# Weekly Model Watch #6

**WEEKLY · MODEL WATCH** · September 15, 2026 · MarketMania Research · Weekly series

Window: **Sep 7-13, 2026 UTC** · Cutoff: **Mon Sep 14, 2026, 16:00 UTC** · Methodology **v1.1 (2026-08-10)**, hash `e66c7e8c864a2233`

PDF: https://marketmania.ai/research/reports/model-watch-2026-09-07.pdf · Open data (JSON): https://marketmania.ai/research/reports/model-watch-2026-09-07.json

> Research question: *"who led the week, which tickers were readable, who saw the turn first, and does a model agreeing with itself across timeframes/horizons make its call stronger?"*

## Research Snapshot

| RESEARCH SNAPSHOT |  |  |  |
|---|---|---|---|
| **4,468** directional calls scored | **7** models (stable lineup) | of **9,301** mature forecasts | window **Sep 7-13, 2026 UTC** |
| hit = direction rule: tp1/tp2->hit, sl->miss, expiry->sign of gross pnl | FH gate **1h / 4h / 1d** | base field hit **39.6%** | cutoff **Mon Sep 14, 2026, 16:00 UTC** |
| market: BTC net **-4.36%** (prior +3.42%) | ann vol **19.7%** (was 43.4%) | TOP5 volume **$16.05B**, -1% w/w | pairwise corr **0.74** (was 0.78) |

## Key finding

> **KEY FINDING · OBSERVATION (one weekly window)**
>
> ### Field accuracy went 42.8% -> 39.6% (-3.3pp) with 1 of 7 models improving, and no weekly title was awarded, for the sixth issue running.

## TL;DR

- **OBSERVATION -- No weekly title again.** gemini-3.1-pro leads descriptively at 41.8% [38.1%, 45.5%]; the gap to #2 (deepseek-v4-pro, 41.4%) is 0.4pp. 1 of 7 models improved week over week; issue #5's leader gemini-3.1-pro is now 1 of 7.
- **No ticker cleared 51%.** BNB is the week's most readable coin (46.9%, n=963) and ETH the hardest at 29.5%; 0 of the 5 assets cleared 51%. Pair of the week: gemini-3.1-pro x BNB 50.0% (n=148); worst: grok-4.6 x ETH 25.6% (n=82).
- **1 of 2 qualifying turns found a caller.** ETH long +4.19%, 11 caller(s), first gemini-3.1-pro; ETH short -2.11%, 0 caller(s), first NOBODY.
- **Self-agreement split across the four cells.** Cross-TF/FH lifts ran -6.0pp to +0.3pp (issue #5: -17.4pp to -2.5pp). On trend days, 4h TF-agreement hit 52.8% (n=195) against 21.2% (n=325) on flat days.

### Week-over-week chart (see PDF for the grouped bar chart)

Directional hit-rate (%) by model, week over week, ranked by this week's rate. Short names: opus-5 = claude-opus-5, fable-5 = claude-fable-5, gemini-3.1 = gemini-3.1-pro, gpt-5.6 = gpt-5.6-sol, qwen-3.8 = qwen-3.8-max, deepseek = deepseek-v4-pro. Field base 42.8% -> 39.6%. The grok column compares two pure grok-4.6 weeks: the 4.5 -> 4.6 flip sits two windows back. "Was" values are the numbers published in issue #5; the leaderboard's W/w column is computed on unrounded rates and can differ by 0.1pp from the difference of the rounded columns.

## Why it matters

MarketMania scores 7 model lines against the same market, hour after hour. Weekly Model Watch asks four practical questions about that week of calls: who led the week, which tickers were readable, who saw the turn first, and does a model agreeing with itself across timeframes/horizons make its call stronger? Each question gets its own block below, with its own table, its own N, and its own honesty caveats. The grok row is a pure grok-4.6 week in this issue and was one in issue #5 as well -- the 4.5 -> 4.6 flip sits two windows back -- so read the grok week-over-week row as a line, not a single model build.

## How to read this

Every block below is scored on the same pool: matured directional calls inside the 1h/4h/1d FH gate, hit by the methodology v1.1 direction rule. The weekly title is awarded only on a >=5pp gap AND non-overlapping 95% Wilson CIs with N>=10 per side -- a descriptive lead is not a title. Reversals use the 4h grid with a >=2.0% counter-move sustained 8h, BTC/ETH only in v1. Cross-TF/FH blocks ask whether a model agreeing with itself does any better, and are split trend vs flat wherever the week supplies both sides. The grok 4.5 -> 4.6 flip (Aug 24, 2026 09:22 UTC) sits two windows back (issue #4): both weeks of this issue are pure grok-4.6, with no 4.5-era rows in either window, so the grok week-over-week row is the first pure-4.6 against pure-4.6 comparison of the series.

## Leaderboard of the week

| Model | N | Cov. | Hit rate | 95% CI | W/w pp |
|---|---|---|---|---|---|
| gemini-3.1-pro | 673 | 50.6% | 41.8% | 38.1%-45.5% | -4.8 |
| deepseek-v4-pro | 727 | 54.7% | 41.4% | 37.9%-45.0% | -2.8 |
| gpt-5.6-sol | 709 | 53.4% | 40.1% | 36.5%-43.7% | -1.6 |
| qwen-3.8-max | 707 | 53.4% | 40.0% | 36.5%-43.7% | -2.4 |
| grok-4.6 | 518 | 39.0% | 39.6% | 35.5%-43.9% | +2.0 |
| claude-fable-5 | 624 | 46.9% | 36.5% | 32.9%-40.4% | -6.3 |
| claude-opus-5 | 510 | 38.4% | 36.5% | 32.4%-40.7% | -5.9 |

No weekly title is awarded this issue -- for the sixth time. Title rule: **>=5pp gap AND non-overlapping 95% CI, N>=10** per side. gemini-3.1-pro's lead over deepseek-v4-pro is 0.4pp and their CIs overlap, so neither condition is met. The grok 4.5 -> 4.6 flip (Aug 24, 2026 09:22 UTC) sits two windows back (issue #4): both weeks of this issue are pure grok-4.6, with no 4.5-era rows in either window, so the grok week-over-week row is the first pure-4.6 against pure-4.6 comparison of the series.

## Ticker of the week

| Symbol | N | Hit rate | 95% CI | Issue #5 |
|---|---|---|---|---|
| BNB | 963 | 46.9% | 43.8%-50.1% | 47.3% |
| XRP | 937 | 42.2% | 39.0%-45.3% | 44.2% |
| BTC | 891 | 41.4% | 38.2%-44.7% | 40.6% |
| SOL | 875 | 36.0% | 32.9%-39.2% | 38.7% |
| ETH | 802 | 29.5% | 26.5%-32.8% | 43.1% |

BNB was the field's most readable ticker (46.9%) and ETH the hardest (29.5%); 0 of 5 assets cleared 51% this week (issue #5: 0 of 5). Issue #5's best ticker was BNB and its hardest SOL -- this issue's ordering is BNB > XRP > BTC > SOL > ETH.

## Pair of the week

Top 6 of the 12 model x ticker pairs tracked this week (best hit-rate), then the 5 worst in their own table -- the pair list is split rather than paged so neither table breaks across a page. Pairs shown require N of 8 or more calls.

| Model (top 6) | Symbol | N | Hit rate |
|---|---|---|---|
| gemini-3.1-pro | BNB | 148 | 50.0% |
| deepseek-v4-pro | BNB | 162 | 49.4% |
| grok-4.6 | BNB | 112 | 49.1% |
| gpt-5.6-sol | BNB | 157 | 48.4% |
| claude-opus-5 | BNB | 118 | 45.8% |
| deepseek-v4-pro | BTC | 153 | 45.1% |

| Model (5 worst) | Symbol | N | Hit rate |
|---|---|---|---|
| gemini-3.1-pro | ETH | 122 | 30.3% |
| claude-opus-5 | ETH | 95 | 29.5% |
| gpt-5.6-sol | ETH | 121 | 28.1% |
| deepseek-v4-pro | ETH | 132 | 28.0% |
| grok-4.6 | ETH | 82 | 25.6% |

gemini-3.1-pro x BNB (50.0%, n=148) is the week's best pair and grok-4.6 x ETH (25.6%, n=82) the weakest. Pair history accumulates across issues -- read this as the sixth data point, not a ranking.

## Who saw the reversal first

Reversal definition: 4h grid; trend = sign of prior 24h; counter-move of 2.0% or more (BTC/ETH) sustained 8h. A model 'calls' the reversal if it has a matured directional hit call in the new direction, FH 4h or 1d, in the slot window [T-12h, T+FH].

| Symbol | Time (UTC) | New side | Move (8h) | Callers | First caller |
|---|---|---|---|---|---|
| ETH | Fri Sep 11, 08:00 | Long | +4.19% | 11 | gemini-3.1-pro |
| ETH | Fri Sep 11, 16:00 | Short | -2.11% | 0 | NOBODY |

The ETH long turn drew 11 caller(s), first gemini-3.1-pro on a 1d/1d call at Sep 12, 00:01 with confidence 70; the ETH short turn drew NOBODY. Scope in v1 remains BTC/ETH only; the event count across six issues is still far too small to characterise turn-detection skill either way.

## Cross-TF and cross-FH confirmation: does agreeing with yourself help?

| Block | FH | vs. | Agree n | Agree hit | Dis n | Dis hit | Base hit | Lift |
|---|---|---|---|---|---|---|---|---|
| cross-TF | 4h | TF 4h vs 1h | 520 | 33.1% | 10 | 80.0% | 35.4% (704) | -2.3 |
| cross-TF | 1d | TF 1d vs 4h | 84 | 11.9% | 5 * | 40.0% | 16.7% (132) | -4.8 |
| cross-FH | 4h | by FH 1h | 485 | 35.7% | 6 * | 83.3% | 35.4% (704) | +0.3 |
| cross-FH | 1d | by FH 4h | 75 | 10.7% | 8 * | 12.5% | 16.7% (132) | -6.0 |

* = N<10 (insufficient); insufficient cells are never used to rank models. Mixed pairs (a model with only one side of the comparison present): 173 / 43 / 211 / 49. Pooled across all four cell families the lifts run -6.0pp to +0.3pp this week (issue #5: -17.4pp to -2.5pp).

**Regime control.** Methodology calls for a trend-vs-flat split on this table. This window has 2 trend days of 7: 09-07 -1.52% (trend), 09-08 -0.72%, 09-09 -0.25%, 09-10 -2.14% (trend), 09-11 +0.85%, 09-12 +0.00%, 09-13 -0.57%. On trend days, 4h TF-agreement hit 52.8% (n=195) against 21.2% (n=325) on flat days.

## Side-mix and the week's failure

| Model | N | Long | Short | Sideways | L:S |
|---|---|---|---|---|---|
| claude-fable-5 | 1,330 | 20.7% | 26.2% | 53.1% | 0.79 |
| claude-opus-5 | 1,330 | 17.9% | 20.4% | 61.7% | 0.87 |
| deepseek-v4-pro | 1,330 | 19.9% | 34.7% | 45.3% | 0.57 |
| gemini-3.1-pro | 1,330 | 19.3% | 31.3% | 49.4% | 0.62 |
| gpt-5.6-sol | 1,327 | 16.4% | 37.0% | 46.6% | 0.44 |
| grok-4.6 | 1,330 | 11.4% | 27.5% | 61.1% | 0.42 |
| qwen-3.8-max | 1,324 | 20.5% | 32.9% | 46.6% | 0.63 |

L:S = each model's own long-share divided by its short-share; the field average of the per-model ratios is 0.62 this week. 'Sideways' was 52.0% of mature forecasts (4,833 of 9,301). Consensus skew: 533 symbol/FH/TF cells had 5 or more models on the same side and hit 38.1% (n=3,347) against the 39.6% field base.

| Model | Symbol | Side | Conf | FH / TF | Slot (UTC) | Exit | Net PnL |
|---|---|---|---|---|---|---|---|
| gpt-5.6-sol | BTC | Short | 82 | 1h / 1h | Sep 8, 07:01 | EXPIRY | -$0.23 |

**Fail of the week.** The week's single highest-confidence individual miss, exit reason expiry. Listed as a single card, never as a model ranking.

## Calibration bridge

The hit-rate leader this week is gemini-3.1-pro (41.8%); the best-calibrated line is qwen-3.8-max (Brier 0.2794, gap +18.8pp). Per-model ok-rate this window ran 99.59% to 100.00%. The full confidence-bucket analysis lives in Weekly Calibration #6.

## Market check: a down week on a quieter tape

> **MARKET CHECK · OBSERVATION (two adjacent calendar weeks)**
>
> OBSERVATION — mean |1d move| 2.10% -> 1.35% (-36% rel), BTC realized vol 35.2% -> 31.4% (ann., hourly); field directional accuracy 42.8% -> 39.6% (-3.3pp), 1 of 7 models improved, sim win-rate up for 1 of 7, field sim PnL -$669.03 -> -$1018.93 (adjacent calendar weeks Aug 31-Sep 6 vs Sep 7-13; descriptive, one pair of weeks, not a claim).
>
> Robustness: raw price-sign accuracy 51.6% -> 38.7% — the same direction as the trade-based hit rule.

| Measure | Aug 31-Sep 6 (week A) | Sep 7-13 (week B) | Change |
|---|---|---|---|
| Mean \|1d move\| (5 assets) | 2.10% | 1.35% | -36% rel |
| BTC realized vol (ann., hourly) | 35.2% | 31.4% | -3.8pp |
| Field directional accuracy | 42.8% | 39.6% | -3.3pp |
| Models improving hit-rate | -- | 1 of 7 | -- |
| Raw price-sign accuracy | 51.6% | 38.7% | -12.8pp |
| Field sim win-rate | 35.6% | 32.1% | -3.5pp |
| Field sim net PnL | -$669.03 | -$1018.93 | -- |

Week A is exactly the issue-#5 window, so this slice and every week-over-week column in this report compare the same two back-to-back weeks. Descriptive, one pair of weeks; the hit rule is methodology v1.1 (trade-based) and the raw price-sign row is the robustness check (hourly closes at :00 against slots at :01). BTC realized vol here is the hourly-return series from the alive slice; the snapshot row's figure is the audited daily-candle estimate -- different estimators, both reported as measured. Lineage splice ACTIVE: grok-4.5 rows are aggregated into grok-4.6 (flip 2026-08-24T09:22:00+00:00, two windows back, inside issue #4's window). Week A (2026-08-31..2026-09-06) is pure grok-4.6 (it reproduces the published issue #5); week B is pure grok-4.6.

## Practical implications

- Do not read this week's leaderboard as a skill ranking: no title was awarded and the top CIs overlap.
- Treat per-ticker readability as weekly weather. Six issues, and the best ticker changed hands in every one until this week (SOL -> BNB -> BTC -> SOL -> BNB -> BNB): BNB led issue #5 and leads again at 46.9%.
- Self-agreement lifts run -6.0pp to +0.3pp with small disagree cells; carry the regime split into the monthly test, do not act on it.

## Limitations

- Single week (Sep 7-13, 2026 UTC). Descriptive for this window only; no claim about next week or any model's underlying skill. This window ran 2 trend days of 7 (issue #5: 3 of 7); every comparison with issue #5 is a comparison across regimes as well as across weeks.
- Observations inside one window are not independent, and the 95% Wilson CIs shown throughout are descriptive, not inferential.
- The reversal detector remains v1 and BTC/ETH-only, on price series reconstructed from trade entry prices. 12 events across six issues cannot characterise detector performance either way.
- Cross-TF and cross-FH disagreement cells are small (n=10, 5, 6, 8); cells under N=10 are insufficient and never used to rank models. Week-over-week deltas are computed on unrounded rates and may differ by 0.1pp from the difference of the rounded columns.
- Model lines in this window: claude-fable-5, claude-opus-5, deepseek-v4-pro, gemini-3.1-pro, gpt-5.6-sol, grok-4.6, qwen-3.8-max. Series density and lineage notes (wave 6): (1) rolling-1w forecasts emit daily since Aug 22, 2026 and rolling-1M daily since Aug 23, 2026, so inside this window (Sep 7-13) daily coverage is FULL, as it was in issues #4 and #5: the 1w series has slots on 7 of 7 days and the 1M series on 7 of 7 days; both sit outside this report's FH gate (1h/4h/1d) and appear only in the exclusions counter (1w 490, 1M 490). (2) The grok line flipped 4.5 -> 4.6 at Aug 24, 2026 09:22 UTC, two windows back (inside the issue-#4 window): this window carries 1,470 grok rows and none from the 4.5 era, and neither does week A (the issue-#5 window), so every grok number in this issue and in issue #5 is a pure grok-4.6 line -- the grok week-over-week row is the first pure-4.6 against pure-4.6 comparison of the series. (3) qwen-3.8-max succeeded qwen-3.7-max on Aug 5, 2026 (fully before this window).
- Market-state row is computed on a single-exchange (binance) daily candle series, as in issue #5 after the audit that found the raw table mixes two exchanges; issue #5 values are as published.
- Research-to-date counter: directional forecasts resolved inside the FH gate (1h/4h/1d) since Jul 11, counted at the issue cutoff (Sep 14 16:00 UTC) -- the definition pinned in issue #3 and carried by issues #4 and #5, which printed 42,905 at the Sep 7 16:00 cutoff. The pack reproduces that pin at the previous cutoff (control OK), so this issue's 47,373 is an additive step under one definition; counter deltas against issue #4 and earlier remain definitional.

## Counters & lineage

| Counter | Value | Counter | Value |
|---|---|---|---|
| Forecasts total | 10,290 | Mature (scored pool) | 9,301 |
| OK in gate | 9,301 | -- of them directional | 4,468 |
| Out of gate (1w / 1M) | 490 / 490 | -- of them sideways | 4,833 |
| Invalid | 9 | Pending (next issue) | 0 |
| Late closes | 0 | Uptime, all grids | 266 / 266 slots |
| Source file | weekly_metrics_2026-09-07.json | Report cutoff | Mon Sep 14, 2026, 16:00 UTC |
| Generated at (pipeline) | 2026-09-15 15:24 UTC | Methodology | v1.1 (2026-08-10) · hash e66c7e8c864a2233 |

**THIS REPORT: 4,468 scored observations.** MARKETMANIA RESEARCH TO DATE (as of Sep 14, 2026 cutoff): 47,373 directional forecasts resolved since Jul 11 · 7 models tracked (current line-up; earlier versions folded into their successors' lineage) · 5 assets · 5 horizons · hourly · 29 published reports.

> **Issue #6.** Weekly Model Watch is a living series. Issue #5's open questions -- does any model hold the top rank two issues running and does the title rule ever fire, and does the reversal detector find callers at all, on a non-ETH turn? -- read this issue as: gemini-3.1-pro at the top (41.8%), the line that also led issue #5 (gemini-3.1-pro), so the rank held two issues running; no weekly title was awarded, for the sixth issue running, and 1 of 2 qualifying turns found a caller, on turns in ETH only. The grok row is pure grok-4.6 this issue and was pure grok-4.6 in issue #5; the 4.5 -> 4.6 flip (Aug 24, 2026 09:22 UTC) sits two windows back. Engine 1.1 has powered the sandbox since Aug 18, i.e. before this window; no cross-engine PnL comparisons are claimed.

## What we're testing next

- Next issue: does gemini-3.1-pro hold the top rank for a third issue, and does the title rule ever fire?
- Does a qualifying reversal ever fire outside ETH, and does a second turn in the same pair ever draw a caller?
- Monthly series: cross-TF/FH lifts with regime control across mixed weeks (trend vs. flat), block bootstrap.

## Related research

| Report | Direct PDF link |
|---|---|
| Consensus Watch #6 | https://marketmania.ai/research/reports/consensus-watch-2026-09-07.pdf |
| Weekly Calibration #6 | https://marketmania.ai/research/reports/weekly-calibration-2026-09-07.pdf |
| Weekly Model Watch #5 | https://marketmania.ai/research/reports/model-watch-2026-08-31.pdf |

The three weekly reports publish together as one issue each week; each links straight to the others' PDF and to its own previous issue. Direct links are the posting rule from wave 2 on.

## Cite this report

```bibtex
@misc{mm_modelwatch_2026w37,
  title  = {Weekly Model Watch #6: weekly leaderboard, ticker/pair reads and cross-confirmation, Sep 7-13 2026},
  author = {{MarketMania Research}},
  year   = {2026}, month = {September}, day = {15},
  url    = {https://marketmania.ai/research/reports/model-watch-2026-09-07.pdf},
  note   = {Methodology v1.1 (2026-08-10), hash e66c7e8c864a2233; window Sep 7-13, 2026 UTC; source weekly_metrics_2026-09-07.json}
}
```