Tech stock research desk

Hermes stock picker

A tech-stock research and paper-trading lab for watchlists, thesis logs, source citations, signal scoring, risk controls, and eventual approval-gated broker automation.

Back to hub
RESEARCH ONLYNO BROKER APIPAPER PORTFOLIO FIRSTQQQ BENCHMARK

market console

Tech ideas do not become trades until they beat the benchmark on paper.

moduleTech universe, source list, and score schema
moduleBullish and bearish score cards with cited catalysts
modulePaper portfolio journal with 1 day, 3 day, 1 week, and 2 week outcomes
moduleApproval gated broker investigation, no order placement by default

buildout checklist

Phase 1: research dashboard
  • Limit initial universe to 30 to 50 technology stocks across mega-cap tech, semiconductors, SaaS, AI infrastructure, and speculative tech.
  • Add market signals: price, volume, relative volume, moving averages, RSI, MACD, Bollinger Bands, gaps, volatility, beta, trend strength, and support or resistance.
  • Add news and event signals: earnings calendar, guidance, SEC filings, analyst changes, product launches, AI partnerships, regulation, layoffs, and major catalysts.
  • Add social listening: Reddit, Stocktwits, Google Trends, Hacker News, YouTube, Bluesky, and X if available, with emphasis on mention acceleration rather than raw hype.
  • Add fundamentals: growth, margins, free cash flow, cash and debt, valuation multiples, revisions, dilution, stock-based compensation, and guidance quality.
Phase 2: scoring and reports
  • Create separate bullish and bearish composite scores from -100 to +100 using technicals, sentiment and news momentum, fundamentals, catalyst quality, options liquidity, and risk adjustment.
  • Generate top 5 bullish tech setups, top 5 bearish setups, top unusual social or news movers, options watchlist, avoid list, open paper trades, and post-close review.
  • Every idea must include direction, confidence, time horizon, thesis, source trail, invalidation level, target, risk notes, and review date.
  • Prefer watch or wait calls over constant buy signals. A stock only becomes actionable when the thesis, timing, liquidity, and risk/reward all clear the filter.
Phase 3: paper trading
  • Run a simulated $200 to $500 paper portfolio before risking real money.
  • Benchmark every signal and trade against QQQ and relevant sector baskets.
  • Track 1 day, 3 day, 1 week, and 2 week outcomes for each recommendation.
  • Measure hit rate, average return after signal, high-confidence precision, false positives, expected value, max drawdown, profit factor, and QQQ-relative performance.
Phase 4: risk and execution guardrails
  • For a $500 real-money test, risk only about 1 to 3 percent per trade and keep positions tiny.
  • Avoid margin shorts and naked options. Prefer fractional shares, paper bearish calls, long puts, or defined-risk spreads only after liquidity checks.
  • Reject options with poor liquidity, wide bid/ask spreads, unmanaged earnings IV crush, or unclear catalyst timing.
  • Require max open positions, max daily loss, no-trade conditions, and an emergency kill switch before any broker automation.
Phase 5: possible broker automation
  • Keep E-Trade as the manual execution venue unless API access is verified and explicitly authorized.
  • Use Alpaca paper trading as the easiest first automation target if the user wants agent-placed paper trades.
  • Move to approval-based real trade tickets before any fully autonomous order placement.
  • Only consider limited autonomous execution after the paper system shows a real, logged edge over QQQ.