branch fix/capital-reconcile-all-bots commit 6ee796d built last · auto: refresh dashboard data (2026-07-03 06:20)

Live status refreshed at commit time (2026-06-01 08:03 Madrid) · live capital total €107.28

Note: capital values reflect what each bot has IN ITS sub-account at the last 15min bar. V2 has a state file (precise to 6 decimals); V3-V6 read from log lines (integer dollars). Trade count = realized round-trip trades since deploy.
Bot Capital Peak Trades State Regime DRSI Cooldown Last BTC Last bar Live
V2 €22.17 €23.32 1 INVESTED $73296 2026-06-01 06:00:00 47d
V3 €23.00 0 CASH $73294 2026-06-01 06:00:00 47d
V4 €22.56 €23.57 0 INVESTED $73294 2026-06-01 06:00:00 47d
V5 €18.81 €20.16 0 INVESTED $73294 2026-06-01 06:00:00 42d
V6 €20.75 €20.75 0 INVESTED $73294 2026-06-01 06:00:00 42d

System architecture · GRU 2×128 ensemble (V2-V6 shared)

INPUT 36 features × 300 event-bars (lb=300) NORM per-sequence MinMax 5-seed GRU ensemble 2 layers × 128 hidden · dropout 0.2 · seeds: 42, 123, 456, 789, 1024 seed_42 → P(SL), P(timeout), P(TP) seed_123 → P(SL), P(timeout), P(TP) seed_456 → P(SL), P(timeout), P(TP) seed_789 → P(SL), P(timeout), P(TP) seed_1024 → P(SL), P(timeout), P(TP) VOTE ≥3/5 agree on danger P(TP)<0.22 ∨ P(SL)>0.70 STRATEGY YAML (per-bot config) V2 ensemble V3 + DRSI>80 V4 adaptive V5 V115_cmp ★ V6 V66 cd(4,48) ★ exit cascade: init_sl → trail → DRSI → GRU danger → peak_drop All 5 bots share the SAME 20 GRU checkpoints (5 seeds × 4 folds). Differences are policy, not model.

All 5 production bots share one GRU 2×128 ensemble. The diff between V2 and V6 is only YAML config.

V2 · GRU Ensemble (always-invested)

backtest config: config/production_v2.yaml

+7,641% compound naked model baseline — no daily filter
EDGE THESIS

Raw 5-seed GRU vote, always-invested. Exits only when ≥3/5 models flag danger. The 'naked model' baseline — useful as a control for V3-V6 attribution: any of them that doesn't beat V2 is not earning its complexity.

WEAKNESS — DON'T COPY IF

Time-in-market % — should be ~95%+. Also: min α per fold (undocumented). If min α ≤ B&H on any fold, the whole GRU thesis is fragile.

LIVE capital
€22.17
LIVE trades
1
LIVE position
INVESTED
Days running
47d
Backtest compound
Backtest min α

V3 · GRU + DailyRSI>80 + Trail 10%

backtest config: config/production_v3_hybrid.yaml

+450 billion% compound regime-dependent — likely path-lucky
EDGE THESIS

GRU + DailyRSI>80 overheat exit + 10% trailing. Designed to surf bulls and step aside on parabolic exhaustion. The astronomical compound number is misleading without fold attribution.

WEAKNESS — DON'T COPY IF

Single-fold contribution share. If >70% of compound comes from one fold, this strategy is one regime away from being worthless. Show the fold decomposition explicitly.

LIVE capital
€23.00
LIVE trades
0
LIVE position
CASH
Days running
47d
Backtest compound
Backtest min α

V4 · GRU 3-regime adaptive (V5.4 Robust-5)

backtest config: config/production_v4_hybrid_v5.yaml

+117% min α first per-regime mechanism
EDGE THESIS

Robust-5 adaptive ATR exits keyed to bull/bear/lateral regime. The first strategy in the family with a real per-regime mechanism rather than uniform parameters. R095 confirmed +53% improvement via adaptive exit on this bot.

WEAKNESS — DON'T COPY IF

Lateral-regime PnL alone. V4's whole bet is regime detection; if its lateral sub-period PnL is < V6's lateral sub-period PnL, the regime mechanism is cosmetic.

LIVE capital
€22.56
LIVE trades
0
LIVE position
INVESTED
Days running
47d
Backtest compound
Backtest min α

V5 · V115_cmp (GRU + peak_drop + ratchet + regime cooldown)

backtest config: config/production_v5_hybrid_v115.yaml

+168,759% compound THE COMPOUND KING
EDGE THESIS

GRU + peak_drop trail + ratchet + per-regime cooldown (greedy bull/bear, 30h punishment lateral). Squeezes maximum compounding from trending phases by re-entering aggressively after pullbacks. Reproduces external agent's V115 result at the dollar.

WEAKNESS — DON'T COPY IF

Min α +238% (vs V6's +245%). V5 trades 7pp of consistency for compound. MDD per fold — if any fold MDD > 30%, V5 may not be tradable at scale.

LIVE capital
€18.81
LIVE trades
0
LIVE position
INVESTED
Days running
42d
Backtest compound
Backtest min α

V6 · V66 Cooldown(4,48) — uniform thresholds, extreme cooldown

backtest config: config/production_v6_hybrid_v66.yaml

+245% min α THE CONSISTENCY KING · CURRENT CHAMPION
EDGE THESIS

Same GRU, uniform thresholds, extreme cooldown (4 bars + 48 extra after loss, all regimes). Optimized for the WORST fold, not the best. The benchmark every R&D candidate must beat. Compound +70,576% is the 'price' paid for the +245% min α record.

WEAKNESS — DON'T COPY IF

Opportunity cost in mega-bulls. V6 will systematically underperform V5 and B&H in raging bulls. If you want an 'always wins' strategy, V6 is not it.

LIVE capital
€20.75
LIVE trades
0
LIVE position
INVESTED
Days running
42d
Backtest compound
Backtest min α

V7 · V5 V115_cmp + vol-Kelly continuous sizing (margin)

backtest config: config/production_v7_vol_kelly.yaml

✅ LIVE since 2026-05-31 ★ FIRST MARGIN BOT IN PRODUCTION · paired with V3 sub-account
EDGE THESIS

Identical V5 cascade for entry/exit. NEW: position size scales with inverse 30-day volatility via Binance USDC borrow. leverage = clip(0.030 / vol_30d_shifted, 0.5, 1.5). Auto-deleverages in high-vol regimes (slip protection). R203 buggy +981K confirmed by R204 honest at 102% retention. R217b confirms V15 aggressive variant (clip[0.5,2.0]) gives +8.8M%/+919 with 0 liquidations — future upgrade candidate after V7 paper trade. V7 inherited 0.0002 BTC + $7.26 USDC as synthetic entry from V3 margin wallet at $73,572.

WEAKNESS — DON'T COPY IF

Margin interest cost (~0.45% APR). Liquidation risk in flash crashes — backtest 0 liquidations across 20 stress scenarios but real-world tail untested. Production uses cross margin in V3's sub-account (not V5 — V3 had pre-funded USDC from May-28 test trades).

LIVE capital
LIVE trades
0
LIVE position
?
Days running
0d
Backtest compound
Backtest min α

V8 · LLM Hybrid Trader (V5 cascade + Opus committee filter)

backtest config: PENDING — production_v8_llm_hybrid.yaml (creation blocked by hook)

PENDING SUB-ACCOUNT scaffold complete, awaiting llm_hybrid Binance sub-account
EDGE THESIS

V5 V115_cmp generates base signal. Every 2h an LLM committee (Opus orchestrator + 3 Sonnet specialists: technical/macro/risk) views multi-timeframe BTC charts (15m/1h/4h/1d with SMA/EMA/BB/RSI) and VETOES entries it finds unsafe. Adds vision-based pattern recognition the GRU lacks.

WEAKNESS — DON'T COPY IF

API cost ~$3-4/day. Committee can FALSE-NEGATIVE good GRU entries. Latency budget — committee cycle must complete in <90s or block bar processing.

LIVE capital
LIVE trades
0
LIVE position
?
Days running
0d
Backtest compound
Backtest min α

V9 · LLM Pure Trader (no GRU, committee decides directly)

backtest config: PENDING — production_v9_llm_pure.yaml

PENDING SUB-ACCOUNT scaffold complete, awaiting llm_pure Binance sub-account
EDGE THESIS

NO GRU model. Pure LLM committee every 2h drives ENTER/HOLD/EXIT decisions. Position size scales with committee confidence × kelly_fraction_max (0.5). The most experimental of the 3 candidates.

WEAKNESS — DON'T COPY IF

Variance — only the LLM filters. No GRU to disagree. Each cycle costs ~$0.30. Slow cadence (every 2h) means missed micro-moves.

LIVE capital
LIVE trades
0
LIVE position
?
Days running
0d
Backtest compound
Backtest min α