Long-horizon benchmark by tier: T+60 tested, T+90 pending

Live benchmark · results separated by tier

Long horizon.
Tier by tier.

Lite, Pro, Max, and Ultra-Max are reported as four separate cohorts. T+60 and T+90 are the primary evidence windows. T+30 remains visible for audit transparency, but is treated as exploratory noise—not the performance conclusion.

LAST FULL MARKET CLOSE · 31 JUL 2026

Lite · T

267

signals · 60 equities

Pro · T

382

signals · 247 equities

Max · T

914

signals · 342 equities

Ultra-Max · T

signals · 40 equities

The long-horizon verdict

Results below are intentionally not added together. A decisive hit rate is right / (right + wrong) for directional calls only. Flat realized moves, neutral calls, abstentions, and missing prices remain visible. The baseline is the better of always-up or always-down within that exact tier and window.

60 / 90

No mature tier beat its exact-cohort majority baseline at T+60; T+90 had no mature signals by the cutoff. At T+60, Max recorded 80 right and 46 wrong for 63.49%, but trailed baseline by 5.56 points. Lite recorded 11 right and 6 wrong for 64.71%, but trailed baseline by 23.53 points. Pro and Ultra-Max were not mature at T+60. T+30 is retained below as exploratory noise only.

Lite

T · 267 signals · 60 equities
87 up · 55 down · 13 neutral · 112 abstain

Checkpoint	Eligible / priced	Direction R / W / flat	Hit rate	Neutral R / W	Abstain	Majority baseline	Lift
T+60	106 / 103	11 / 6 / 9	64.71%	0 / 0	77	88.24% down	−23.53 pp
T+90	0 / 0	Not mature at cutoff
Through 31 Jul	267 / 265	48 / 46 / 46	51.06%	6 / 7	112	70.21% down	−19.15 pp
T+30 · noise	151 / 147	25 / 8 / 15	75.76%	0 / 0	99	78.79% down	−3.03 pp

Lite at T+60: 64.71%—11 right and 6 wrong—but 23.53 points below its bearish cohort baseline. This is not a benchmark-beating result. T+90 was not mature. The through-date mark mixes holding periods; T+30 is retained as exploratory noise only.

Lite · T+60 · 9 stocks · 11 right calls

Lite · T+90 · no mature right calls

Lite · Through 31 Jul · 27 stocks · 48 right calls

Lite · T+30 exploratory/noise · 17 stocks · 25 right calls

Pro

T · 382 signals · 247 equities
168 up · 102 down · 112 neutral

Checkpoint	Eligible / priced	Direction R / W / flat	Hit rate	Neutral R / W	Abstain	Majority baseline	Lift
T+60	0 / 0	Not mature at cutoff
T+90	0 / 0	Not mature at cutoff
Through 31 Jul	382 / 382	126 / 62 / 82	67.02%	30 / 82	0	57.98% up	+9.04 pp
T+30 · noise	371 / 371	112 / 66 / 81	62.92%	39 / 73	0	62.36% up	+0.56 pp

Pro had no mature T+60 or T+90 cohort by 31 July. Its 67.02% through-date mark mixes holding periods and cannot replace a fixed-horizon test. T+30 is retained as exploratory noise only.

Pro · T+60 · no mature right calls

Pro · T+90 · no mature right calls

Pro · Through 31 Jul · 115 stocks · 126 right calls

Pro · T+30 exploratory/noise · 107 stocks · 112 right calls

Max

T · 914 signals · 342 equities
571 up · 288 down · 55 neutral

Checkpoint	Eligible / priced	Direction R / W / flat	Hit rate	Neutral R / W	Abstain	Majority baseline	Lift
T+60	161 / 150	80 / 46 / 24	63.49%	0 / 0	0	69.05% down	−5.56 pp
T+90	0 / 0	Not mature at cutoff
Through 31 Jul	914 / 892	393 / 215 / 229	64.64%	19 / 36	0	55.10% down	+9.54 pp
T+30 · noise	847 / 822	356 / 205 / 207	63.46%	19 / 35	0	51.52% up	+11.94 pp

Max at T+60: 63.49%—80 right and 46 wrong—but 5.56 points below its bearish cohort baseline. This is not a benchmark-beating result. T+90 was not mature. T+30 is retained as exploratory noise and is not used to name a winning tier.

Max · T+60 · 38 stocks · 80 right calls

Max · T+90 · no mature right calls

Max · Through 31 Jul · 182 stocks · 393 right calls

Max · T+30 exploratory/noise · 182 stocks · 356 right calls

Ultra-Max

T · 71 signals · 40 equities
28 up · 43 down

Checkpoint	Eligible / priced	Direction R / W / flat	Hit rate	Neutral R / W	Abstain	Majority baseline	Lift
T+60	0 / 0	Not mature at cutoff
T+90	0 / 0	Not mature at cutoff
Through 31 Jul	71 / 71	27 / 14 / 30	65.85%	0 / 0	0	60.98% down	+4.88 pp
T+30 · noise	0 / 0	Not mature at cutoff

Ultra-Max had no mature T+60 or T+90 cohort by 31 July. Its 65.85% through-date mark mixes holding periods and cannot be treated as fixed-horizon evidence. T+30 was also immature and is retained only as an exploratory audit row.

Ultra-Max · T+60 · no mature right calls

Ultra-Max · T+90 · no mature right calls

Ultra-Max · Through 31 Jul · 17 stocks · 27 right calls

Ultra-Max · T+30 exploratory/noise · no mature right calls

A ticker appears only within the tier and window where at least one timestamped up or down call was correct. ×N exposes repeated correct events inside that same tier. Appearance does not mean every call on that stock was right. Correct neutral calls are shown in the tables but excluded from the directional stock lists. T+60 tapes are shown first; T+90 has no mature events. T+30 tapes are collapsed and kept only for audit transparency.

Read the prior frozen benchmark

The event-level audit ledger is retained internally and is not offered as a public download.

How the ledger was locked

PopulationLive product-tier equity signals only. Every backdate-* replay, ETF, leveraged ETF, fund, and crypto vehicle is excluded.

PricesSplit/dividend-adjusted Yahoo daily bars. Entry is the first regular-session open actually available after issuance.

DirectionCanonical GMM μ sign first, stored chip only when μ is absent. Realized moves inside ±5% are flat/inconclusive.

CutoffsT+30, T+60, and T+90 exit at the first close on or after the calendar horizon. No T+90 signal was mature by 31 July. Through-date exits at the latest close by the cutoff; holding periods vary.

Audit hashes: equity cohort b59d5dff…e6a37c · adjusted price cache 42a3c841…18d9d · right-call CSV 6d5857a6…fa6899.

Important: This is a descriptive paper signal-direction study, not customer returns, a trading strategy, or investment advice. No trade was executed. Figures omit fees, spreads, slippage, taxes, position sizing, and portfolio construction. Prediction events are correlated, the through-date windows vary, and AI-generated research may be wrong. Past observed results do not indicate future results.