Earnings Preview Report for 5 Holdings — Before the Market Opens

The Problem

Earnings week. Five of my holdings reporting in the same window. NVDA on Tuesday. GOOGL and META on Wednesday. TSLA on Thursday. PLTR on Friday.

Each stock needs prep. Consensus estimates. Historical surprise rates. Implied move from the options chain. The 2-3 metrics that actually matter for that specific company. If you're doing this properly, that's 3-4 hours per stock. Minimum.

Five stocks. Fifteen to twenty hours of analyst work. For a holding company operator who also runs three portfolio companies, that time doesn't exist.

Most people solve this by skipping the prep entirely. They watch earnings, react, and wonder why they're always a step behind. Or they pay for a research service that gives them generic reports written for everyone and tuned for no one.

I wanted something different. I wanted my reports. For my holdings. With my risk parameters. Delivered before I pour my Monday morning coffee.

The Solution

Two Mr.Chief skills working together: US Stock Analysis and SEC Filing Watcher. The SEC watcher tracks filings in real-time and catches any last-minute 8-Ks or amendments. The stock analysis skill pulls market data, options chains, and historical earnings data.

Hari — my investment advisor agent — orchestrates the whole thing. Every Sunday at 21:00 UTC during earnings season, Hari checks which of my holdings report that week, then generates a consolidated earnings preview.

The Process

The configuration is straightforward:

# earnings-preview.yaml
schedule: "0 21 * * 0"  # Every Sunday 9PM UTC
holdings:
  - NVDA
  - GOOGL
  - META
  - TSLA
  - PLTR
analysis_per_stock:
  - consensus_estimates: [revenue, EPS, guidance]
  - historical_surprise: last_8_quarters
  - options_implied_move: front_week_straddle
  - key_metrics:
      NVDA: [data_center_revenue, gross_margin, china_exposure]
      GOOGL: [cloud_revenue, search_growth, capex_guidance]
      META: [DAU_growth, ARPU, reality_labs_loss]
      TSLA: [deliveries, auto_margin, energy_storage]
      PLTR: [commercial_revenue, rule_of_40, net_retention]
  - sec_recent_filings: last_30_days
output: markdown + telegram_alert

Hari runs through each stock sequentially:

1. Pull consensus — aggregates from multiple financial data sources via Alpha Vantage and web scraping
2. Calculate historical surprise — maps last 8 quarters of beats/misses and the magnitude
3. Read the options chain — front-week straddle price tells you what the market expects the stock to move
4. Extract key metrics — per-stock custom metrics that actually drive the thesis
5. Check SEC EDGAR — any recent filings, insider transactions, or amendments that could matter

The SEC Filing Watcher runs continuously, so by Sunday, Hari already has a cache of anything filed that week.

# Earnings preview template structure
preview = {
    "stock": ticker,
    "report_date": earnings_date,
    "timing": "AMC" | "BMO",  # After market close or before market open
    "consensus": {
        "revenue": {"estimate": "$38.2B", "range": "$37.1B - $39.8B"},
        "eps": {"estimate": "$0.89", "range": "$0.82 - $0.96"},
    },
    "surprise_history": {
        "beat_rate": "7/8 quarters",
        "avg_surprise": "+8.3%",
        "last_quarter": "+11.2% EPS beat"
    },
    "implied_move": {
        "straddle_price": "$14.20",
        "implied_pct": "±6.8%",
        "historical_avg_move": "±8.1%"
    },
    "key_watch": ["Data center $30B+ threshold", "Gross margin >74%"],
    "risk_flags": ["China revenue disclosure", "Blackwell ramp timeline"]
}

The output hits my Telegram as a single consolidated brief. One message. Five stocks. Everything I need.

The Results

The real value isn't speed — it's consistency. Week after week, I get the same depth of analysis. No shortcuts because I'm tired. No missed filings because I forgot to check EDGAR. The agent doesn't have bad weeks.

Last earnings season: the NVDA preview flagged a discrepancy between consensus revenue and the implied options move. The market was pricing a ±6.8% move on a stock where the last 4 quarters averaged ±8.1%. That was a signal. Not a trade recommendation — Hari doesn't do that — but information I would have missed doing this manually at midnight.

Try It Yourself

Install the US Stock Analysis and SEC Filing Watcher skills on Mr.Chief. Configure your holdings list and custom key metrics per stock. Set the cron schedule for Sunday evening. The earnings calendar is pulled automatically — no manual date entry needed.

Start with 2-3 holdings. Get comfortable with the output format. Then expand. The marginal cost of adding another stock is essentially zero.

The market doesn't care how many hours you spent preparing. It cares whether you were prepared.