Grading My Prediction Market Election Scorecard

The odds called it. The scorecard didn’t go their way.

Kalshi had Democrats taking both chambers of Congress at about 70% going into November 3, and that’s what happened. So the favorite won. The question my election series has been asking is different: did knowing that make the election any less uncertain for investors?

In my election scorecard, I wrote down four checks before any results, using what S&P 500 options charge for election night as the measuring stick. In one last check before the vote, I graded the two that finished early. Here’s the full scorecard with the last two graded, and what I make of it. I held it for a couple of weeks to see how the whole post-election week played out, since the day after turned out to be a poor guide to the rest of it.

The scorecard

Check 1: Passed

Test: In the final week, the election-night premium in options should follow Kalshi’s odds.

Result: It did, more closely than I expected. That’s the one place my call was wrong. I covered it in the pre-election update.

Check 2: Failed

Test: The expected result should make for an ordinary Wednesday, a move of less than 1.0%.

Result: SPY rose 1.65%, a bit more than the 1.42% options had priced as of Tuesday’s close.

For comparison, 2018 started from the same place, the expected result happened, and the S&P rallied 2.1% the next day. 2026 landed in the same neighborhood.

Check 3: Failed

Test: Options should price election night clearly below 2022, at 1.2 normal days or less over the final five readings.

Result: 2026 averaged about 1.55, against 1.57 in 2022, the last midterm before Kalshi listed control of Congress. I covered it in the pre-election update.

Check 4: Failed

Test: The index should stay quiet, under 1.0%, while at least 3 of the 11 sector funds move the other way by more than 0.5%.

Result: The index moved 1.65%, and only two sector funds went the other way.

Energy fell 4.74% as oil dropped on hopes for de-escalation with Iran, and industrials fell 1.09%. Tech was flat. The other eight rose, from 0.44% for utilities to 4.26% for consumer discretionary. That’s a rising tide, not a reshuffle.

That’s three out of four against, which matches my October call of “not visibly.” It’s also consistent with what I found for the Fed, where options keep charging about 1.3 normal days even when Kalshi has the decision as a near-lock.

Then Thursday erased $2.9 trillion

The scorecard ends on Wednesday, but the market didn’t.

On Thursday the S&P 500 fell 4.4%, and tech had what Fortune called its biggest selloff since Covid, as investors weighed what a Democratic Congress, and a wave of local votes against data centers, would mean for the AI buildout. By MarketWatch’s estimate, the day erased about $2.9 trillion from the S&P 500, the largest one-day loss in market value ever. As of Tuesday’s close, options had priced Thursday at about 1.06%, an ordinary day.

Nvidia fell harder than most, about 8%, with headlines of its own that day, including a congressional request for Jensen Huang to brief lawmakers. But it wasn’t the whole story. All 11 sector funds fell, and the AI trade swooned across the board. By Friday’s close, the index was down about 2.5% from election eve.

That’s an odd way to greet a result everyone saw coming. If $2.9 trillion isn’t reason enough to take the odds seriously, what is?

The 30% problem

Here’s why the drop matters for the odds:

  1. Kalshi had Democrats taking both chambers at about 70% on the eve of the vote, up from the low 60s a week earlier. The next most likely outcome, a split Congress, was at about 24%.
  2. If investors believed that, stocks should already have been most of the way toward pricing in a Democratic Congress. Only the remaining 30% should still have been news.
  3. For 30% of a surprise to knock 2.5% off the index, the gap between that outcome and the alternatives would have to be worth about 8% of the S&P 500.

If it was worth that much, someone should have been paying up to protect against it. Check 3 says nobody was.

Uncertainty went up, not down

The clearest picture comes from what options charged for the six weeks through December 18:

  • Election eve: about 15.7%, with election night still included.
  • Wednesday: 13.1%, as if the uncertainty had been settled.
  • Friday: 19%, with no election left in it.
What S&P 500 options charged for the stretch through December 18 at each day’s close. The shaded days still had election night ahead of them.

To be fair, volatility usually rises when stocks fall, and some of that jump is just Thursday’s drop showing up again. Options expiring in 2027 rose much less. So I won’t claim the election created a lasting new source of uncertainty.

The narrower claim is the one that matters: the expected outcome didn’t reduce it. And to keep the timeline straight, all of this came before Nvidia’s leadership crisis broke open the following week.

The case against me

The best objection is that Thursday was about what the result meant, not whether it would happen. Investors trusted the odds and just hadn’t worked out the consequences for AI spending.

Maybe. But this wasn’t a thin call. A 70% favorite, with the main alternative at 24%, was the scenario the market had priced as most likely, and it’s exactly the one you’re supposed to have worked out before the vote. Doing that homework is what trusting a probability looks like.

The other objection is that the win was bigger than the odds let on. Kalshi’s contract was about who would control Congress, not by how much, so these markets alone can’t rule that out. But it doesn’t rescue them either. If they got the headline right and missed the part that moved stocks, they still didn’t deliver the information that mattered.

So why didn’t anyone use it?

My read is that investors don’t trust the prices yet. A price set by a fairly small group of traders, on contracts that are cheap to push around, isn’t something a portfolio manager will cut hedges on.

The odds are also built mostly from the same polls and forecasts the professionals already read. And almost nobody with real exposure hedges there. The money that protects portfolios still lives in options and futures, and in the week of the vote it priced the election as though the odds didn’t exist.

I build in this space and I’m not neutral about it. I’d like these markets to work. But I’m skeptical they will until they list questions specific enough, and deep enough, that a fund would rather hedge there than in options.

Where I land

Prediction markets got the midterms right, and so far that hasn’t made elections any less uncertain for stock investors. Until investors trust the odds enough to act before the vote, prediction markets are just another forecast. What would it take for you to act on the odds before an event like this?

Author: Ed Kaim

Ed Kaim is the founder of Quantcha, which has built options analytics since 2014 and licenses its tools and data to market makers, quant funds, and hedge funds. He writes about how options, stocks, and event contracts fit together, and has filed a formal comment with the CFTC on its prediction markets rulemaking. Earlier, he was a product manager at Microsoft. More at edkaim.com.