Split Congress: R Senate, D House?

The evidence leans toward the 40.5% Polymarket price being too high for this specific "classic split" outcome. The two forecasters that publish full joint, correlated simulations, Decision Desk HQ and Nate Silver's Silver Bulletin, both imply a probability in the 21% to 24% range once you account for how tightly House and Senate outcomes move together in the same national environment, a full 16 to 20 points below the market. The long run historical base rate for any divided Congress since 1945 is also close to 25%, reinforcing the model estimates. The one real counterweight is recency: three of the last four midterms (2010, 2018, 2022) produced exactly this structural pattern, opposition wins House, president's party holds Senate, which is a small sample but argues the true rate could run above the 1945-to-present average. On balance, the quantitative anchors point lower than 40.5%, though the gap is a judgment call about correlation rather than a settled fact.

• Decision Desk HQ (July 2026): Democrats about 61-62% to win the House, Republicans about 57% to hold the Senate, but DDHQ states there is a 3-in-4 chance one party sweeps both chambers. Back-solving those numbers gives an implied classic-split (GOP Senate, Dem House) probability of about 22%, versus the 40.5% market price.
• Nate Silver's Silver Bulletin gives Democrats about a 40% chance of sweeping both chambers and Republicans about a 36% chance of sweeping both, for roughly 24% combined divided-government probability. Silver has said the reverse split (GOP House, Dem Senate) is only about 1%, implying the classic split alone is close to 23%. Two independently built joint models converge on roughly 21-23%.
• Historical base rate: since 1945, the House and Senate have been held by different parties in only about 10 of roughly 40 Congresses, close to 25%. That figure sits near the model estimates and well under the 40.5% market price.
• Countervailing signal: in the last four midterms (2010, 2014, 2018, 2022), the specific pattern of the opposition flipping the House while the president's party holds the Senate occurred three times (2010, 2018, 2022), a small-sample 75% hit rate. The 2026 Senate map, where the only real Republican targets (Georgia, Michigan) are Democrat-held seats and Republicans have almost nothing exposed, structurally resembles those cycles and is the strongest argument for a higher-than-historical-average rate.
• Marginal (non-joint) probabilities alone would suggest a higher number: Silver's 85-90% Dem House and up to 60% GOP Senate multiply to roughly 50%, and Kalshi's simple marginals (about 84% Dem House, about 49-57% GOP Senate depending on date) multiply to roughly 40-48%. The entire gap between 'in line with market' and 'model says overpriced' comes down to how strongly correlated you believe the two chambers are, and Silver explicitly describes them as almost perfectly correlated.
• Cross-market check: Kalshi's own combo contract for this exact outcome traded near 31% in late May 2026, and Polymarket's split-outcome pricing has moved roughly between 36% and 42% in recent weeks, bracketing the current 40.5% and sitting above both DDHQ's and Silver Bulletin's implied joint estimates.

The exact outcome this Polymarket contract prices (Republicans hold the Senate, Democrats flip the House) is not a hypothetical: it happened once before, in the 116th Congress elected during Trump's first midterm in 2018, when Democrats netted 41 House seats while Republicans still netted 2 Senate seats on an unusually lopsided map. The 2026 map inverts that lopsidedness in the GOP's favor (Republicans defend 22 of 35 Senate seats versus Democrats' 13), which is the structural reason a Senate hold is plausible even in a hostile political environment. That environment currently looks similar to, or worse than, October 2018: Trump's approval was 37.1% as of mid-July 2026 versus 38 to 41% just before the 2018 midterm, and the generic congressional ballot favors Democrats by roughly mid-single digits nationally. Historical base rates (presidents under 50% approval have averaged 37-seat House losses at midterm) point toward a House flip, while Senate forecasters still see Democrats needing a net gain of four seats against a map built to protect Republicans. This is research on historical patterns and current polling, not a prediction or financial advice.

• The precise split this market prices (GOP Senate + Dem House) happened only once since 1984: the 116th Congress, elected in Trump's first midterm in November 2018, which was the first Congress since the 99th (elected 1984) to pair those two outcomes.
• In 2018, Democrats netted 41 House seats (their biggest House gain since 1974) while Republicans still netted 2 Senate seats, because Democrats were defending 26 of 35 seats that cycle on what FiveThirtyEight called the most lopsided Senate map either party has ever faced.
• The 2026 Senate map mirrors that dynamic in reverse: Republicans are defending 22 of 35 seats up (versus Democrats' 13), which structurally cushions a GOP hold even in a bad national environment for the party; Democrats need a net gain of 4 seats for a majority.
• From 1934 to 2018, the president's party averaged losing 28 House seats and 4 Senate seats at midterm, with only three post-1932 exceptions (1934, 1998, 2002), each tied to unusual circumstances (New Deal popularity, impeachment backlash, a post-9/11 rally) not obviously present in 2026.
• Approval is historically the best single predictor of loss size: presidents below 50% approval have averaged 37-seat House losses versus 14 for those above 50%. Trump's approval was 37.1% approve / 59.0% disapprove (net -17) as of July 14, 2026, close to the 38-41% Gallup readings recorded just before the 2018 midterm.
• Sabato's Crystal Ball currently rates only Alaska and Ohio as Senate toss-ups and recently moved North Carolina from toss-up to lean Democratic, a modest tightening, but the underlying seat math still favors a Republican hold given how few genuinely competitive GOP seats are in play.

The consensus "GOP Senate, Dem House" split looks defensible on paper (favorable 2026 Senate map, D+6ish generic ballot, weak Trump approval) but sits on a battlefield that has shifted meaningfully in just the last two weeks. Maine's Graham Platner, Democrats' best shot at a GOP-held seat, withdrew around July 8-10 amid a rape allegation, with a replacement not chosen until a July 25 party convention, a live catalyst that cuts toward this contract's YES side. At the same time Texas and Iowa, historically safe Republican seats, are polling as toss-ups, widening the Senate map in a way that raises the tail probability of a full Democratic sweep, which would resolve this specific contract NO even if the House goes as expected. Georgia's December 1 runoff rule is a distinct resolution-timing risk: if that seat decides Senate control and no candidate clears 50% on Nov 3, this market may not settle on the date assumed in the brief. Redistricting litigation (a Voting Rights Act-gutting Supreme Court ruling, Missouri's map facing a possible referendum suspension) and a September 30 government funding deadline five weeks before the election are additional live variables that could still move the House and Senate environment before votes are counted.

• Resolution mechanics: Polymarket settles this market once AP, Fox News, and NBC all call both chambers, falling back to official certification if they disagree, and to the party of the first-elected Speaker or Senate Majority Leader if control is structurally ambiguous. This is a real edge case in a Congress this narrowly divided, not a formality.
• Georgia timing risk: winning the Nov 3 Ossoff vs. Mike Collins Senate race requires a majority, not a plurality. Falling short sends it to a Dec 1 runoff. If that seat is the deciding vote for Senate control, resolution could slip roughly a month past the date implied in the brief.
• Maine catalyst (days old as of this writing): Graham Platner, who had led Susan Collins by about 2 points in a late-June/early-July NYT/Siena poll, suspended his campaign around July 8-10, 2026 amid a rape allegation and other controversies. Maine Democrats pick a replacement nominee at a July 25, 2026 convention, an active, unresolved variable in the GOP's most vulnerable held seat.
• Battlefield widening: Texas (Paxton vs. Talarico tied 47-47 in a late-June UT/Texas Tribune poll) and Iowa (Hinson 48, Turek 46) are polling competitive, seats Republicans have not had to defend in decades. If that holds, the tail scenario is not just a split outcome but a full Democratic Senate-and-House sweep, which resolves this specific contract to NO.
• Redistricting is not fully settled: the Supreme Court's Louisiana v. Callais ruling (April 29, 2026) narrowed Voting Rights Act Section 2 protections against gerrymandering claims, and Missouri's new pro-GOP map faces a citizen referendum that could suspend it for 2026, live legal wildcards on top of an already-priced-in net Republican redistricting edge (roughly a mid-single-digit House seat swing nationally).
• Macro catalysts ahead: a Sept 30, 2026 government funding deadline (following two shutdowns already this year) sits five weeks before the election as a blame-game risk for both parties, and the generic ballot itself ranges from D+3 to D+11 depending on pollster, meaning the House-flip leg of this bet is favored but not locked in.