Atlas

Every exhibit the math has drawn, in one place. Each picture below is computed from the corpus — canonical days averaged from hundreds of real sessions, recurring shapes mined from years of tape, physics numbers measured over 180,000+ labeled sessions — never sketched by hand. Everything describes finished history; nothing forecasts. The full story of how each one is built lives in Shape Math.

The canonical days

The archetype of each vetted pattern family — the one curve whose warped distance to every dominant member is smallest, event-aligned so the defining moment survives the averaging. The caption on each card states the whole funnel: how many days carried the tag, how many were genuinely dominated by the pattern, how many the canon averages.

The opening reversal (up)
Sell-off from the open into a reversal extreme that holds all session, then a grind higher.
the aligned average of the 116 closest of 230 dominant members — computed, not drawn · dashed line = the alignment anchor (the family's defining event)
the three most textbook real examples
2020-03-30
2021-11-24
2025-01-28
The opening reversal (down)
Pop from the open into a reversal extreme that holds all session, then a slide.
the aligned average of the 77 closest of 153 dominant members — computed, not drawn · dashed line = the alignment anchor (the family's defining event)
the three most textbook real examples
2020-04-03
2021-03-04
2024-04-12
The buy-the-close climax
A run that accelerates INTO the bell and finishes at the session high — the event anchor is the close itself.
the aligned average of the 23 closest of 45 dominant members — computed, not drawn · dashed line = the alignment anchor (the family's defining event)
the three most textbook real examples
2022-07-05
2023-03-24
2020-03-04
Trend from the open (up)
One-directional from the first bar — membership from the day-type labels, aligned on the open itself.
the aligned average of the 99 closest of 196 dominant members — computed, not drawn · dashed line = the alignment anchor (the family's defining event)
the three most textbook real examples
2020-05-18
2025-07-10
2020-04-29
Trend from the open (down)
The bear mirror, from the day-type labels.
the aligned average of the 92 closest of 183 dominant members — computed, not drawn · dashed line = the alignment anchor (the family's defining event)
the three most textbook real examples
2025-04-23
2023-05-19
2025-09-24

The shapes the market keeps drawing

Matrix-profile motifs: for six tickers, the two-hour windows closer to each other than anything else in years of tape, with measured occurrence counts — and the discords, the stretches whose nearest neighbor is farthest: the weirdest tape in the archive.

SPY — the 2-hour shapes it keeps drawing
every 24-bar window of 1,668 sessions self-joined; windows spanning the overnight seam excluded · occurrences measured within 1.25× the pair distance
motif #1 · printed ×3 · d=0.531
2022-04-28 · 11:40 ET
2023-04-17 · 13:55 ET
motif #2 · printed ×2 · d=0.551
2021-09-01 · 11:25 ET
2026-04-15 · 13:25 ET
motif #3 · printed ×2 · d=0.557
2021-09-10 · 14:00 ET
2026-02-12 · 9:30 ET
motif #4 · printed ×3 · d=0.598
2023-12-20 · 13:10 ET
2024-04-04 · 12:25 ET
the weirdest 2 hours in the archive (discords)
2021-09-17 · 13:55 ET · d=4.541
2020-08-10 · 12:40 ET · d=4.523
2026-02-19 · 9:40 ET · d=4.403
QQQ — the 2-hour shapes it keeps drawing
every 24-bar window of 1,682 sessions self-joined; windows spanning the overnight seam excluded · occurrences measured within 1.25× the pair distance
motif #1 · printed ×2 · d=0.542
2022-04-28 · 12:00 ET
2025-05-15 · 10:30 ET
motif #2 · printed ×4 · d=0.55
2023-04-17 · 14:00 ET
2023-07-07 · 11:15 ET
motif #3 · printed ×2 · d=0.56
2022-01-10 · 12:50 ET
2024-01-18 · 13:10 ET
motif #4 · printed ×3 · d=0.592
2022-09-02 · 11:20 ET
2022-10-26 · 11:45 ET
the weirdest 2 hours in the archive (discords)
2022-07-27 · 12:35 ET · d=4.43
2024-09-23 · 14:00 ET · d=4.328
2023-04-19 · 12:40 ET · d=4.278
IWM — the 2-hour shapes it keeps drawing
every 24-bar window of 1,682 sessions self-joined; windows spanning the overnight seam excluded · occurrences measured within 1.25× the pair distance
motif #1 · printed ×2 · d=0.543
2022-07-29 · 13:10 ET
2024-02-15 · 12:40 ET
motif #2 · printed ×2 · d=0.57
2023-10-09 · 13:00 ET
2025-06-18 · 9:35 ET
motif #3 · printed ×2 · d=0.572
2021-09-10 · 14:00 ET
2025-02-21 · 11:15 ET
motif #4 · printed ×6 · d=0.59
2022-04-28 · 11:40 ET
2023-02-15 · 12:40 ET
the weirdest 2 hours in the archive (discords)
2021-10-04 · 13:15 ET · d=4.766
2022-11-21 · 10:25 ET · d=4.416
2026-06-25 · 12:35 ET · d=4.397
NVDA — the 2-hour shapes it keeps drawing
every 24-bar window of 1,683 sessions self-joined; windows spanning the overnight seam excluded · occurrences measured within 1.25× the pair distance
motif #1 · printed ×2 · d=0.576
2020-07-30 · 10:50 ET
2024-09-24 · 10:50 ET
motif #2 · printed ×2 · d=0.587
2019-12-16 · 13:15 ET
2022-09-01 · 9:45 ET
motif #3 · printed ×3 · d=0.612
2022-09-13 · 14:00 ET
2022-09-20 · 11:45 ET
motif #4 · printed ×2 · d=0.626
2020-01-28 · 13:30 ET
2023-01-20 · 13:15 ET
the weirdest 2 hours in the archive (discords)
2019-11-07 · 10:35 ET · d=4.324
2025-07-08 · 9:45 ET · d=4.232
2025-01-03 · 13:50 ET · d=4.212
TSLA — the 2-hour shapes it keeps drawing
every 24-bar window of 1,688 sessions self-joined; windows spanning the overnight seam excluded · occurrences measured within 1.25× the pair distance
motif #1 · printed ×2 · d=0.503
2023-12-20 · 13:40 ET
2025-04-28 · 10:00 ET
motif #2 · printed ×3 · d=0.557
2024-12-24 · 11:50 ET
2026-04-15 · 10:20 ET
motif #3 · printed ×2 · d=0.582
2025-12-12 · 11:35 ET
2026-04-27 · 11:30 ET
motif #4 · printed ×4 · d=0.587
2020-01-08 · 10:40 ET
2022-12-19 · 11:35 ET
the weirdest 2 hours in the archive (discords)
2025-01-08 · 10:05 ET · d=4.535
2022-07-26 · 13:55 ET · d=4.353
2022-03-03 · 11:20 ET · d=4.325
BTCUSD — the 2-hour shapes it keeps drawing
every 24-bar window of 2,373 sessions self-joined; windows spanning the overnight seam excluded · occurrences measured within 1.25× the pair distance · mined from 2020+ (earlier feed data fails bar-quality checks)
motif #1 · printed ×2 · d=0.418
2022-04-08 · 12:05 ET
2023-07-20 · 9:30 ET
motif #2 · printed ×3 · d=0.56
2021-10-13 · 11:40 ET
2021-11-20 · 12:20 ET
motif #3 · printed ×2 · d=0.572
2022-09-27 · 10:35 ET
2023-04-26 · 14:00 ET
motif #4 · printed ×2 · d=0.584
2021-01-24 · 12:35 ET
2022-09-06 · 11:30 ET
the weirdest 2 hours in the archive (discords)
2023-07-26 · 12:10 ET · d=4.374
2020-02-04 · 13:50 ET · d=4.34
2020-01-22 · 11:55 ET · d=4.279

The snippet that gives the day away

Learned shapelets: for each day type, the short stretch of tape that best splits that type from everything else — found by information gain, judged on held-out sessions against shuffled-label chance. The receipts on each card are out-of-sample; where a type is marked by NOT printing the snippet, the card says so.

TD+ bull trend day the anti-tell two hours
from SPY 2023-12-13 at 11:30 ET · found by information gain, judged on held-out sessions with the training threshold frozen
printed by only 48% of held-out TD+ days vs 90% of everything else — this type is marked by not drawing it · gain 0.13 vs chance 0.00
TD- bear trend day the anti-tell two hours
from SPY 2019-12-30 at 10:30 ET · found by information gain, judged on held-out sessions with the training threshold frozen
printed by only 33% of held-out TD- days vs 81% of everything else — this type is marked by not drawing it · gain 0.14 vs chance 0.00
TFO+ bull trend-from-the-open the telltale two hours
from SPY 2020-05-20 at 9:30 ET · found by information gain, judged on held-out sessions with the training threshold frozen
printed by 93% of held-out TFO+ days vs 73% of the rest · gain 0.03 vs chance 0.00
TFO- bear trend-from-the-open the telltale two hours
from QQQ 2026-06-10 at 10:30 ET · found by information gain, judged on held-out sessions with the training threshold frozen
printed by 96% of held-out TFO- days vs 60% of the rest · gain 0.09 vs chance 0.00
TR trading range day the telltale two hours
from QQQ 2023-02-06 at 11:30 ET · found by information gain, judged on held-out sessions with the training threshold frozen
printed by 96% of held-out TR days vs 84% of the rest · gain 0.02 vs chance 0.00
TTR+ bull trending trading range the telltale two hours
from IWM 2023-11-27 at 12:30 ET · found by information gain, judged on held-out sessions with the training threshold frozen
printed by 100% of held-out TTR+ days vs 90% of the rest · gain 0.02 vs chance 0.01
TTR- bear trending trading range the telltale two hours
from QQQ 2020-12-18 at 9:30 ET · found by information gain, judged on held-out sessions with the training threshold frozen
printed by 89% of held-out TTR- days vs 67% of the rest · gain 0.03 vs chance 0.00

Every gain above beats its shuffled-label ceiling — the signals are real — but they are small against the 0.81-bit maximum a perfect separator would score: one snippet tells you a little about a day type, never the whole story. The honest reading of the pair of trend-day cards: a trend day is largely defined by the two-sided coil it refuses to print.

The confusable pairs

The sharper question: what separates the day types a trader actually mixes up — trend day or trend-from-the-open, range or trending range. Sessions restricted to the two types, candidates harvested from both sides, same out-of-sample discipline.

TD+ vs TFO+ — what separates them
from QQQ 2021-08-24 at 10:30 ET · sessions restricted to the two types, candidates from both sides
the separator lives in TFO+ tape: printed by 70% of held-out TFO+ days vs 49% of TD+ days · gain 0.03 of 1 bit vs chance 0.00
TD- vs TFO- — what separates them
from IWM 2023-05-08 at 11:30 ET · sessions restricted to the two types, candidates from both sides
the separator lives in TFO- tape: printed by 97% of held-out TFO- days vs 83% of TD- days · gain 0.04 of 1 bit vs chance 0.01
TR vs TTR+ — what separates them
from QQQ 2024-02-15 at 11:30 ET · sessions restricted to the two types, candidates from both sides
the separator lives in TTR+ tape: printed by 62% of held-out TTR+ days vs 36% of TR days · gain 0.05 of 1 bit vs chance 0.00
TR vs TTR- — what separates them
from QQQ 2024-01-05 at 10:30 ET · sessions restricted to the two types, candidates from both sides
the separator lives in TTR- tape: printed by 55% of held-out TTR- days vs 36% of TR days · gain 0.03 of 1 bit vs chance 0.00

These are the discriminations a trader actually argues about — trend day or trend-from-the-open, range or trending range. All four separators are real out of sample but small (0.03–0.05 of the 1-bit ceiling): the confusable pairs are confusable for a reason. A consistent quirk worth knowing: every winning snippet prints MORE on the milder type — the sharper day is again marked by what it does not draw.

The order of events

Path-signature fingerprints: order-aware coordinates — was the day front- or back-loaded, did churn build before the move or chase it — that a correlation cannot see. Reverse a day and every sign flips.

day typesessionsnet move (med, day-sigmas)front-loading (med, IQR)churn-first (med, IQR)volume-first (med)
TD+ bull trend day1,1243.400.34 (-0.1–0.8)-0.40 (-2.0–1.1)-1.46
TD- bear trend day823-3.390.14 (-0.5–0.7)-0.31 (-1.9–1.5)-1.33
TFO+ bull trend-from-the-open6502.250.69 (-0.0–1.2)-1.71 (-3.8–1.0)-2.28
TFO- bear trend-from-the-open604-2.200.72 (0.1–1.3)-2.41 (-4.4–0.1)-2.92
TR trading range day753-0.190.51 (-0.1–1.1)-1.79 (-3.6–0.1)-1.94
TTR+ bull trending trading range5631.29-0.55 (-1.1–0.3)2.56 (-0.4–4.5)2.21
TTR- bear trending trading range515-1.14-0.45 (-1.1–0.9)1.57 (-3.3–3.9)1.19

Level-2 path-signature coordinates (Lévy areas) over 5,032 labeled sessions (SPY/QQQ/IWM), oriented per session by the day’s own direction. Front-loading > 0 means the day’s move came early; churn-first > 0 means two-sided churn built up before the move, < 0 means price ran ahead of the churn. The clean read: trend-from-the-open days are the most front-loaded and their price leads the churn; trending trading ranges are their mirror image — back-loaded, churn first, the trend leg emerging from the range. These are order-aware numbers a correlation cannot see — reversing a day flips the sign. TR rows are oriented by a near-zero net move — read their magnitudes, not their signs. All of it describes finished sessions; none of it forecasts.

How many regimes a day runs through

Bayesian online change-point detection: at every bar, the calibrated probability that the tape just changed character — no hindsight. The census below counts how many return-regimes each day type actually runs through, on returns deflated by the everyday loudness curve.

day typesessionsone regime all dayone breaktwo or moremedian first break
TD+ bull trend day1,12472%25%4%13:15 ET
TD- bear trend day82365%33%2%13:30 ET
TFO+ bull trend-from-the-open65072%25%3%13:25 ET
TFO- bear trend-from-the-open60470%28%2%12:55 ET
TR trading range day75361%35%4%13:20 ET
TTR+ bull trending trading range56364%30%6%13:20 ET
TTR- bear trending trading range51558%37%5%13:30 ET

Bayesian online change-point detection over 5,032 labeled sessions (SPY/QQQ/IWM), on returns deflated by the time-of-day loudness curve (the open prints ~2.2× the bar-to-bar movement of the 14:00 lull) — without that deflation the detector rediscovers the close’s volatility ramp every single day and calls it a regime. The gradient that remains is the day types’ own: trend days mostly hold one return-regime all session; trending ranges break the most. Two honest caveats. Early breaks are structurally harder to detect than mid-session ones (the posterior needs tape to establish a regime before it can reject one), so first-break times skew toward midday. And the breaks are measured INDEPENDENT of the FLUSS shape seams — where both fire, the median gap (16.5 bars) matches random placement (17): a return-regime change and a shape change are different events, and neither instrument substitutes for the other. Describes finished sessions; forecasts nothing.

The weirdness meter

Compression novelty: how many bytes a day costs on top of the archive before it. The hall of the days history could not compress — and, beside it, the null this instrument published about everything it cannot see.

The days the archive could not compress
novelty = the share of a day’s bytes the previous 300 sessions failed to save — pattern-new tape costs full price
SPY2022-03-02TD+0.864
SPY2025-04-23TFO-0.844
SPY2021-09-17TD-0.841
SPY2022-11-15TR0.841
SPY2022-12-29TD+0.841
SPY2023-12-14TR0.841
SPY2024-07-24TD-0.841
QQQ2021-03-08TD-0.841
Face validity, unprompted: the top day is the week Russia invaded Ukraine, and 2021-09-17 — the quad-witching session — is also the matrix profile’s number-one SPY discord. Two unrelated instruments crowning the same tape.
And the null, published
The same instrument sees no day-type structure at all. Median novelty is identical across every one of the seven day types (0.75). Pairwise, two days of the same type compress together no better than two days of different types (0.689 vs 0.689) — and a two-channel variant (returns + bar ranges, twice the material) lands on the same null. At 77 symbols a session, compression distance is a coarse instrument: it can rank the extremes, and that is all it can do here. The day types live in geometry the deflate dictionary does not speak — which is precisely why the DTW engine, and not a compressor, runs the twin search. Measured on 4,132 sessions (SPY/QQQ/IWM); describes finished sessions; forecasts nothing.

Which timescale is doing the work

Wavelet variance split by clock: how much of each day type’s movement lived at ten minutes, forty, the hour-plus trend. Same arc, different animal — in percentages.

day typesessions10m20m40m80mhour-plus trend
TD+ bull trend day1,1242.4%3.3%5.3%9.8%78.2%
TD- bear trend day8233.0%4.3%6.9%12.4%72.6%
TFO+ bull trend-from-the-open6503.5%4.8%8.0%14.9%67.2%
TFO- bear trend-from-the-open6043.8%5.3%8.8%16.6%63.9%
TR trading range day7534.1%5.8%9.8%18.5%61.1%
TTR+ bull trending trading range5633.8%5.3%9.0%16.7%63.9%
TTR- bear trending trading range5153.6%5.1%9.3%17.1%63.5%

Haar wavelet (MODWT) variance split of the close path over 5,032 labeled sessions (SPY/QQQ/IWM); the energy identity is exact and bench-pinned (a pure oscillation lands on its matching scale). Read the SPREAD, not the absolute levels — coarse scales dominate any random-walk-like series. The spread is a clean gradient: a bull trend day parks 78% of its variance in the hour-plus trend with 2.4% at the 10-minute clock, while a trading-range day nearly doubles the fast-clock share (4.1%) and gives up a sixth of the trend’s coarse energy. On a trend day the 5-minute chart is nearly decoration; on a range day the fast clocks are where the day actually happens. Analysis starts at 10:45 ET — the coarsest filter spends the first 75 minutes as warmup, a stated tradeoff rather than invented pre-open tape. Describes finished sessions; forecasts nothing.

The dictionary of days

SAX words: every session spelled as eight letters, so same-shaped days print the same word and the archive greps by string equality. The ten most-printed silhouettes, with day-type mixes the words were never shown.

aaabcddd
printed ×79 · 66% TD+ + 20% TFO+
latest: SPY 2026-06-29
aabcccdd
printed ×69 · 78% TD+ + 17% TFO+
latest: SPY 2026-06-30
aabccddd
printed ×61 · 67% TD+ + 30% TFO+
latest: QQQ 2026-06-29
aabbcddd
printed ×60 · 67% TD+ + 25% TFO+
latest: QQQ 2026-06-04
aabbccdd
printed ×51 · 82% TD+ + 14% TFO+
latest: IWM 2026-05-13
aabbbcdd
printed ×40 · 70% TD+ + 23% TFO+
latest: QQQ 2026-05-06
dddcbaaa
printed ×39 · 64% TD- + 26% TFO-
latest: SPY 2026-05-07
aaabccdd
printed ×36 · 67% TD+ + 14% TFO+
latest: IWM 2026-04-13
ddccbaaa
printed ×33 · 70% TD- + 15% TFO-
latest: IWM 2025-11-13
ddcbbbaa
printed ×29 · 66% TD- + 31% TFO-
latest: QQQ 2026-06-16

8 segments × 4 letters over 5,032 labeled sessions (SPY/QQQ/IWM) — a vocabulary sized by measurement: fine enough for 2,444 distinct silhouettes, coarse enough that the common ones actually repeat (the ten above cover 10% of all days). The dictionary validates against labels it was never shown: the perfect staircase is 82% bull trend days, and every top climb has its mirror decline. Words are exact-match keys — grep the archive for a chart, literally. Descriptive of finished sessions; forecasts nothing.

The persistence of swings

Every pullback’s exact prominence — how far the market rallied before that low’s basin was absorbed — via 0-dimensional persistent homology. The census of swings with real teeth, per day type.

day typesessionstoothy dips (med)toothy rallies (med)deepest dip (med)deepest rally (med)days with no toothy dip
TD+ bull trend day1,124441.34σ1.46σ4%
TD- bear trend day823441.55σ1.46σ3%
TFO+ bull trend-from-the-open650552.12σ1.87σ1%
TFO- bear trend-from-the-open604561.75σ2.33σ0%
TR trading range day753662.62σ2.7σ0%
TTR+ bull trending trading range563662.14σ2.97σ1%
TTR- bear trending trading range515662.52σ2.4σ0%

0-dimensional sublevel persistence of the close path over 5,032 labeled sessions (SPY/QQQ/IWM): every pullback low paired with the rally that absorbed its basin, its PROMINENCE exact by construction (the machinery is benched to recover constructed dips to machine precision, and jitter of size e can only fake prominence 2e — noise separation is the method’s core guarantee). A swing has “teeth” at 0.5σ of prominence. Read ACROSS types — within one, the trend itself inflates prominences. The gradient is the price-action claim, measured: trend days print the fewest toothy swings and the shallowest counter-swings (and 4% of bull trend days print no toothy pullback at all — the never-give-an-entry day), trading ranges print six a side with counter-swings twice as prominent. This is the H0 half of topological data analysis; the H1 half — a double top as a long-lived loop — is NOT shipped and stays on the roadmap. Describes finished sessions; forecasts nothing.

The session as a state machine

A fitted hidden Markov model: four states that named themselves — quiet, bull drive, bear drive, loud chop — and how much of each day type lives in each. Trend days run their drive two-to-one; ranges sit dead-balanced.

day typesessionsquietloud 0.76 · drift 0.05bear driveloud 1.27 · drift -0.5bull driveloud 1.27 · drift 0.56loud choploud 2 · drift -0.13
TD+ bull trend day1,12431%17%38%15%
TD- bear trend day82328%38%16%18%
TFO+ bull trend-from-the-open65032%21%33%13%
TFO- bear trend-from-the-open60429%34%22%16%
TR trading range day75331%27%27%15%
TTR+ bull trending trading range56331%24%29%15%
TTR- bear trending trading range51531%29%24%16%

A four-state hidden Markov model over the tape’s half-hour character (trailing means of seasonally deflated loudness and drift), fitted once on 600 sessions with ten seeded restarts (six land on the same optimum) and decoded over all 5,032. The state names fell out of the parameters — quiet, two drives, loud chop — and the census reads exactly as a price-action reader would put it: a bull trend day spends 38% of its bars in bull drive against 17% in bear drive, the bear day mirrors it, a trading range sits dead-balanced at 27/27, and the trending ranges lean their way. Three design iterations were required and are part of the record: per-bar features produced a memoryless mixture (median dwell one bar — 5-minute returns barely remember the last bar), and mean-subtracting each session blinded the drives to their own trend (bull days “spent more time in bear drive” — an impossibility that exposed the bug). Median dwell now sits at the half-hour window’s floor, so read OCCUPANCY for structure, not dwell. Describes finished sessions; forecasts nothing.

The loop that will not die

H1 persistent homology of the delay-embedded session: a double top that keeps mattering is a long-lived loop, and the census says trading ranges literally contain loops (median one) while trend days literally do not (median zero).

day typesessionsloopiness (med)loopiness (p90)real loops (med)
TD+ bull trend day1,1240.140.340
TD- bear trend day8230.160.320
TFO+ bull trend-from-the-open6500.200.430
TFO- bear trend-from-the-open6040.210.420
TR trading range day7530.290.561
TTR+ bull trending trading range5630.270.541
TTR- bear trending trading range5150.270.541

H1 persistent homology of each session’s delay embedding (z-close against itself 30 minutes back) over 5,032 labeled sessions (SPY/QQQ/IWM). Loopiness = the lifetime of the longest-lived loop; a “real loop” lives ≥0.25 z-units. The machinery is exact (benched against a hand-computable case to machine precision) and noise-immune by theorem — jitter can only make short-lived loops. The gradient is the chapter’s claim delivered: trading ranges literally contain loops (median one per session) and trend days literally do not (median zero), with every one of the loopiest days in the archive a range day. Cross-instrument check: day-scale double-structure tags do NOT separate loopiness (0.21 vs 0.21 median, n=1170/330) — a null, published. Describes finished sessions; forecasts nothing.

The character of a day, in numbers

Dynamical-systems measurements per day type: how fast stretches away from the mean decay, how often reversion fails to register at all, how stuck and how orderly the tape runs.

day typereversion half-life (med, IQR)no measurable reversionlaminarity (med)perm. entropy (med)Hurst (med)
TD+ bull trend day47m(2883m)7%0.8390.880.54
TD- bear trend day49m(2987m)9%0.8370.8910.534
TFO+ bull trend-from-the-open32m(2054m)1%0.8120.8920.54
TFO- bear trend-from-the-open32m(2054m)1%0.8130.8970.542
TR trading range day29m(1945m)0%0.7990.8980.544
TTR+ bull trending trading range39m(2466m)3%0.8140.8940.543
TTR- bear trending trading range39m(2467m)3%0.8150.8960.537

Dynamical-systems measurements over 180,591 labeled sessions (117 tickers): how many minutes a stretch away from the mean takes to decay (Ornstein–Uhlenbeck half-life), how often a day trends so cleanly reversion is unmeasurable, how much time the tape spends stuck in one neighborhood (RQA laminarity), how orderly the bar sequence is (permutation entropy), and trend persistence (DFA Hurst). The half-life column is the separator — trend days’ stretches decay slowest and are the only types where reversion regularly fails to register; ranges snap back fastest. Hurst, honestly, barely separates the types at this bar size — a measured null, reported rather than hidden. Character describes a finished session; it forecasts nothing.

Base rates, hallmarks, and the full day-type taxonomy live on Day Types.

More exhibits

Computed galleries that live on their own pages.