How to compare earnings reactions with Claude or ChatGPT

An earnings reaction is a stock’s price move after a company reports its results. This scan anchors the move to the company’s SEC 8-K item 2.02, Results of Operations and Financial Condition, and separates the opening gap from the moves that followed.

The rule Monk.trade uses: the latest 8-K item 2.02 in the last 10 sessions. A filing counts on the first daily bar ending at or after Monk.trade has it. Before-open filings count on the coming session; after-close filings count on the next trading day. Measure from the close before that session to today’s close.

Your app writes this MonkQL from the prompt. Each line does one thing.

1. from stocks

   Scan every active US stock.

2. let age = since(filings(form = "8-k", item = "2.02"), 10)

   Find the latest 8-K item 2.02 in the last 10 sessions, counting from when Monk.trade has the filing.

3. include type == "common" and sma(close * volume, 20)[1] >= 2e7

   Keep common stock averaging at least $20M traded a day over the prior 20 sessions.

4. where age >= 0

   Keep stocks with a results filing in that period.

5. show age as sessions, 100 * (open[age] / close[age + 1] - 1) as gap_pct unit percent, 100 * (close / open[age] - 1) as since_pct unit percent, 100 * (close / close[age + 1] - 1) as total_pct unit percent

   Show sessions since the filing’s daily bar, the opening gap, the move since that open and the total move from the prior close.

6. sort abs(total_pct) desc

   Largest total move first, whether up or down.

7. take 5

   The five largest.

• Gapped up, then faded

  “Using Monk.trade, which stocks opened higher on the day their latest 8-K item 2.02 counted in the last 10 sessions, then fell from that open? Keep common stock averaging at least $20M traded a day, largest falls first.”

  from stocks
  let age = since(filings(form = "8-k", item = "2.02"), 10)
  include type == "common" and sma(close * volume, 20)[1] >= 2e7
  where age >= 0 and open[age] > close[age + 1] and close < open[age]
  show age as sessions, 100 * (open[age] / close[age + 1] - 1) as gap_pct unit percent, 100 * (close / open[age] - 1) as since_pct unit percent, 100 * (close / close[age + 1] - 1) as total_pct unit percent
  sort since_pct asc
  take 5
  

• Gapped down, then recovered

  “Using Monk.trade, which stocks opened lower on the day their latest 8-K item 2.02 counted in the last 10 sessions, then rose from that open? Keep common stock averaging at least $20M traded a day, largest recoveries first.”

  from stocks
  let age = since(filings(form = "8-k", item = "2.02"), 10)
  include type == "common" and sma(close * volume, 20)[1] >= 2e7
  where age >= 0 and open[age] < close[age + 1] and close > open[age]
  show age as sessions, 100 * (open[age] / close[age + 1] - 1) as gap_pct unit percent, 100 * (close / open[age] - 1) as since_pct unit percent, 100 * (close / close[age + 1] - 1) as total_pct unit percent
  sort since_pct desc
  take 5
  

334 stocks matched; 23 had too little price history to check. Each row’s evidence confirms an 8-K with item 2.02; the anchor is when the filing counts, which can be later than the results release. Run 01a17f1e-0951-7901-89ba-c47fa4f92ccc

1. Oct 24, 2026
   How did Comfort Systems stock react to its Q3 earnings?

   Q3 revenue of $3.77B and $13.91 a diluted share in its Oct 22, 2026 8-K; the stock fell 9.3% to $1,160.86 on Oct 23.

• Some companies release results without an 8-K item 2.02, or file it after the release.
• A filing during regular hours counts on that day’s bar, so its opening gap predates the filing.
• Other news can also move prices over the measured period.
• Past reactions do not predict future prices.