Hey traders,
Every October of a midterm year, the same chart shows up.
Some strategist on CNBC, suit a little too crisp, voice pitched at that register that says I’ve seen a few cycles, kid, pulls up the S&P 500, boxes off the five sessions before Election Day and the three after, and lets the number land like he’s revealing a magic trick.
Eighteen of the last nineteen.
Since 1950.
Median gain about 3%.
Bank of America ran it with a slightly different ruler and got 22 of 23. Yahoo Finance crowned it the best eight-session stretch of a midterm year. The window opened at Tuesday’s close, and I swear you could hear the collective exhale of every financial advisor in America.
So how’s it going? Is this the same old thing we’ve seen before?
Friday the S&P 500 dropped 1.8% and snapped its 100-day moving average like a dry twig.
The Nasdaq 100 fell 3.5% and went sniffing around correction territory.
Intel lost 29% in a single session, a day the company hadn’t seen since Gerald Ford was in office.
Reddit lost 29%, too.
Amazon slipped on a capex number that made shareholders feel like they’d been handed the check at a dinner they didn’t order.
For the month, the Nasdaq Composite finished down 5.3%, its worst since March 2025. The Dow, that creaky old thing, finished up 1.5%.
Read that again: The growth index lost five and the old-economy index… gained one and a half?
That’s a market changing its mind about what it wants to own, a week before the country changes who’s in power. (That’s the whole point of what Garrett Baldwin is running Monday, Tuesday and Wednesday at 10 a.m. Eastern.)
But does that mean the market will perform differently than in the past? Let’s think about that stat I mentioned earlier for a second…
It describes the index. It says nothing about who’s inside it.
Think back to 2022. The stat technically worked that year, but you’d never know it from living through it. Powell walked out on November 2nd, hiked 75 basis points, and when somebody floated the word “pivot” he swatted it down as “premature.”
Stocks got smoked the Thursday before the vote. The window only turned green because a soft CPI print landed two days after the election and the market ripped. If you’d bought the chart on that Tuesday, you spent a week staring at red waiting for a rescue that had nothing to do with Congress.
2018 ran the tape backward. October was a rout, the week before the vote was a relief rally, the strategist got to take his bow. Then the S&P fell 7.6% from Halloween to New Year’s Eve, the worst post-midterm stretch in the entire sample.
The stat paid you for eight sessions and then came back for the money with interest.
And the bottoms?
Three of the last nine midterm years, 1990, 2002 and 2022, put in their low between October 9th and October 12th. Weeks before anybody voted. The market doesn’t wait for results. It front-runs them. By the time the famous window opens, the big money konows what the election means for its book, and the tape you’re squinting at is the residue of that decision.
Which brings us back to this week…
Tech didn’t sell off because earnings were bad. Amazon and Alphabet grew. Tech sold off because the spending outlook got heavier at the exact moment the odds of the party that likes to regulate, tax and subpoena large companies went to roughly nine in ten on the House.
Rates sit near 5%.
Oil’s been working. Healthcare was Thursday’s best sector. None of that is a coincidence, and none of it shows up in a bar chart of S&P closes.
I’ll be blunt about the alternative.
You can hold the index, lean on a stat that’s been right 18 of 19 times, and hope this is the 19th time and not the 20th. The stat might even hold. The S&P could finish next week higher, and the guy in the crisp suit will be back on Thursday to remind you he told you so. And you could still be sitting on the wrong half of a market that just told you, in the loudest voice it’s used since June, that it’s rotating.
The midterm pattern is real.
It just isn’t a plan.
Have a good spooky weekend.
And I'll see ya Monday!
Charles Delvalle
Editorial Director, Option Pit