NVIDIA reports Wednesday after the close. The stock's down six straight sessions into the print — peaked at $225.30 on August 13, closed Friday at $214.72. That's what we call a soft run-in: negative five sessions before earnings.
The obvious read is the market's pricing in bad news, so maybe there's upside surprise potential. We checked twenty-six years of August fiscal Q2 reports. Nine were soft run-ins. The answer is no edge — but the distribution is the story.
Five of those nine rallied +14% to +22% over the next twenty days. Two cratered −23% to −27%. Two went flat. Almost nothing in between. It's a barbell, not a bias. Median is +14%, but that's just the middle value of a nine-point spread — not a typical outcome. Mean +3.45%, standard deviation near 17 points. Hit rate 56% versus 53% for hot run-ins. Coin flip.
Correlation between the run-in and what follows is basically zero: +0.05 at T+5, −0.02 at T+10, −0.32 at T+20 (p-value 0.12 at n=26). Nothing to trade.
This year's −4.64% run-in sits on the deeper end historically — comparable to 2019 (−6.00%), 2022 (−6.07%), 2010 (−4.69%). Those three went +22.42%, −27.04%, +14.61% at T+20. The three closest analogues span a fifty-point range. A deeper run-in doesn't tilt the odds, it just signals the move afterward will be large, in a direction the setup doesn't disclose.
One wrinkle: the true run-in window closes Wednesday at 4pm. $NVDA sits −1.31% against the August 19 base. A ~1.3% gain across Monday/Tuesday/Wednesday flips this into a hot run-in. Soft is likely, not locked.
Twenty-six years of August reports say six straight down closes tell you far less than it looks like they should.
The obvious read is the market's pricing in bad news, so maybe there's upside surprise potential. We checked twenty-six years of August fiscal Q2 reports. Nine were soft run-ins. The answer is no edge — but the distribution is the story.
Five of those nine rallied +14% to +22% over the next twenty days. Two cratered −23% to −27%. Two went flat. Almost nothing in between. It's a barbell, not a bias. Median is +14%, but that's just the middle value of a nine-point spread — not a typical outcome. Mean +3.45%, standard deviation near 17 points. Hit rate 56% versus 53% for hot run-ins. Coin flip.
Correlation between the run-in and what follows is basically zero: +0.05 at T+5, −0.02 at T+10, −0.32 at T+20 (p-value 0.12 at n=26). Nothing to trade.
This year's −4.64% run-in sits on the deeper end historically — comparable to 2019 (−6.00%), 2022 (−6.07%), 2010 (−4.69%). Those three went +22.42%, −27.04%, +14.61% at T+20. The three closest analogues span a fifty-point range. A deeper run-in doesn't tilt the odds, it just signals the move afterward will be large, in a direction the setup doesn't disclose.
One wrinkle: the true run-in window closes Wednesday at 4pm. $NVDA sits −1.31% against the August 19 base. A ~1.3% gain across Monday/Tuesday/Wednesday flips this into a hot run-in. Soft is likely, not locked.
Twenty-six years of August reports say six straight down closes tell you far less than it looks like they should.