drjones
|
b1c871c198
|
fix: knife-catch gate (RSI<8 + downtrend = SKIP), BTC macro gate both directions. YES bets were 15% WR losing -4. Root cause: buying dips in downtrend.
|
2026-08-03 02:30:28 -07:00 |
|
drjones
|
0427f015bc
|
fix: all-local 2-model ensemble (MacBook qwen + GamingPC ornith), no containers, disagreement=SKIP, tight RSI gate with momentum confirmation
|
2026-08-03 02:02:48 -07:00 |
|
drjones
|
820c542d89
|
session complete: 8-coin fleet, tight RSI gate, watchdog test 70/72, 33/35 unit
|
2026-08-03 00:26:36 -07:00 |
|
drjones
|
f3d9cafbab
|
untrack DBs/logs, add new coin configs
|
2026-08-03 00:26:18 -07:00 |
|
drjones
|
4006d52ba0
|
gitignore + watchdog fixes
|
2026-08-03 00:25:57 -07:00 |
|
drjones
|
7e6b2dd1ca
|
watchdog test script + log filter fix
|
2026-08-03 00:25:30 -07:00 |
|
drjones
|
06ff6e8c76
|
session complete: 3-model ensemble, exposure cap, rebalance, pattern matching, dashboards, docs
|
2026-08-02 22:36:42 -07:00 |
|
drjones
|
32cfdfe842
|
feat: 3-model ensemble (qwen3.5+ornith+llama3.2), WR-based stake rebalancing, 25% exposure cap, rebalance ceiling .50, exposure throttling
|
2026-08-02 21:51:00 -07:00 |
|
drjones
|
0408ff026f
|
fix: Kalshi market data nested under 'market' key — resolve_bets now correctly settles orders. fix: BTC Binance proxy. All 8 data sources verified. PNL tracking live.
|
2026-08-02 20:27:32 -07:00 |
|
drjones
|
c817216ab1
|
feat: order management (cancel/amend), stale order cleanup before market close, enriched LLM reasoning (RSI+BTC+regime), fleet dashboard with SVG charts/gauges/donuts
|
2026-08-02 20:10:24 -07:00 |
|
drjones
|
8e860361ec
|
feat: 5-coin fleet (DOGE/SOL/ETH/XRP/BTC), enriched context (multi-TF trend, orderbook depth, BTC corr, WR history, regime), memory system (learned patterns), self-adapting overrides, fleet dashboard with per-coin RSI/mom
|
2026-08-02 19:25:58 -07:00 |
|
drjones
|
a4def4f72e
|
docs: README with architecture, data inputs, decision engine, infrastructure map
|
2026-08-02 19:13:18 -07:00 |
|
drjones
|
c14bb98f7f
|
K4LSH1_OPS: multi-coin LLM bot — DOGE/SOL/ETH, two-tier qwen+ornith, self-adapting
|
2026-08-02 19:10:27 -07:00 |
|