Market Calendar

■ Bull day (SPY up) ■ Bear day (SPY down) ■ Weekend/holiday

🎬 YouTube to AI

Paste YouTube links below. Each video will be transcribed, converted into a structured trading strategy note, linked into the Obsidian knowledge graph, and tagged with Pine Script code.

Pipeline Agents

🎬 Agent 1: Transcript
YouTube captions + Whisper fallback
📝 Agent 2: Editor
LLM extraction → structured Markdown
📂 Agent 3: Importer
Obsidian vault with collision protection
🔗 Agent 4+6+7: Linker
Graph links + Pine Script + Regime
📊 Agent 8: Pidge DB
Master strategy ingestion (74 files)
⚡ Agent 9: Breakers
2% max, 5/day cap, 5% DD halt
🛡 Fail-Safes (6)
Health, stagnation, manifest, drawdown

Queue Status

Pending
0
Processing
0
Completed
0
Failed
0
📈
Trading AI
🏗
System Architecture & Improvement
🤖
Managed Agents
📊
Homer Marketing Department
📋 PAPER TRADING — open P&L: — Unrealized: — Combined P&L: — Loss limit: Unlimited
📌 End of session: log your trades via Interactive Brokers — Homer pulls them automatically once the IBKR Flex Query ID is set.
🛑 Daily loss limit: stop trading for the day after losing $450 (3 × the $150 risk per trade).
📋 PAPER TRADING — open P&L: — Unrealized: — Combined P&L: — Loss limit: Unlimited
🤖 AI Pidge Stock Test 3/6+ · +15% target · -5% stop · Paper only Feed source: Twelve Data live quotes, under 15s old
0 trades today — — 0 open
SymbolScoreMarketEntryStopTarget SharesEntry timeStatusExitP&LAction
🔔 AI alerts — open P&L: — Unrealized: — Combined P&L: — Loss limit: Unlimited
🔔 AI alerts Every alert · +15% target · -5% stop · Paper only Feed source: Twelve Data live quotes, under 15s old
0 trades today — — 0 open
SymbolScoreMarketAlertEntryStopTarget SharesEntry timeStatusExitP&LAction
🏦 IBKR — really traded — open P&L: — Unrealized: —
🏦 IBKR paper account Read from IBKR fills · real prices, real fills, real P&L · read-only
0 trades today — — 0 open
SymbolEntryStopTarget SharesCostEntry timeStatusExitP&LFills
Score bands — does a higher score mean a better result?
Source: the analytics engine, from the two paper-trade ledgers (each trade's Pidge score). Root data: the scanner board — TradingView's scanner list, Twelve Data prices.
Win rate and profit/loss by how many of the 6 Pidge filters a stock passed, for both paper test books.
News catalysts — which kind of news moves a stock most?
Source: catalyst-impact.py, every 15 min, from the Pidge news archive plus the scanner's day files. Root data: TradingView news headlines, and stock moves from the scanner board (Twelve Data prices).
Every headline in the news archive, grouped by category and ranked by average peak move (highest first). The count next to each category is how many headlines support it.
Reconciliation — did our paper trades really happen at IBKR?
Source: reconcile-paper-vs-ibkr.py, every 15 min — the two paper ledgers checked against IBKR's own order book. Root data: the IBKR paper account DUR241952, read through the paper gateway.
Green = the ledger and IBKR agree. Anything else is a difference worth looking at.
Cumulative equity curve
Source: the analytics engine, rebuilt every 5 min, from the two paper-trade ledgers. Root data: the scanner board — the stock list from TradingView's scanner and live prices from Twelve Data.
Variable performance heatmap
Source: the analytics engine, rebuilt every 5 min, from the two paper-trade ledgers. Root data: the scanner board — the stock list from TradingView's scanner and live prices from Twelve Data.
Win rate for every pair of criteria. Darker/red = worse, brighter/green = better.
Win / loss distribution
Source: the analytics engine, rebuilt every 5 min, from the two paper-trade ledgers. Root data: the scanner board — the stock list from TradingView's scanner and live prices from Twelve Data.
How it was analyzed
Source: the analytics engine, rebuilt every 5 min, from the two paper-trade ledgers. Root data: the scanner board — the stock list from TradingView's scanner and live prices from Twelve Data.
Top combinations
Source: the analytics engine, rebuilt every 5 min, reading ONLY the active book's own live trade ledger — no backup files and no other book. Every combination that was tested is listed; none are capped or hidden. Root data: the scanner board — the stock list from TradingView's scanner and live prices from Twelve Data.
Paper P&L — the running total behind each day
Each paper test separately, from its own trade record. Built nightly at 9:00 PM ET by paper-pnl-db.py; the same numbers are saved into the Obsidian database.
0 selected
Date Priority Owner Approve Title Notes File
Recent Sessions
Loading…
Messages
Select a session on the left.
Loading…