Confluence

Today’s live states, matched against every settled month since Aug 1981 that read the same — computed only from what was knowable at the time on the history side. Matched exactly, listed in full. We present the record; we do not assert predictive skill.

Today AS OF 2026-09-15 · LIVE

The four states below define "a month like today." Every month since Aug 1981 that read the same on all four is listed in the register.

today's states: live · analog months: settled month record through 2026-08

Today's other readings — live, and not used to select the months below. They appear on each register row (as of their own record months) so you can compare then with now:

Margins × valuation — the double-stretch, counted

both stretched margins +1.41σ · CAPE +1.60σ

66 of 278 quarters since 1957Q1 read like today, across 14 spells; the current spell is 3 quarters old (since 2025Q4).

statequartersspells+1y median+3y medianworst +3y
both stretched ← today6614+13.5% (N=63)+35.1% (N=61)-37.6%
one stretched9424+12.5% (N=93)+35.7% (N=87)-38.7%
neither6621+14.9% (N=66)+38.8% (N=66)-25.1%
either compressed528+18.4% (N=52)+50.7% (N=52)-24.5%

S&P total return · Shiller monthly convention · quarterly join — CAPE at each quarter's end month · the shared house ladder, deep + pale folded · counts, never probabilities · this card describes — the analog matcher does not consume it · margins AS OF 2026Q2 · CAPE AS OF 2026-06

S&P 500 — the analog months shaded

1997-032000-022003-012005-112008-102011-082014-062017-052020-042023-042026-088001,0002,0003,0004,0005,0007,000

Shaded spans are the months matching today on all four dimensions; the open run is the brighter shade. Log scale.

69 months · 30 prior episodes since Aug 1981

The table is complete; nothing is chosen for you.

episodethen — entry-month readingswhat followed — S&P total return
spanmonthsspellbft/HEAMRFS·M·IRM/IR trajEB thenCPS thenduringworst aft.+1y+3y+5ybecame
2026-04 → nowopen52expensive / expensive3 of 855 · 55 · 19stable / deteriorating+54.2%+1.41σ+11.7%same spell
2025-08 → 2025-1252expensive / expensive3 of 847 · 42 · 22deteriorating / deteriorating+33.3%+0.95σ+7.3%Disinflation
2025-05 → 2025-0511expensive / expensive4 of 844 · 64 · 37stable / stable+8.3%+0.83σ+29.1%Transitional
2025-02 → 2025-0213expensive / expensive5 of 845 · 60 · 33stable / improving-12.5%+0.92σ+15.6%same spell
2024-11 → 2024-1112expensive / expensive4 of 842 · 51 · 27stable / deteriorating+25.0%+1.02σ+15.1%Overheating
2024-02 → 2024-0767expensive / expensive3 of 840 · 59 · 30improving / stable+16.7%+0.86σ+11.2%+15.1%Transitional
2023-07 → 2023-0716expensive / expensive1 of 834 · 58 · 18stable / stable+12.5%+1.05σ-5.1%+24.8%+73.3%Disinflation
2021-12 → 2021-1218expensive / expensive1 of 868 · 84 · 4stable / deteriorating+58.3%+1.22σ-19.3%-15.0%+34.5%same spell
2021-07 → 2021-1043expensive / expensive1 of 869 · 73 · 25deteriorating / deteriorating+25.0%+1.42σ+2.6%-15.2%-15.2%+35.9%same spell
2020-01 → 2020-0111expensive / neither4 of 850 · 55 · 53stable / stable+4.2%+0.43σ-18.8%+17.9%+26.8%+97.0%same spell
2019-04 → 2019-0411expensive / expensive2 of 849 · 56 · 46stable / deteriorating+29.2%+1.07σ-7.0%-3.0%+58.9%+90.4%same spell
2018-08 → 2018-09226expensive / expensive2 of 862 · 69 · 81stable / stable+25.0%+1.07σ+1.7%-11.1%+4.8%+61.6%+65.4%same spell
2018-03 → 2018-05321expensive / expensive1 of 865 · 61 · 62deteriorating / deteriorating+62.5%+1.12σ+0.3%-3.9%+7.7%+63.0%+67.2%same spell
2018-01 → 2018-01119expensive / expensive2 of 866 · 68 · 71stable / improving+33.3%+1.12σ-6.4%-4.7%+43.9%+54.7%same spell
2017-10 → 2017-10116expensive / expensive1 of 863 · 59 · 56stable / deteriorating+50.0%+1.09σ+1.6%+11.0%+41.5%+58.9%same spell
2017-07 → 2017-08213expensive / expensive3 of 860 · 59 · 63stable / stable+33.3%+1.06σ+0.2%+1.7%+18.5%+46.2%+84.6%same spell
2017-02 → 2017-0548expensive / expensive1 of 858 · 57 · 78stable / improving+41.7%+1.18σ+3.3%+1.8%+14.9%+29.1%+84.1%same spell
2007-01 → 2007-0115neither / expensive3 of 856 · 57 · 65stable / improving+22.7%+1.42σ-44.2%-1.4%-15.4%+1.8%same spell
2004-12 → 2004-12113neither / neither0 of 476 · 70 · 75stable / improving+47.6%+2.04σ-2.4%+7.1%+30.1%+2.8%same spell
2004-04 → 2004-0415neither / neither0 of 476 · 69 · 83stable / improving+52.4%+2.25σ-3.4%+4.5%+36.1%-17.3%same spell
2004-02 → 2004-0213neither / neither0 of 473 · 63 · 59stable / deteriorating+42.9%+2.34σ-4.0%+6.7%+33.1%-22.4%same spell
2003-12 → 2003-1211neither / neither1 of 470 · 63 · 54stable / deteriorating+38.1%+1.75σ+1.9%+12.8%+38.0%-10.8%same spell
1999-11 → 2000-01314expensive / expensive0 of 258 · 73 · 89stable / improving+42.9%+0.43σ+2.7%-37.8%-5.2%-34.5%-10.7%same spell
1999-06 → 1999-0729expensive / expensive0 of 263 · 64 · 87stable / improving+38.1%+0.91σ+4.5%-32.0%+7.9%-32.0%-13.9%same spell
1999-01 → 1999-0114expensive / expensive0 of 258 · 66 · 81stable / stable+52.4%+1.04σ-13.6%+15.6%-5.2%-2.6%same spell
1998-02 → 1998-06524— / expensive0 of 270 · 71 · 84stable / stable+28.6%+1.14σ+8.8%-7.6%+21.0%+16.1%-4.4%same spell
1997-11 → 1997-12221— / expensive0 of 274 · 71 · 87stable / stable+47.6%+1.97σ+2.6%+0.2%+25.5%+43.8%0.0%same spell
1997-04 → 1997-09614— / expensive0 of 274 · 71 · 92stable / improving+52.4%+2.06σ+23.5%+0.5%+10.6%+63.0%-0.9%same spell
1996-11 → 1996-1229— / neither0 of 266 · 57 · 88deteriorating / stable+31.6%+1.83σ+1.2%+3.3%+31.8%+101.0%+65.2%same spell
1996-08 → 1996-0926— / neither0 of 271 · 68 · 86stable / deteriorating+47.4%+1.75σ+2.0%+4.1%+41.4%+104.7%+66.2%same spell
1996-03 → 1996-0311— / neither0 of 258 · 61 · 93improving / stable+26.3%+1.98σ+0.2%+25.0%+108.5%+97.6%same spell
S&P 500 median (closed episodes)+2.7% (N=14)-4.0% (N=25)+12.8% (N=29)+38.0% (N=25)+28.8% (N=22)

Returns are S&P total return · Shiller monthly convention (dividends reinvested monthly — see "The total-return basis" under Methodology). "During" runs from the episode's first month-end to its last; a one-month episode has no span at month-end anchoring and shows "—" (its median counts multi-month episodes only). "Worst aft." is the deepest monthly total-return level within the 36 months following the episode's last month, from that month's level; +1y/+3y/+5y are measured from the episode's last month. An episode is the matching run itself — not the surrounding regime spell. Context columns show each episode's entry-month readings (MR denominator = relationships with data then; "—" where a series' history doesn't reach). Historical record, not a forecast.

The shape of what followed

one dot = one closed episode

-44%0%+109%
Worst aft.
+1y
+3y
+5y

worst aft.: 16 of 25 saw a decline · +1y: 6 of 29 ended lower · +3y: 4 of 25 · +5y: 8 of 22

Methodology →