# Weekly Model Watch #5

**WEEKLY · MODEL WATCH** · September 8, 2026 · MarketMania Research · Weekly series

Window: **Aug 31-Sep 6, 2026 UTC** · Cutoff: **Mon Sep 7, 2026, 16:00 UTC** · Methodology **v1.1 (2026-08-10)**, hash `e66c7e8c864a2233`

PDF: https://marketmania.ai/research/reports/model-watch-2026-08-31.pdf · Open data (JSON): https://marketmania.ai/research/reports/model-watch-2026-08-31.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 |  |  |  |
|---|---|---|---|
| **4,520** directional calls scored | **7** models (stable lineup) | of **9,264** mature forecasts | window **Aug 31-Sep 6, 2026 UTC** |
| hit = direction rule: tp1/tp2->hit, sl->miss, expiry->sign of gross pnl | FH gate **1h / 4h / 1d** | base field hit **42.8%** | cutoff **Mon Sep 7, 2026, 16:00 UTC** |
| market: BTC net **+3.42%** (prior -0.07%) | ann vol **43.4%** (was 30.9%) | TOP5 volume **$16.24B**, -18% w/w | pairwise corr **0.78** (was 0.85) |

## Key finding

> **KEY FINDING · OBSERVATION (one weekly window)**
>
> ### Field accuracy went 42.9% -> 42.8% (-0.1pp) with 1 of 7 models improving, and no weekly title was awarded, for the fifth issue running.

## TL;DR

- **OBSERVATION -- No weekly title again.** gemini-3.1-pro leads descriptively at 46.5% [43.0%, 50.1%]; the gap to #2 (deepseek-v4-pro, 44.2%) is 2.3pp. 1 of 7 models improved week over week; issue #4's leader deepseek-v4-pro is now 2 of 7.
- **No ticker cleared 51%.** BNB is the week's most readable coin (47.3%, n=916) and SOL the hardest at 38.7%; 0 of the 5 assets cleared 51%. Pair of the week: deepseek-v4-pro x BNB 53.5% (n=144); worst: grok-4.6 x SOL 29.5% (n=105).
- **0 of 2 qualifying turns found a caller.** ETH long +4.61%, 0 caller(s), first NOBODY; ETH short -2.13%, 0 caller(s), first NOBODY.
- **Self-agreement read as a penalty in every cell.** Cross-TF/FH lifts ran -17.4pp to -2.5pp (issue #4: -7.6pp to -1.9pp). On trend days, 4h TF-agreement hit 33.6% (n=214) against 36.8% (n=247) 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 42.9% -> 42.8%. The grok column compares this window's pure 4.6 line against issue #4's spliced 4.5+4.6 line. "Was" values are the numbers published in issue #4; 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. Issue #5's grok row is a pure grok-4.6 week while issue #4's was a lineage splice, 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 inside the PREVIOUS window (week A, issue #4): every issue-#4 grok number is the lineage splice "grok-4.6 (incl. 4.5-era, Aug 24 00:00-09:22)"; this window is pure grok-4.6, so the grok week-over-week row compares a pure-4.6 week with a spliced week.

## Leaderboard of the week

| Model | N | Cov. | Hit rate | 95% CI | W/w pp |
|---|---|---|---|---|---|
| gemini-3.1-pro | 761 | 57.3% | 46.5% | 43.0%-50.1% | +2.3 |
| deepseek-v4-pro | 724 | 54.6% | 44.2% | 40.6%-47.8% | -0.4 |
| claude-fable-5 | 628 | 47.2% | 42.8% | 39.0%-46.7% | -0.2 |
| qwen-3.8-max | 697 | 53.5% | 42.5% | 38.9%-46.2% | -0.1 |
| claude-opus-5 | 545 | 41.1% | 42.4% | 38.3%-46.6% | +0.0 |
| gpt-5.6-sol | 689 | 51.8% | 41.6% | 38.0%-45.4% | -1.0 |
| grok-4.6 | 476 | 36.0% | 37.6% | 33.4%-42.0% | -2.3 |

No weekly title is awarded this issue -- for the fifth 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 2.3pp and their CIs overlap, so neither condition is met. The grok 4.5 -> 4.6 flip (Aug 24, 2026 09:22 UTC) sits inside the PREVIOUS window (week A, issue #4): every issue-#4 grok number is the lineage splice "grok-4.6 (incl. 4.5-era, Aug 24 00:00-09:22)"; this window is pure grok-4.6, so the grok week-over-week row compares a pure-4.6 week with a spliced week.

## Ticker of the week

| Symbol | N | Hit rate | 95% CI | Issue #4 |
|---|---|---|---|---|
| BNB | 916 | 47.3% | 44.1%-50.5% | 41.2% |
| XRP | 909 | 44.2% | 41.0%-47.5% | 40.3% |
| ETH | 982 | 43.1% | 40.0%-46.2% | 43.3% |
| BTC | 780 | 40.6% | 37.2%-44.1% | 43.7% |
| SOL | 933 | 38.7% | 35.6%-41.9% | 45.8% |

BNB was the field's most readable ticker (47.3%) and SOL the hardest (38.7%); 0 of 5 assets cleared 51% this week (issue #4: 0 of 5). Issue #4's best ticker was SOL and its hardest XRP -- this issue's ordering is BNB > XRP > ETH > BTC > SOL.

## 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 |
|---|---|---|---|
| deepseek-v4-pro | BNB | 144 | 53.5% |
| claude-opus-5 | BNB | 112 | 49.1% |
| claude-fable-5 | BNB | 127 | 48.0% |
| gemini-3.1-pro | BNB | 155 | 47.7% |
| deepseek-v4-pro | ETH | 164 | 46.9% |
| gemini-3.1-pro | XRP | 147 | 46.9% |

| Model (5 worst) | Symbol | N | Hit rate |
|---|---|---|---|
| claude-opus-5 | SOL | 112 | 36.6% |
| deepseek-v4-pro | SOL | 148 | 36.5% |
| grok-4.6 | BTC | 70 | 35.7% |
| grok-4.6 | ETH | 99 | 35.4% |
| grok-4.6 | SOL | 105 | 29.5% |

deepseek-v4-pro x BNB (53.5%, n=144) is the week's best pair and grok-4.6 x SOL (29.5%, n=105) the weakest. Pair history accumulates across issues -- read this as the fifth 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 | Thu Sep 3, 08:00 | Long | +4.61% | 0 | NOBODY |
| ETH | Fri Sep 4, 08:00 | Short | -2.13% | 0 | NOBODY |

The ETH long turn drew NOBODY; the ETH short turn drew NOBODY. Scope in v1 remains BTC/ETH only; the event count across five 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 | 461 | 35.4% | 10 | 90.0% | 39.9% (651) | -4.6 |
| cross-TF | 1d | TF 1d vs 4h | 67 | 28.4% | 12 | 83.3% | 43.8% (128) | -15.4 |
| cross-FH | 4h | by FH 1h | 401 | 37.4% | 20 | 75.0% | 39.9% (651) | -2.5 |
| cross-FH | 1d | by FH 4h | 57 | 26.3% | 12 | 83.3% | 43.8% (128) | -17.4 |

* = N<10 (insufficient); insufficient cells are never used to rank models. Mixed pairs (a model with only one side of the comparison present): 177 / 49 / 227 / 59. Pooled across all four cell families the lifts run -17.4pp to -2.5pp this week (issue #4: -7.6pp to -1.9pp).

**Regime control.** Methodology calls for a trend-vs-flat split on this table. This window has 3 trend days of 7: 08-31 +1.19%, 09-01 -1.50% (trend), 09-02 -0.06%, 09-03 +4.98% (trend), 09-04 -1.97% (trend), 09-05 +0.24%, 09-06 +0.58%. On trend days, 4h TF-agreement hit 33.6% (n=214) against 36.8% (n=247) on flat days.

## Side-mix and the week's failure

| Model | N | Long | Short | Sideways | L:S |
|---|---|---|---|---|---|
| claude-fable-5 | 1,330 | 32.6% | 14.6% | 52.8% | 2.24 |
| claude-opus-5 | 1,326 | 29.5% | 11.6% | 58.9% | 2.54 |
| deepseek-v4-pro | 1,325 | 31.2% | 23.4% | 45.4% | 1.34 |
| gemini-3.1-pro | 1,329 | 31.6% | 25.7% | 42.7% | 1.23 |
| gpt-5.6-sol | 1,330 | 26.4% | 25.4% | 48.2% | 1.04 |
| grok-4.6 | 1,322 | 15.4% | 20.6% | 64.0% | 0.74 |
| qwen-3.8-max | 1,302 | 34.0% | 19.5% | 46.5% | 1.74 |

L:S = each model's own long-share divided by its short-share; the field average of the per-model ratios is 1.55 this week. 'Sideways' was 51.2% of mature forecasts (4,744 of 9,264). Consensus skew: 521 symbol/FH/TF cells had 5 or more models on the same side and hit 41.3% (n=3,248) against the 42.8% field base.

| Model | Symbol | Side | Conf | FH / TF | Slot (UTC) | Exit | Net PnL |
|---|---|---|---|---|---|---|---|
| gemini-3.1-pro | BTC | Long | 80 | 4h / 4h | Sep 4, 12:01 | SL | -$1.60 |

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

## Calibration bridge

The hit-rate leader this week is gemini-3.1-pro (46.5%); the best-calibrated line is qwen-3.8-max (Brier 0.2755, gap +16.6pp). Per-model ok-rate this window ran 98.10% to 100.00%. The full confidence-bucket analysis lives in Weekly Calibration #5.

## Market check: an up week on lower volume

> **MARKET CHECK · OBSERVATION (two adjacent calendar weeks)**
>
> OBSERVATION — mean |1d move| 2.35% -> 2.10% (-11% rel), BTC realized vol 37.9% -> 35.2% (ann., hourly); field directional accuracy 42.9% -> 42.8% (-0.1pp), 1 of 7 models improved, sim win-rate up for 1 of 7, field sim PnL -$473.59 -> -$669.03 (adjacent calendar weeks Aug 24-30 vs Aug 31-Sep 6; descriptive, one pair of weeks, not a claim).
>
> Robustness: raw price-sign accuracy 44.2% -> 51.6% — the opposite direction to the trade-based hit rule.

| Measure | Aug 24-30 (week A) | Aug 31-Sep 6 (week B) | Change |
|---|---|---|---|
| Mean \|1d move\| (5 assets) | 2.35% | 2.10% | -11% rel |
| BTC realized vol (ann., hourly) | 37.9% | 35.2% | -2.7pp |
| Field directional accuracy | 42.9% | 42.8% | -0.1pp |
| Models improving hit-rate | -- | 1 of 7 | -- |
| Raw price-sign accuracy | 44.2% | 51.6% | +7.4pp |
| Field sim win-rate | 37.0% | 35.6% | -1.4pp |
| Field sim net PnL | -$473.59 | -$669.03 | -- |

Week A is exactly the issue-#4 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, inside week A). Week A (2026-08-24..2026-08-30) mixes the 4.5-era (Aug 24 00:00-09:22) with 4.6 and reproduces the published issue #4; 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. Five issues, and the best ticker has changed hands every time (SOL -> BNB -> BTC -> SOL -> BNB).
- Self-agreement lifts run -17.4pp to -2.5pp with small disagree cells; carry the regime split into the monthly test, do not act on it.

## Limitations

- Single week (Aug 31-Sep 6, 2026 UTC). Descriptive for this window only; no claim about next week or any model's underlying skill. This window ran 3 trend days of 7 (issue #4: 3 of 7); every comparison with issue #4 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. 10 events across five issues cannot characterise detector performance either way.
- Cross-TF and cross-FH disagreement cells are small (n=10, 12, 20, 12); 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 5): (1) rolling-1w forecasts emit daily since Aug 22, 2026 and rolling-1M daily since Aug 23, 2026, so inside this window (Aug 31-Sep 6) daily coverage is FULL, as it was in issue #4 (the first full week of the series): 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, inside the PREVIOUS window (week A, issue #4): this window carries 1,470 grok rows and none from the 4.5 era, so every grok number in this issue is a pure grok-4.6 line, while every issue-#4 grok number is the lineage splice "grok-4.6 (incl. 4.5-era, Aug 24 00:00-09:22)" -- the grok week-over-week row compares a pure-4.6 week with a spliced week. (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 #4 after the audit that found the raw table mixes two exchanges; issue #4 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 7 16:00 UTC) -- the definition pinned in issue #3 and carried by issue #4, which printed 38,385 at the Aug 31 16:00 cutoff. The pack reproduces that pin at the previous cutoff (control OK), so this issue's 42,905 is an additive step under one definition; counter deltas against issue #3 and earlier remain definitional.

## Counters & lineage

| Counter | Value | Counter | Value |
|---|---|---|---|
| Forecasts total | 10,290 | Mature (scored pool) | 9,264 |
| OK in gate | 9,264 | -- of them directional | 4,520 |
| Out of gate (1w / 1M) | 490 / 490 | -- of them sideways | 4,744 |
| Invalid | 46 | Pending (next issue) | 0 |
| Late closes | 0 | Uptime, all grids | 266 / 266 slots |
| Source file | weekly_metrics_2026-08-31.json | Report cutoff | Mon Sep 7, 2026, 16:00 UTC |
| Generated at (pipeline) | 2026-09-08 08:08 UTC | Methodology | v1.1 (2026-08-10) · hash e66c7e8c864a2233 |

**THIS REPORT: 4,520 scored observations.** MARKETMANIA RESEARCH TO DATE (as of Sep 7, 2026 cutoff): 42,905 directional forecasts resolved since Jul 11 · 7 models tracked (current line-up; earlier versions folded into their successors' lineage) · 5 assets · 5 horizons · hourly · 25 published reports.

> **Issue #5.** Weekly Model Watch is a living series. Issue #4'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 outside ETH? -- read this issue as: gemini-3.1-pro at the top (46.5%) where issue #4 had deepseek-v4-pro, no weekly title was awarded, for the fifth issue running, and 0 of 2 qualifying turns found a caller, on turns in ETH only. The grok row is pure grok-4.6 this issue; issue #4's grok row was the lineage splice "grok-4.6 (incl. 4.5-era, Aug 24 00:00-09:22)". 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 any model hold the top rank two issues running, and does the title rule ever fire?
- Does the reversal detector find callers at all, and on a non-ETH turn?
- 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 #5 | https://marketmania.ai/research/reports/consensus-watch-2026-08-31.pdf |
| Weekly Calibration #5 | https://marketmania.ai/research/reports/weekly-calibration-2026-08-31.pdf |
| Weekly Model Watch #4 | https://marketmania.ai/research/reports/model-watch-2026-08-24.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_2026w36,
  title  = {Weekly Model Watch #5: weekly leaderboard, ticker/pair reads and cross-confirmation, Aug 31-Sep 6 2026},
  author = {{MarketMania Research}},
  year   = {2026}, month = {September}, day = {8},
  url    = {https://marketmania.ai/research/reports/model-watch-2026-08-31.pdf},
  note   = {Methodology v1.1 (2026-08-10), hash e66c7e8c864a2233; window Aug 31-Sep 6, 2026 UTC; source weekly_metrics_2026-08-31.json}
}
```