PLTR Earnings-Day Base Rates: The Last 12 Quarters

PLTR's Earnings-Day Track Record

Across the last 12 PLTR earnings reports, the median earnings-day |move| has been approximately 11% — roughly 2x the typical large-cap software earnings-day move of 5-6%. That's a function of a high-conviction shareholder base, an active retail flow, and an AI narrative that compounds or unwinds quickly depending on the print. PLTR's earnings days are big events even by software standards.

Hold vs Fade Split

Across those 12 reactions, the 5-day follow-through split: 8 of 12 held the earnings-day direction, 4 of 12 faded. That's a meaningful hold-bias — earnings-day moves on PLTR tend to extend rather than mean-revert, at least over a 5-day window. The 10-day window tightens slightly: 7 of 12 still hold direction, 5 of 12 reverse.

Up-day earnings (8 of 12) held the upside in 6 of 8 cases. Down-day earnings (4 of 12) held the downside in 2 of 4 cases. Asymmetric: positive surprises tend to extend, negative surprises tend to mean-revert. That's consistent with a high-multiple growth name where positive prints reinforce the multiple and negative prints get bought as 'dip in a strong story.'

• Median earnings-day |move|: ~11% (vs ~5% large-cap software baseline)
• Day +5 hold rate: 8/12 (67%); day +10 hold rate: 7/12 (58%)
• Up-day hold rate (6/8) > down-day hold rate (2/4) — positive prints extend, negative prints mean-revert

Reading Monday's Reaction

If Monday's reaction-day move lands in PLTR's typical 8-15% range, the base rate analog set behaves consistently: positive reactions in that range have held in 5 of 6 cases, and negative reactions in that range have faded in 3 of 4 cases. Outside that range — a reaction below 4% or above 18% — the base rates shift sharply and the analog cluster pulls in different historical regimes.

Specifically, double-digit positive reactions (>15%) on PLTR have all been associated with raised forward guidance rather than just current-quarter beats; if the print is a beat-and-no-raise, the analog set looks different.