GioModel freezes its forecasts before the outcome exists and publishes them here in full, immediately. A frozen number cannot be edited after the fact, only graded.
This rule changed on 2026-08-26. Until then a release published no numbers for 30 days after its freeze. That lag was never about secrecy — fixing the disclosure date in advance made it impossible to show only the forecasts that worked. Immediate disclosure keeps that guarantee by a different route: every release must be public, in full, and a withheld one now fails the build rather than waiting. The full policy history is recorded below rather than overwritten.
Each release fixes its data panel, anchor price, P10–P90 range per horizon, and probability — hashed and dated. Members see it the day it is made.
A horizon of h sessions resolves on exactly the h-th completed NYSE/Nasdaq session after the freeze session — holidays and weekends counted properly, never a calendar-day approximation. The realized close is the official close of that session and nothing else.
On the unlock date the frozen numbers and their outcomes appear below, unchanged. Outcomes are written once to an append-only ledger; a later disagreement raises an error rather than silently correcting the record.
A site that argues its claims are checkable cannot quietly edit a printed commitment. Every version of the disclosure rule is kept here, oldest first. Past entries are appended to, never rewritten.
A frozen release published no numbers at all until 30 calendar days after its freeze date.
Fixing the disclosure date in advance, before any outcome existed, made it impossible to publish only the forecasts that turned out well.
A frozen release is published in full immediately, on the day it is frozen.
Changed at the owner's direction so current forecasts are visible rather than sealed for a month. The anti-cherry-picking guarantee is preserved by completeness instead of by delay: every release must be public in full, and a withheld release now fails the build.
What the change costs: The old rule proved its point without needing to be trusted: a date fixed in advance cannot be gamed. The new rule depends on the completeness invariant actually holding, which is a weaker form of proof because it has to be checked continuously rather than once.
Every row below was frozen and hashed before its outcome existed, after the public protocol began. Nothing from model development appears here, however favourable it looked — that evidence is kept separately, and labelled as development evidence, on each model's stock page. Keeping the two apart is the whole point: a backtest that is allowed to migrate into a track record stops being either.
25 frozen releases on record · 25 public · 0 inside the 0-day window
| Model | Ticker | Frozen | Public on | Anchor | 1-wk band | 2-wk band | 1-mo band | Status |
|---|---|---|---|---|---|---|---|---|
| 01 | ASTS | Jul 29, 2026 | Jul 29, 2026 | $53.03 | $45.43–$65.632026-08-05 close $68.38 · OUTSIDE · median err +28.95% | $42.21–$73.89resolves 2026-08-12 (Wed) | $38.93–$86.23resolves 2026-08-27 (Thu) | PUBLIC |
| 02 | HOOD | Jul 29, 2026 | Jul 29, 2026 | $89.84 | $81.05–$102.262026-08-05 close $92.80 · inside · median err +2.62% | $77.73–$108.26resolves 2026-08-12 (Wed) | $75.62–$126.29resolves 2026-08-27 (Thu) | PUBLIC |
| 03 | NKE | Jul 30, 2026 | Jul 30, 2026 | $42.29 | $40.01–$44.272026-08-06 close $42.00 · inside · median err −0.51% | $39.29–$44.96resolves 2026-08-13 (Thu) | $37.51–$46.85resolves 2026-08-28 (Fri) | PUBLIC |
| 04 | UBER | Jul 31, 2026 | Jul 31, 2026 | $70.36 | $66.69–$74.73resolves 2026-08-07 (Fri) | $65.35–$77.49resolves 2026-08-14 (Fri) | $64.01–$79.56resolves 2026-08-31 (Mon) | PUBLIC |
| 05 | NVDA | Jul 31, 2026 | Jul 31, 2026 | $200.75 | $183.54–$222.21resolves 2026-08-07 (Fri) | $179.56–$230.55resolves 2026-08-14 (Fri) | $177.39–$246.31resolves 2026-08-31 (Mon) | PUBLIC |
| 06 | RDDT | Jul 31, 2026 | Jul 31, 2026 | $140.67 | $122.06–$160.01resolves 2026-08-07 (Fri) | $107.75–$170.59resolves 2026-08-14 (Fri) | $100.06–$175.70resolves 2026-08-31 (Mon) | PUBLIC |
| 07 | PLTR | Jul 31, 2026 | Jul 31, 2026 | $123.06 | $108.93–$138.40resolves 2026-08-07 (Fri) | $100.40–$149.52resolves 2026-08-14 (Fri) | $91.15–$169.10resolves 2026-08-31 (Mon) | PUBLIC |
| 08 | FUTU | Jul 31, 2026 | Jul 31, 2026 | $105.12 | $95.75–$118.31resolves 2026-08-07 (Fri) | $90.92–$128.27resolves 2026-08-14 (Fri) | $86.49–$139.53resolves 2026-08-31 (Mon) | PUBLIC |
| 09 | PYPL | Aug 5, 2026 | Aug 5, 2026 | $57.93 | $54.01–$61.36resolves 2026-08-12 (Wed) | $52.36–$63.05resolves 2026-08-19 (Wed) | $49.51–$65.64resolves 2026-09-03 (Thu) | PUBLIC |
| 10 | CRM | Aug 5, 2026 | Aug 5, 2026 | $192.98 | $181.31–$204.55resolves 2026-08-12 (Wed) | $176.71–$209.61resolves 2026-08-19 (Wed) | $169.35–$218.69resolves 2026-09-03 (Thu) | PUBLIC |
| 11 | NFLX | Aug 5, 2026 | Aug 5, 2026 | $74.20 | $69.84–$78.96resolves 2026-08-12 (Wed) | $68.00–$81.46resolves 2026-08-19 (Wed) | $65.40–$85.43resolves 2026-09-03 (Thu) | PUBLIC |
| 12 | ORCL | Aug 5, 2026 | Aug 5, 2026 | $144.39 | $136.01–$152.62resolves 2026-08-12 (Wed) | $132.38–$158.10resolves 2026-08-19 (Wed) | $127.13–$166.01resolves 2026-09-03 (Thu) | PUBLIC |
| 13 | AMT | Aug 5, 2026 | Aug 5, 2026 | $168.07 | $160.44–$175.69resolves 2026-08-12 (Wed) | $157.32–$179.33resolves 2026-08-19 (Wed) | $152.56–$184.65resolves 2026-09-03 (Thu) | PUBLIC |
| 14 | WMT | Aug 5, 2026 | Aug 5, 2026 | $112.34 | $109.17–$116.46resolves 2026-08-12 (Wed) | $107.76–$118.83resolves 2026-08-19 (Wed) | $106.16–$122.53resolves 2026-09-03 (Thu) | PUBLIC |
| 15 | AMD | Aug 5, 2026 | Aug 5, 2026 | $482.05 | $444.97–$532.40resolves 2026-08-12 (Wed) | $431.59–$559.50resolves 2026-08-19 (Wed) | $413.93–$606.07resolves 2026-09-03 (Thu) | PUBLIC |
| 16 | CRWD | Aug 5, 2026 | Aug 5, 2026 | $209.86 | $192.64–$232.24resolves 2026-08-12 (Wed) | $187.37–$242.72resolves 2026-08-19 (Wed) | $180.23–$262.61resolves 2026-09-03 (Thu) | PUBLIC |
| 17 | SOFI | Aug 5, 2026 | Aug 5, 2026 | $18.25 | $16.31–$20.38resolves 2026-08-12 (Wed) | $15.61–$21.43resolves 2026-08-19 (Wed) | $14.66–$23.18resolves 2026-09-03 (Thu) | PUBLIC |
| 18 | RKLB | Aug 5, 2026 | Aug 5, 2026 | $74.82 | $66.56–$86.02resolves 2026-08-12 (Wed) | $63.26–$91.71resolves 2026-08-19 (Wed) | $58.28–$102.32resolves 2026-09-03 (Thu) | PUBLIC |
| 19 | ONON | Aug 5, 2026 | Aug 5, 2026 | $38.19 | $35.10–$41.52resolves 2026-08-12 (Wed) | $33.87–$43.09resolves 2026-08-19 (Wed) | $32.06–$45.77resolves 2026-09-03 (Thu) | PUBLIC |
| 20 | CELH | Aug 5, 2026 | Aug 5, 2026 | $29.15 | $26.16–$32.96resolves 2026-08-12 (Wed) | $24.98–$35.10resolves 2026-08-19 (Wed) | $23.21–$38.96resolves 2026-09-03 (Thu) | PUBLIC |
| 21 | NET | Aug 5, 2026 | Aug 5, 2026 | $292.96 | $264.44–$327.37resolves 2026-08-12 (Wed) | $253.48–$344.91resolves 2026-08-19 (Wed) | $240.65–$376.38resolves 2026-09-03 (Thu) | PUBLIC |
| 22 | PATH | Aug 5, 2026 | Aug 5, 2026 | $13.82 | $12.38–$15.14resolves 2026-08-12 (Wed) | $11.96–$15.69resolves 2026-08-19 (Wed) | $10.99–$16.61resolves 2026-09-03 (Thu) | PUBLIC |
| 23 | ZS | Aug 5, 2026 | Aug 5, 2026 | $161.65 | $148.06–$176.16resolves 2026-08-12 (Wed) | $141.81–$184.72resolves 2026-08-19 (Wed) | $134.03–$196.52resolves 2026-09-03 (Thu) | PUBLIC |
| 24 | DKNG | Aug 5, 2026 | Aug 5, 2026 | $21.76 | $19.63–$24.04resolves 2026-08-12 (Wed) | $18.61–$25.07resolves 2026-08-19 (Wed) | $17.33–$26.79resolves 2026-09-03 (Thu) | PUBLIC |
| 25 | RIVN | Aug 5, 2026 | Aug 5, 2026 | $15.56 | $13.71–$17.52resolves 2026-08-12 (Wed) | $12.89–$18.36resolves 2026-08-19 (Wed) | $11.72–$19.71resolves 2026-09-03 (Thu) | PUBLIC |
A horizon resolves on exactly the h-th completed trading session after the freeze session, and is graded against the official close of that session — no calendar-day approximation and no substitute close. Outcomes are written once to an append-only ledger and never revised. As of 2026-08-26 every frozen release is published in full on the day it is frozen; the previous 30-day disclosure lag, and the reason it was replaced, are recorded in disclosurePolicyHistory below. Horizons are counted in trading sessions on the NYSE/Nasdaq session calendar (weekends, scheduled holidays with observed-date rules, and declared unscheduled closures). Releases join this ledger the day they are frozen; nothing is added retroactively and nothing leaves.
NO-GO does not mean the research contains no information. It means GioModel refuses to convert incomplete evidence into a trading recommendation. Members receive the ranges, the probabilities, the diagnostics and the evidence — everything the model produced. What they do not receive is a claim that the model has earned directional deployment, because the deployment gate is independent of payment. No subscription moves a gate. Read the nine gates and the method in full — they are free.
The first build's five-model search produced a best post-hoc AUC of 0.458 across 740 out-of-sample rows — worse than a coin flip once selection effects were counted, with fold AUCs swinging from 0.320 to 0.675. A Phase B challenger tuned on the same development dates reached 2 of 9 gates, but reusing those dates disqualified it as fresh confirmation. Verdict: benchmark only. The specification was rejected; the process was kept.
The selected Signed Core Ridge family passed 2 of 9 gates. The recurring failure was calibration: quantile-bin expected calibration error stayed above the declared threshold in every repair path, and six bounded Phase B repair attempts were spent without converting development promise into gate passes. Publishing the failure was the deliverable.
The third build added a terminal holdout, next-session filing availability, actual ex-dividend dates, and a traceable source ledger — and still passed only G3 and G6. Its corrected protocol was frozen before the rerun, but that is not independent confirmation. Improved craft does not create a deployable model; only evidence does.
The 2026-08-05 seven-ticker release lands 5–6 of 9 gates per model. WMT and AMD carry an extra open item: no independently verified close block exists for them, so the cross-verification gate cannot pass yet. Every model in the release stays unapproved until its remaining gates — including post-freeze forward evidence — are earned.