earnings-play

name: earnings-play description: Framework for trading around earnings — detecting the earnings lock (R6), picking IV-aware structures, and recording post-earnings lessons.

Earnings plays

Earnings is the best-understood edge source on single names (vol expansion / crush, gap mean-reversion, post-earnings drift) AND the most-common source of catastrophic losses (directional wrong + IV crush).

The earnings lock (R6)

Within 2 trading days of next earnings (earnings_dte ∈ [0, 2]), the PreToolUse hook rejects any order whose strategy_label does not start with earnings_. Rationale: normal directional trades get blown up by IV crush on the print.

earnings_dte is not looked up automatically — you must pass it into the order tool. If you're using /research or /enter, check the earnings calendar manually (via EDGAR 10-Q/8-K dates or a public calendar) and pass it explicitly when the print is imminent.

Strategy labels for earnings setups

Use these verbatim — post-mortem aggregates on the exact label:

• earnings_directional_debit_spread — long call/put + short higher-strike call / lower-strike put at the same expiry. Neutralizes most of vega; caps profit but caps loss too. The safer pre-print bet when you have direction conviction.
• earnings_iv_drop — buy ATM/slightly-ITM call/put after the print once IV collapses, betting on mean-reversion of the gap.
• earnings_long_call / earnings_long_put — single-leg long premium through the print. Highest risk: IV crush typically eats 30-50% of the debit regardless of direction. Only attempt when IV is already low before the print and the expected move is larger than historical.
• earnings_post_drift — held after the print to capture post-earnings announcement drift (PEAD). 1-3 week holding period on beats with strong guide.

Pre-earnings checklist

Before filing an earnings thesis:

1. Consensus + whisper: what are analysts expecting? Any recent guide update in 8-K? (Use edgar-mcp.get_recent_filings_for_ticker.)
2. Last 4 reactions: did the stock go up on beats? Down on misses? Or is the reaction uncorrelated with the surprise?
3. Implied move: straddle price ÷ stock price. If 8% is priced in and the historical average is 4%, options are expensive — prefer spreads or skip.
4. IV rank: if IV rank > 80 pre-print, long premium is almost always a net loss due to crush. Either spread it or stay out.
5. Sector context: did peers report? Did they beat/miss? Earnings cycles cluster.
6. Insider activity: edgar-mcp.get_insider_transactions — Form 4 within 30 days of the print is a weak but real signal.

Post-earnings playbook

After the print, new thesis + new trade. Do NOT carry a pre-print position through unless the original thesis explicitly stated "through earnings" (rare and usually a mistake).

1. Gap beat + guide up → consider earnings_post_drift long for 1-3 weeks.
2. Gap miss + guide cut → consider short stock or long puts. (Short premium — CSPs etc. — is currently hard-blocked at the order tools until multi-leg combos get atomic sizing.)
3. In-line + muted reaction → no edge; move on.
4. Big gap + IV still elevated → earnings_iv_drop after the first 1-2 days once IV normalizes.

Known lesson (seed this into notes early)

Long single-leg calls through earnings on mega-caps (AAPL/MSFT/NVDA) historically lose 25-35% of the debit to IV crush alone when the direction is correct by 2%. Either size the trade assuming you lose 30% on the vol move, or use a spread. This lesson should be appended to notes with source="post_mortem" and tags including earnings_iv_crush so it surfaces on semantic search next time.

Logging earnings trades

When you record a thesis for an earnings play:

• timeframe = "through_earnings" or "post_earnings_1-3w"
• invalidation = concrete trigger. Examples:
  • "pre-print: close below $250 OR implied move prices in > 10%"
  • "post-drift: close below 5-day SMA OR 15% adverse move"
• max_loss_pct — for single-leg long premium, set this equal to the full debit (you can lose 100%).

After the trade closes, run /weekly-post-mortem or manually append_note with tags lesson,earnings,<ticker>,<strategy_label> so the next research cycle picks it up.