Commit Graph

9 Commits

Author SHA1 Message Date
tinkle-community
eabd279d10 feat: show hold duration in the recent trades panel
Adds a HOLD column (45m / 2h10 / 1d3h) plus a header row to the dashboard's
recent-closes table. Makes the hold-time-vs-PnL pattern (short holds bleed,
1h+ holds carry the edge) visible per trade instead of only in the aggregate
Edge Profile panel.
2026-07-15 19:36:58 +09:00
tinkle-community
8c8cd9b61f feat: show entry price and notional size in the live positions panel
The panel only showed symbol/side/leverage/PnL — entry and position size
are the first things a trader checks. Side and leverage share a column to
make room; prices format with magnitude-aware precision (64,187 · 184.2 ·
0.0067); size is |qty| × mark notional.
2026-07-10 18:56:15 +09:00
tinkle-community
390db72cdb feat: remove the last beginner walls — USDC sourcing help, reachable deposit page, first-run reassurance
- The /welcome deposit page (QR + auto-refreshing balance + key backup) was
  gated behind a legacy 'beginner mode' localStorage flag that nothing sets
  anymore — every deposit link silently bounced to /traders. Gate removed:
  any signed-in user reaches it.
- 'Don't have USDC yet?' helper on the deposit page: buy on Binance/OKX/
  Coinbase, withdraw picking the Base network, arrives in ~1 minute. This is
  the wall every true newcomer hits first.
- Autopilot step 3 now says exactly where trading money goes
  (app.hyperliquid.xyz → Deposit, USDC on Arbitrum).
- Register page: 'NEW_USER ONBOARDING / Initializing Registration Sequence'
  jargon replaced with what actually happens next (guided launch, ~$13,
  five minutes).
- Dashboard shows a calm first-run note while a fresh autopilot runs its
  first cycle ('reading the whole market — first decision lands in a minute
  or two'), so the initial silence doesn't read as broken.
- Hyperliquid connect steps renamed to plain language ('Approve it in your
  wallet (trade-only, cannot withdraw)' etc.).
2026-07-07 23:03:20 +09:00
tinkle-community
c53a563bff feat: data-driven autopilot profitability tuning + edge profile panel
Analysis of 410 live closed trades found the edge is real but was being
destroyed by execution: gross +$267 vs $245 fees; trades held <1h were net
negative (the <15m bucket alone: -$48 on $66 fees) while 1h+ holds carried
+$78; shorts lost $72 while longs made $94 — and both the prompt and the
engine were manufacturing those losers.

Changes, each tied to the data:
- Prompt: removed the 'MUST open at least one long AND one short every
  cycle' mandate (forced weak-signal shorts); direction is now data-driven
  with 'never open just to balance the book'. Added fee-awareness (round
  trip ≈ 0.1% notional, require expected move ≥ 3x cost) and aligned the
  hold discipline with the backend throttle.
- Forced book-balance opens now require |board z-score| ≥ 0.75 — the engine
  previously force-opened full-size 10x positions on near-neutral signals
  with hardcoded confidence 70. DirectionalCandidates now carries scores.
- Min AI-managed hold raised 45m → 60m (the 15-60m bucket still bled after
  the earlier throttle landed).
- Legacy prompt hygiene: vergex path drops long-only-era custom prompts and
  zh-era configs fall back wholesale to built-in English sections — fixes
  the two long-failing kernel prompt tests.
- New Edge Profile dashboard panel: net after fees by hold-time bucket and
  side, computed from recent closed trades, with an automatic takeaway line
  — the fee/churn regression detector, always visible. Fixed the
  HistoricalPosition timestamp types (epoch ms, not strings).
2026-07-07 19:21:49 +09:00
tinkle-community
026c5d3089 fix: dashboard metrics — real drawdown baseline, realized/unrealized split, risk radar fields
- Max drawdown was double-broken: the backend built the equity curve on a
  hardcoded $10k baseline (understating a small account's drawdown ~20x) and
  the frontend multiplied the already-percent value by 100 again (0.87% shown
  as -86.9%). The curve now starts from the trader's real initial balance
  (10k fallback when unknown) and the UI renders the percent once; the demo
  engine and type docs are aligned to percent semantics.
- Header now separates equity-based 'Total P/L (incl. unrealized)' from
  'Realized P/L (closed trades)' so it no longer contradicts profit factor /
  win rate, and the stats strip shows the fee-drag chain
  (gross - fees = net) plus a per-trade-labelled sharpe.
- Risk radar read a non-existent account field (total_unrealized_profit) so
  unrealized PnL always showed $0; small PnLs also render with cents now.
- Nav 'Connect Hyperliquid' turns into a green connected chip when the server
  already holds a fully-authorized exchange, instead of nagging forever from
  a browser without the localStorage flow state.
2026-07-06 00:01:44 +09:00
tinkle-community
a744328313 feat: unify autopilot launch on server preflight with live balance polling
- New shared launch module (web/src/lib/launch): preflight client, model/
  exchange resolution, and a single launchAutopilot orchestrator used by both
  Strategy Studio and the guided launch panel — the two previously duplicated
  and divergent launch implementations are gone.
- Preflight runs BEFORE any strategy mutation, so a failed launch no longer
  rewrites/activates the strategy as a side effect; readiness and minimums
  come from the server (live balances) instead of stale client caches.
- Guided panel polls preflight every 20s so deposits are detected without
  manual refresh; step metas show live balances and required minimums; the
  beginner wallet page polls its balance while the deposit screen is open.
- Every failure now routes to a guided fix: hyperliquid-funds gets its own
  setup anchor (scrolls to the launch panel), start rejections surface the
  server's preflight reason (including manual restarts from the trader list),
  and structured error bodies flow through ApiError.errorData.
- Terminal dashboard shows a persistent runtime banner when the AI fee wallet
  runs low/empty or the trader enters safe mode.
- Vitest coverage for setup-target routing, readiness helpers, and the
  launch orchestration order (no side effects before preflight).
2026-07-05 20:38:51 +09:00
tinkle-community
49613830ce feat: client-side simulation layer for terminal dashboard panels
Adds a local synthetic data engine that can drive all dashboard panels
(candles, order book, liq ladder, stats) without a backend connection,
plus de-synced beam animation phases in the orchestration topology.
2026-07-05 19:13:44 +09:00
tinkle-community
277b5ad9c6 refactor: drop equity curve + AI rationale panels, rebalance dashboard footer
- Remove the equity curve (only ever 2 points → a meaningless straight line);
  also drop the now-unused equity-history SWR poll and path calc
- Remove the Latest Decision / AI rationale panel (redundant with Execution Log)
- Rebalance the footer: Market Net Inflow + By Symbol now form one two-column
  row (1.5fr / 1fr) instead of full-width inflow + a lonely by-symbol column
2026-06-30 16:13:59 +08:00
tinkle-community
110bf52908 feat: cream terminal redesign, English-only UI, autopilot launch fixes
- Redesign dashboard into a cream-paper + vermilion IBM Plex Mono terminal
  (live L2 order book, cost/liq map, WS K-line, signal matrix, orchestration
  topology, risk radar, execution log, current positions, equity curve)
- Convert all user-facing UI and backend strings/prompts from Chinese to
  English (multi-language retained, default English)
- Add /api/statistics/full endpoint + full-stats frontend wiring
- Fix Autopilot launch: reuse the existing trader instead of creating
  duplicates (eliminates repeat ~35s create cost and stale-trader 404s);
  launch sends 5m scan interval
- Fix unreadable toasts: cream theme with high-contrast text + per-type accent
- Silence background dashboard polls (getTraderConfig) to stop error-toast spam
2026-06-30 16:03:52 +08:00