Historical analogs

Past weeks when the chosen indicators had moved most like they have over the last few weeks, and what an outcome series did afterwards. A search of the past by a stated distance, not a forecast.

The last 4 weeks, to 2026-10-09

Data ↓

UST 10YR +28 bpDXY +3.12%Brent +0.11%S&P 500 +2.02%US HY OAS +46 bp

Closest past weeks

Week of	Distance	UST 10YR	DXY	Brent	S&P 500	US HY OAS	JCI (IHSG) +4w	JCI (IHSG) +13w
2024-01-19	1.37	+25 bp	+1.56%	−0.64%	+1.79%	+15 bp	+1.50%	−1.94%
2024-12-27	1.45	+44 bp	+2.14%	+1.69%	−1.02%	+23 bp	+1.84%	−7.47%
2024-04-12	1.49	+19 bp	+2.52%	+5.99%	+0.12%	+9 bp	−2.72%	+0.56%
2025-10-31	1.80	−2 bp	+2.13%	+0.84%	+1.85%	+14 bp	+4.22%	+2.03%
2026-06-05	2.01	+17 bp	+2.28%	−8.10%	−0.21%	−5 bp	+5.02%	+18.62%
2024-11-01	2.12	+39 bp	+1.72%	−6.34%	−0.39%	−6 bp	−5.21%	−5.28%
2025-08-01	2.20	−12 bp	+1.55%	+2.02%	−0.66%	+33 bp	+3.88%	+8.31%
2024-06-14	2.70	−22 bp	+1.05%	−1.62%	+2.42%	+20 bp	+8.80%	+16.00%

Outcome after	n	Mean	Median	Up	Min	Max
+4w	8	+2.17%	+2.86%	75%	−5.21%	+8.80%
+13w	8	+3.85%	+1.29%	62%	−7.47%	+18.62%

"Up" = share of outcomes above zero. n in amber: fewer than 5 observations, read with care. Past outcomes describe the past; they are not a forecast.

Method

Weekly grid (Fridays; each series' last reading on or before, at most 7 days earlier). For each indicator, its change over the chosen weeks (% for prices, bp for rates), standardised by that change's mean and standard deviation over the whole grid (so the scaling uses the full history, later data included). Distance = Euclidean distance between today's standardised changes and each past week's, every indicator weighted equally. Weeks overlapping today's window are excluded and picked weeks are at least 8 weeks apart. Outcome = the outcome series' change from that week to 4 and 13 weeks later. 150 candidate weeks.