasdaq · long · limit 28,800 · stop 27,780 · target 30,690 · until 23 Oct 2026