Live direction ledger · market close 28 Jul 2026
We kept the misses.
Quorum issued the calls. The market supplied the
outcome. We froze every live equity signal, then compared the council
with the simplest honest alternative: always choose the majority
direction in the same cohort.
T+30 decisive directions
493 / 278
right / wrong · 63.94% directional hit rate
Same-cohort baseline
52.66%
always choosing the T+30 cohort's majority direction
T+90
Running
no live product-tier signal was mature at cutoff
The complete T+30 ledger
At T0, the corpus contained 1,644 live equity prediction
events across 342 tickers. At T+30, 1,365 were old enough and
1,336 had usable adjusted prices.
+11.28 pp
descriptive lift over the same-cohort majority baseline.
Quorum scored 63.94% on 771 decisive up/down calls; the cohort's
always-up baseline scored 52.66%. The descriptive Wilson 95%
interval for Quorum was 60.49%–67.26%.
| Checkpoint | Price evaluated | Right |
Wrong | Flat / inconclusive | Neutral R/W |
Abstain | Baseline comparison |
|---|
| T+30 | 1,336 | 493 |
278 | 303 | 58 / 108 |
96 | +11.28 pp |
| T+60 | 249 | 89 |
51 | 32 | 0 / 0 |
77 | −7.14 pp |
| T+90 | 0 | — | — | — |
— | — | not mature |
Why T+60 is not a victory claim: Quorum was right on 89
and wrong on 51 decisive calls, or 63.57%. But this early cohort was
heavily bearish; an always-down rule scored 70.71%. A benchmark
should test discrimination, not reward the regime.
What the four-tier data actually supports
Quorum has four research tiers: Quorum Lite, Quorum Pro,
Quorum Max, and Quorum Ultra-Max. Quorum Pro+ is the
subscription that includes Quorum Max; Team / RIA is a five-seat
Quorum Pro plan, not a fifth research tier.
Free · selective
Quorum Lite
25 right · 7 wrong · 96 abstentions
Only 32 decisive T+30 calls. The raw 78.12%
rate trailed the cohort's always-down baseline, so we do not call
Lite “78% accurate.”
Cloud · breadth
Quorum Pro
77 right · 36 wrong across first signals
The one-first-signal-per-stock check covered
113 decisive stocks: 68.14% versus a 61.06% majority baseline.
Frontier · depth
Quorum Max
356 right · 205 wrong at T+30
The largest mature event cohort: 63.46%
versus a 51.52% majority baseline. Max adds deeper debate; this
observational study does not prove paid tiers cause better calls.
Fourth tier · not mature
Quorum Ultra-Max
71 live signals · T+30 not mature
The fourth tier had 71 live signals across
40 equities at cutoff, but none was old enough for a T+30
performance claim. Quorum Ultra-Max remains invite only; we
publish no accuracy rate before its clock matures.
Calls worth inspecting — including the miss
Max · called up
SEZL+51.03%
4 Jun signal · adjusted T+30 move
Max · called up
CDNA+46.28%
20 Jun signal · adjusted T+30 move
Max · called down
HYLN−50.39%
13 Jun signal · direction right
Max · later called up
HYLN−56.71%
20 Jun signal · direction wrong
A ticker is not permanently right or wrong. Each timestamped
briefing is a separate falsifiable event. Showing only the SEZL,
CDNA, and first HYLN calls would be misleading.
Alphabet: an audit note, not a headline
An experimental Quanta Downside Signal on GOOGL was followed by
a −3.97% move at exact T+30 — inside our ±5% flat band. One session
later it was down −6.06%, making the directional observation
correct at day 31. Quanta experimental is not a public product tier,
and earlier product-tier GOOGL calls were mixed. We therefore do
not publish “Quorum correctly predicted GOOGL at T+30.”
How it was scored
- The four research tiers — Quorum Lite, Quorum Pro, Quorum Max,
and Quorum Ultra-Max; every
backdate-* replay and 146 ETF/leveraged-ETF rows
excluded.
- Split/dividend-adjusted Yahoo daily bars. Entry is the first
regular-session open actually available after the signal.
- Exit is the first close on or after the signal's exchange-local
date plus 30, 60, or 90 calendar days.
- Canonical GMM μ sign determines up/down first; stored chip is
used only when μ is absent.
- Moves inside ±5% are flat/inconclusive. Neutral calls and
abstentions are reported separately.
- Both every prediction event and the earliest mature signal per
ticker within each tier were checked to expose repetition bias.
Important: This is a paper signal-direction study, not a
portfolio return, customer return, recommendation record, or
investable strategy. No trade was executed. Figures omit fees,
spreads, slippage, taxes, position sizing, and portfolio construction.
Prediction events are correlated, so descriptive intervals do not
capture every source of uncertainty. Past observed results do not
indicate future results. Quorum is educational research, not
investment advice, and its outputs may be wrong.