resolve_open_trades walks the daily bars after each open trade and closes it at
the target (target hit) or stop (stop/ambiguous), leaving undecided trades open.
Runs nightly inside the outcome evaluator (so it's coordinated with fresh OHLCV)
and on its manual trigger. New "My Trades" section at the top of Signals → Track
Record shows realized hit-rate, expectancy (avg R), total R, total P&L, and a
closed-trades table — your actual results, separate from the theoretical signal
record below it.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The 500+ row readiness table forced scrolling past it to reach the actual job
controls. Put JobControls first, readiness below.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Richer LLM output (same grounded call, ~no extra cost):
- All providers now also return a recommendation (buy/hold/avoid) and a thorough
reasoning paragraph; Gemini now actually captures reasoning + grounding
citations (it was dropping them). Stored on sentiment_scores (migration 008),
exposed in the API; display-only — NOT fed into the composite/EV.
- Ticker Sentiment panel shows an "LLM view" badge and a "Full analysis & sources"
expander with the complete reasoning + citations.
Search-budget scoping (Gemini grounding free tier = 5000/mo):
- collect_sentiment now targets only watchlist + open paper trades + top-N by
composite, skips tickers refreshed within sentiment_fresh_hours (72h), and caps
per run (sentiment_max_per_run). Once the relevant set is fresh, runs spend 0
searches until it ages out — bounding monthly usage well under the free tier.
- Widened sentiment lookback to 7d (scoring + display) so sparser collection
still feeds the dimension score.
Deploy: alembic upgrade (sentiment_scores.recommendation). Switch provider to
Gemini Flash in Admin for the cost win (grounded, cheapest).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
New paper_trades table (migration 007) + service/router. "Mark as taken" on each
setup card (shares prefilled from position sizing, entry from current price, both
editable) records a simulated trade. Overview gains an Open Trades table that
marks each position to the latest close — P&L in $, %, and R-multiples — with a
total unrealized P&L footer and a Sell button to close at the current price.
Closed trades are retained for future realized-P&L reporting.
Deploy: alembic upgrade (new paper_trades table).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Re-applies the activation gate at several min_target_probability thresholds
(60→30, other conditions fixed) over the already-replayed candidates, so the
trade-off between how many setups qualify and their expectancy is visible in one
table — the cheap "optimize" half of Phase 2. Candidates now carry meets_core +
best_prob so the sweep needs no re-replay. New sweep table in BacktestPanel with
the current threshold starred.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Backtest (32k setups) showed the touch-only probability model was ~2x
over-confident — predicted 70% hit 39%, predicted 88% hit 46% — because it
ignored the competing stop. estimate_probability now multiplies the reach
probability (touch within horizon) by the two-barrier gambler's-ruin ratio
1/(R:R+1) = P(target before stop). A 3:1 setup now reads ~25% base, not ~70%,
which lines up with realized rates. Strength/alignment modulation unchanged.
Recalibrates every probability and the EV ranking; the min_target_probability
gate threshold now means roughly what it says. Re-run the backtest to confirm
the calibration table flattens toward the diagonal.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
New BacktestPanel: shows qualified hit-rate/expectancy vs the all-setups baseline,
a by-direction breakdown, and the probability calibration table (predicted vs
realized, over-confident buckets flagged amber). Includes a "Run backtest" button
that triggers the job and a plain explanation of the method and its limits.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Replays the price-derived engine over stored OHLCV: at each weekly as-of date,
rebuild the setup from bars <= D (no lookahead) and walk the actual forward bars
for the realized outcome. Reports realized hit-rate/expectancy of qualified
setups (and all setups, by direction) plus a probability calibration curve
(predicted target prob vs realized hit rate).
Reuses pure functions throughout; extracted compute_technical_from_arrays /
compute_momentum_from_closes from scoring_service so live and backtest stay in
sync. Runs as a weekly/triggerable 'backtest' job caching the report in a
SystemSetting; GET /backtest/report serves it. Sentiment/fundamentals held
neutral (no point-in-time history) — calibrates the price/S-R/probability machinery.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The Jobs panel only surfaced live progress; once a job finished you couldn't see
when it ran, whether it succeeded, or its message (e.g. a regime/collector error).
Add a "Last run <ago> · <status> — <message>" line per job, colored by status,
from the runtime_* fields the backend already returns.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The spinner arrows clash with the dark glass UI. One base-layer CSS rule removes
them from every type=number input (admin settings, signals filters, data cleanup,
etc.) — values are typed, and type=number still gives the mobile numeric keypad
and validation.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The account/risk inputs are global "set once" settings, so they're moved out of
the panel body into a single compact line in the recommendation header. Replaced
the number-input spinners: risk % is now a segmented preset selector (0.5/1/2/3),
account size a clean text field with a $ prefix. Relabel "at risk" → "max loss".
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Finnhub's earnings calendar now supplies next_earnings_date through the
fundamentals chain; persisted on fundamental_data (migration 006) and exposed in
the fundamentals API. The recommendation panel warns when earnings fall within
the ~30-day target horizon (a report can gap price through stop/target) and
otherwise shows the next date. Informational only.
Deploy: run alembic upgrade (new fundamental_data.next_earnings_date column).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
New market_regime_service computes a benchmark (SPY) trend from its 50/200-day
SMAs, cached in a SystemSetting and refreshed by a nightly job; GET /market/regime
exposes it. Dashboard shows a regime banner; setup cards flag a counter-trend
caution when a setup fights the regime (LONG in a bearish market / SHORT in a
bullish one). Informational only — nothing is suppressed.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Risk-based sizing on each setup card: shares = floor((account × risk%) /
|entry − stop|), with position value and dollars-at-risk. Account size and
per-trade risk % are editable inline and persisted in localStorage. Flags when
a position would exceed the account (needs margin). Frontend-only.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Qualified tickers already get their own "qualified setup" alert, so an S/R
proximity ping on them is redundant noise. Drop the watchlist ∪ qualified scope
(remove now-unused _alert_scope_tickers) and alert only on watchlist tickers.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Trade-setup targets now pre-merge near-duplicate S/R levels into zone
representatives (same 2% clusterer as chart + alerts) before generate_targets
runs. A clustered wall (e.g. 183 + 185) becomes one target carrying the zone's
COMBINED strength (capped 100) instead of two near-identical targets that each
undervalue the wall — which also feeds a more honest reach-probability via the
S/R-strength magnet. Representative price is the zone's near edge; the strongest
constituent's id is retained. Singleton levels pass through unchanged, so the
downstream band-spreading / probability / primary-selection pipeline and its
tests are untouched.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Three fixes to over-firing S/R proximity alerts:
- Route through cluster_sr_zones (the same merger the chart uses) instead of raw
SRLevel rows, so near-duplicate levels (e.g. CVX 183 + 185) collapse into one
zone and one alert.
- Alert only the single NEAREST strong zone per ticker, not every nearby level.
- Scope to watchlist + qualified-setup tickers via _alert_scope_tickers (was
iterating all watchlist entries only; qualified setups are now included too).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Root cause of "price plan needed in bulk but fine on manual reload": on free
tiers FMP returns only market cap (others 402) and the chain merged that as a
partial success — so when the Finnhub/Alpha Vantage fallbacks were rate-limited
during a bulk run, the chain silently returned market-cap-only and the
collector's backoff never engaged. Manual single fetches worked because the
fallbacks weren't throttled at that moment.
Fixes:
- Chain distinguishes RateLimitError from other failures: if a fallback is
rate-limited and fields are still missing, raise RateLimitError (unless
allow_partial=True) so the collector backs off and retries.
- Bulk job paces requests (fundamental_request_spacing_seconds, default 3s) to
stay under Finnhub's ~60/min, and on retry-exhaustion stores partial data and
continues instead of aborting the whole run.
- Manual fetch passes allow_partial=True so a lone 429 doesn't fail the refresh.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Closes the action loop — instead of polling the dashboard, the platform pushes
actionable signals to Telegram. New hourly 'alerts' job dispatches four
toggleable triggers, deduped via a new alert_log table (cooldown-based for
qualified/S-R/digest, watermark-based for score deterioration). Admin → Settings
gains a Telegram panel (write-only bot token, chat ID, per-trigger toggles, Send
Test). Credentials follow DB > env precedence (TELEGRAM_BOT_TOKEN / _CHAT_ID).
Backend: alert_service + AlertLog model + migration 005, scheduler job, admin
endpoints/schema. Frontend: AlertSettings panel, hooks, api, types.
Deploy: run alembic upgrade (new alert_log table).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
AsyncIOScheduler was constructed with no job_defaults, so APScheduler's default
misfire_grace_time of 1s applied. In this single-process app the scheduler shares
one event loop with the API and all other jobs, so when a daily job came due
while the loop was busy (e.g. the scanner mid-run), the fire was processed >1s
late, flagged a misfire, and skipped — while next_run still advanced 24h, making
the job look healthy though it never ran. Set a generous grace window (1h),
coalesce missed runs into a single catch-up, and cap concurrency at 1.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
scan_rr set the total then called scan_all_tickers as one opaque await, so the
runtime snapshot's processed count stayed 0 until the whole scan finished and
jumped straight to 100%. scan_all_tickers now takes an optional progress_callback
invoked per ticker; the scheduler wires it to _runtime_progress so the sidebar's
live indicator advances as tickers are scanned.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Track Record: new "Reset" action (POST /admin/track-record/reset) deletes all
trade setups so stats start fresh after material scoring/setup changes — live
setups regenerate on the next scan. Guarded by a confirm dialog.
Recommendation config: remove distance_penalty_factor, which was exposed in the
admin UI but consumed nowhere (the touch-probability model superseded it). A
knob that silently does nothing is worse than no knob. Remaining defaults are
left as-is — they're reasonable, and the honest way to tune them is backtesting
against accumulated outcomes, not invented "researched" numbers.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Per design decision: the watchlist is now purely user-curated (no auto-seeding
of the top-10), so the auto_populate/dismissed machinery is removed and removals
are plain deletes. Each entry is enriched with latest close + day-over-day move.
Overview now shows two clear blocks: Top Setups (what to trade) and My Watchlist
(my names with current price and today's %). Market watchlist table drops the
now-meaningless auto/manual Type column in favour of Price and Day columns.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Removing a ticker did nothing because get_watchlist re-runs auto_populate on
every read, instantly re-adding any top-ranked ticker the user had just removed.
Removals are now tombstoned as a "dismissed" entry_type: auto-population skips
them, the list hides them, and a later manual add revives the row. Also exposes
an Add/Remove-watchlist toggle in the ticker detail header.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The dashboard Top Setups list showed raw fields in arbitrary order with no
indication of why a ticker was listed or which was best. Now sort by expected
value (R) — probability-weighted payoff per unit risk — so the strongest
opportunity is row 1, badged "Top pick", with a new Exp. Value column that
folds R:R and target probability into one "is this worth taking" number.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Answers "why does a too-far-progressed setup still show": setups are only
recalculated by the scheduled R:R scan and manual fetch; at creation
entry == current price (0% progress), so over-progression is a
between-scans drift effect and must be judged at read time.
- /trades now attaches current_price (latest close per ticker).
- Qualification drops setups whose R:R recomputed from the current price
falls below min_rr — i.e. price already ran toward target (reward
consumed) or through the stop. Reuses the existing min_rr threshold
instead of a separate progress %; far cleaner (a 3:1 is already ~1:1
by 33% progress). Skipped for historical setups (no current_price).
- Fix: useFetchSymbolData now invalidates the trades queries, so a fetch/
recompute actually refreshes confidence/setups in the UI (was the cause
of the stale 100% confidence lingering after recompute).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- Overview Top Setups shows the primary target's probability (concrete,
distance-calibrated) instead of the overlapping confidence number. The
stale 100% confidences were leftovers from the old model and self-heal
on rescan; confidence stays in the detail view + gate.
- Each metric now has one home: composite = ranking, target probability =
actionability, confidence = direction conviction.
- Staleness message states the real basis (% of entry->target distance
already covered), not the raw % from entry, so narrow setups read
correctly ("67% of the move is gone").
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Scores never updated ("101d ago"): get_score only recomputes stale/
missing dimensions, but nothing marked them stale on new data, and there
was no scheduled scoring job.
- Fetch endpoint force-recomputes dimensions + composite.
- Scheduled scan (scan_all_tickers) refreshes scores per ticker, so
scores stay current globally, not just on manual fetch.
Granular fetch: /ingestion/fetch accepts a sources filter; the freshness
bar gets a per-row refresh button (OHLCV/Sentiment/Fundamentals fetch
that provider only — marked paid; S/R/Scores recompute for free). Header
button is now "Fetch All".
Job visibility: GET /jobs/running (any user) + sidebar live indicator
showing running scheduled jobs with progress, polled every 10s.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
MKC showed 5 targets all far/Aggressive: target selection was top-5 by
quality (0.35*R:R + ...), and R:R grows with distance, so far levels
crowded out every nearby one.
generate_targets now selects for spread: always include the nearest
level, plus the best-quality representative from each distance band
(Conservative <=2.9 ATR, Moderate <=4.6 ATR, Aggressive beyond), then
fill remaining slots by quality. Restores a Conservative/Moderate/
Aggressive mix with the nearest target always present.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Fixes nonsensical "Conservative @ 90%" on far targets (AJG: +39% target
labelled Conservative/90%).
- Probability = chance of touching the level within the outcome horizon
under a driftless random walk: 2*(1 - Phi(d / (ATR*sqrt(T)))), T=30d to
match the outcome evaluator. Distance (in ATR) dominates; strength and
alignment modulate modestly. Far targets are now correctly low-prob.
- Classification derived from that probability (>=60 Conservative,
40-60 Moderate, <40 Aggressive) instead of distance-rank.
- Combined with the most-likely-worthwhile primary pick, the nearest
strong resistance becomes the Conservative primary; far levels demote
to low-probability stretch targets.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- JWT now carries a username claim; sidebar shows "Signed in as <name>"
instead of the bare user id (sub). Re-login required for the new claim.
- Signals: Min R:R / Min Confidence inputs reflect the effective filter —
auto-filled from the activation gate when "Qualified only" is on, reset
to 0 when off (no more misleading 0 while the gate is active).
- Signals layout: Run Scanner moved to its own action row (it's a job
trigger, not a filter); qualified toggle grouped with the refinement
filters under one Filters panel.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Triggered by MRK: entry 113 shown with no current price (actually ~119).
- Ticker header shows last close + day change % + "last close · Nd ago"
(the age reveals OHLCV collection lag — why entry looked off)
- Setup cards show Current price and entry drift; flag setups as
stale (price moved >1/3 toward target) or invalidated (past stop)
- Chart: draw only nearest support below + nearest resistance above
current price, plus a prominent "Now" price line (full S/R stays in
the S/R tab)
- Chart overlay is selectable (Auto/Long/Short/None) — only the chosen
setup's entry/stop/target render, instead of everything at once
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Triggered by CNC showing "LONG (High Confidence)" with SHORT reasoning
and no long setup.
- A: recommendation action + reasoning are ticker-level and identical
on both setups; reasoning always matches the shown action
- B: recommended_action only picks a direction with a tradeable setup;
strong bias with no setup (e.g. price at ATH) → NEUTRAL with an
explanatory reason instead of a fake LONG_HIGH
- C: confidence is a directional-agreement model — opposing signals push
it below 50 (SHORT on a 92-technical/99-momentum stock ~0%, not 55%)
- D: fundamental score requires >=2 real metrics (market-cap-only no
longer yields a high score)
- E: RSI score peaks at healthy momentum (~60) and penalizes
overbought/oversold extremes instead of treating RSI 90 as maximal
- F: fundamentals chain merges fields across providers (FMP market cap
+ Finnhub P/E) instead of stopping at the first with any field
- NEUTRAL label: "No Clear Setup" (covers untradeable-bias case)
Scores recompute on next scan/scoring run; C and E shift score
distributions intentionally.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
xAI returned 410 — search_parameters/Live Search is retired. Route xAI
through the Responses API web_search tool instead (same path as OpenAI):
- OpenAISentimentProvider parametrized with base_url / tool_type / source
- xAI builds it against https://api.x.ai/v1 with the web_search tool
- Drop the dead Live Search code from the generic compatible provider
- Frontend label: "xAI Grok — web search"
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Providers (admin-switchable, no redeploy):
- DeepSeek and any OpenAI-compatible endpoint (OpenRouter, Together,
Groq, local Ollama) via a generic Chat Completions adapter + base_url
- xAI Grok with Live Search (search_parameters web+X, citations) —
grounded tier alongside OpenAI and Gemini
- DeepSeek / generic compatible endpoints are ungrounded (no web
search); UI shows an amber warning and labels each provider's grounding
- Optional env fallbacks DEEPSEEK_API_KEY / XAI_API_KEY
UI: replace native <select> (unstyleable white popup on Windows) with a
custom dark Dropdown component everywhere — sentiment provider, scanner
filters, market sort, indicators, admin universe, user role.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Make "qualified" mean an edge candidate, not just R:R + confidence.
The gate now also requires (all admin-configurable, defaults on):
- high conviction: recommended_action LONG_HIGH / SHORT_HIGH only
- clean read: risk_level Low (no contradicting signals)
- probable primary target: best target probability >= min (default 60)
- Shared predicate: app/services/qualification.py +
frontend/src/lib/qualification.ts (mirrored)
- Activation config extended (min_target_probability,
require_high_conviction, exclude_conflicts) with bool-aware
get/update + validation
- /trades/performance switched to ?qualified_only=true, applying
the full gate server-side; confidence breakdown stays unfiltered
- Dashboard "Qualified", Signals "Qualified only" toggle, and
Track Record all use the one gate; Admin gains the new controls
Sentiment provider runtime config (prior change) included.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Admin-configurable thresholds (min R:R, default 2.0; min confidence,
default 70%) defining what counts as an actionable signal:
- Admin Settings: new Activation Thresholds panel
(GET/PUT /admin/settings/activation)
- GET /trades/activation exposes values to all users with access
- Signals/Setups: filters initialize from activation values
- Track Record: "Qualified signals only" toggle (default on) via
min_rr/min_confidence params on /trades/performance; the
confidence breakdown always covers the full population so the
thresholds can be validated against outcomes
- Dashboard: "Qualified" metric and qualified-first Top Setups
- Outcome evaluator unchanged: every setup is still evaluated
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- Technical dimension now uses all directional indicators:
0.30*ADX + 0.20*EMA + 0.20*RSI + 0.15*EMA_Cross (bullish=80 /
neutral=50 / bearish=20) + 0.10*Volume_Profile (POC proximity)
+ 0.05*Pivot_Points (structure confluence); weights re-normalize
when data is insufficient, as before
- ATR stays out of scoring (volatility input for scanner stops,
not a directional signal)
- IndicatorSelector uses the shared Select so the option list is
dark instead of the native white popup
- Update technical scoring tests for the six-component breakdown
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Closes the feedback loop on R:R scanner signals:
- Nightly outcome_evaluator job replays unresolved setups against daily
OHLCV bars: target_hit / stop_hit / ambiguous (same-bar, counted as
loss) / expired after OUTCOME_EVALUATION_MAX_BARS (default 30)
- Migration 004: evaluated_at + outcome_date on trade_setups
- GET /trades/performance: hit rate, expectancy (avg R), total R with
breakdowns by direction, recommended action, and confidence bucket
- New Performance page (stat cards, breakdown tables, Evaluate Now,
methodology disclosure) wired into sidebar and mobile nav
- 17 new unit tests for evaluation logic and stats aggregation
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- Add shared UI primitives: Button, Field/Input/Select, PageHeader,
Section, Callout, Tabs, Disclosure
- Replace gradient buttons with single blue-accent btn-primary
- Reserve gradient text for the brand wordmark only
- Rework Scanner page onto the glass system; collapse explainer and
glossary into a disclosure, move filters into a glass toolbar
- Restructure Ticker Detail into tabs (Analysis / Indicators / S/R)
with chart and recommendation always visible
- Align Watchlist, Rankings, Admin, Login/Register to shared primitives
- Unify stray indigo/violet/gray accents into the blue family
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
dennisthiessen