Scan: Will the Democrats win the Michigan Senate race in 2026?

date	predicted	market price	confidence
2026-08-29	59%	64%	60%
2026-08-22	66%	64%	60%
2026-08-15	54%	56%	52%

1. Planner Fable 5.1 · Fermi decomposition + tool plan

Tools chosen: polymarket_direct kalshi_related polymarket_related article_search claude_news wikipedia code_execution

Sub-questions (Fermi decomposition)

What is the current Polymarket and Kalshi implied probability of a Democratic win in the 2026 Michigan Senate race, and how has it trended?
Who are the leading Democratic candidates (Haley Stevens, Mallory McMorrow, Abdul El-Sayed) and who is favored in the August 2026 primary; does the eventual nominee affect general-election polling?
Is Mike Rogers the presumptive Republican nominee, and how does he poll head-to-head against each potential Democratic nominee?
What is the national environment (generic ballot, Trump approval) and what is the historical base rate for the president's party losing Senate seats in midterms, particularly in Michigan-type swing states?
Are there significant independent or third-party candidates (e.g., a Detroit-area independent) who could split the Democratic vote?
What do forecasters (Cook, Sabato, Inside Elections) rate the race, and how does fundraising compare between parties?

Planner reasoning

This is an open-seat Michigan Senate race (Gary Peters retiring) in a 2026 midterm with a Republican president, which historically favors the out-party. Key drivers are the Democratic primary outcome (Stevens/McMorrow/El-Sayed), Mike Rogers as likely GOP nominee, polling, national environment, and possible independent candidacies. The Polymarket price is the primary anchor, with Kalshi as a cross-check.

2. Research Parallel tool calls

tool	status	time	items	summary
polymarket_direct	OK	0.2s	1	## This Market's Polymarket Data Will the Democrats win the Michigan Senate race in 2026? - Current price (probability): 61.50% - 7-day price change: -3.00% - 30-day price change: +3.00% - Total volume: $455,617 (USD notional) - Price range: 55.50% - 74.00% - Data points: 91 days
kalshi_related	OK	16.9s	1	1 related markets / summaries. keyword 'Michigan Senate': ok \| keyword 'Rogers Michigan': no matches \| keyword 'Stevens McMorrow': no matches
polymarket_related	OK	0.3s	0	Scanned 100 active Polymarket markets, kept 0 matches. keyword 'Michigan Senate': 0 markets \| keyword 'Senate control 2026': 0 markets \| keyword 'Michigan Democratic primary': 0 markets
article_search	OK	46.6s	16	Fetched 16 unique articles across 3 queries (lookback=45d). 'Michigan Senate race 2026 poll Rogers Stevens McMorrow El-Sayed': 10 articles \| 'Michigan Senate Democratic primary 2026': 10 articles \| 'Mike Rogers Senate campaign Michigan': 10 articles
claude_news	OK	18.0s	13	Based on available research: - Race dynamics: Incumbent Sen. Gary Peters (D) is retiring. For most of the primary, El-Sayed, congresswoman Haley Stevens, and state senator Mallory McMorrow sought the Democratic nomination. After McMorrow withdrew in July, El-Sayed and Stevens remained in a tig
wikipedia	OK	0.2s	3	Fetched 3 Wikipedia entries (0 missing pages).
code_execution	OK	58.4s	0	Historical base rate (open Senate seats in swing states, midterm years since 1994) - Constructed a set of 23 open-seat Senate races in competitive/swing states across the 8 midterm cycles (1994, 1998, 2002, 2006, 2010, 2014, 2018, 2022): MI/ME/AZ '94, IN/OH '98, MN '02, MN/TN '06, IL/DE/IN/OH/PA

3. Evidence Brief Sonnet 5 · 6935 chars

# Current state Democratic nominee Abdul El-Sayed defeated Rep. Haley Stevens in the Aug 4, 2026 primary (Mallory McMorrow withdrew in July) and now faces Republican nominee Mike Rogers (unopposed) in the Nov 3, 2026 general election. Polling since the primary shows a near-toss-up, with El-Sayed holding a small aggregate lead; the race resolves only on AP/Fox/NBC race call or official certification. # Timeline of key events - 2026 (pre-July): Three-way Dem primary — El-Sayed, Stevens, McMorrow (confirmed, Wikipedia/CNN). - 2026-07: Mallory McMorrow withdraws from primary (confirmed, CNN/Wikipedia). - 2026-08-04: Michigan Democratic primary — El-Sayed defeats Stevens, 48.5% (confirmed, CBS/Wikipedia). - 2026-08-07: Michigan Dem general-election kickoff rally in Detroit; party unity push (confirmed, CBS). - 2026-08-10: El-Sayed reports call with Obama; Dems seek to unify factions (reported, CBS). - 2026-08-13→14: First post-primary general poll (Fox News) shows race close (reported, WaPo). - 2026-08-13→early Sept: Decision Desk HQ model moves from 51% El-Sayed (8/13) to 66% El-Sayed (early Sept) (reported, claude_news synthesis). - 2026-08-31–09-03: WDIV/Detroit News poll: Rogers 45.8% vs El-Sayed 44.4%, 9% undecided (reported, Newsweek/The Hill). - 2026-09 (various): MSU poll El-Sayed +5 (50-45); Susquehanna El-Sayed +7 (46-39); AARP/Fabrizio-Ward/Impact El-Sayed +1 (48-47); Trafalgar (9/9) El-Sayed +1 (46-45) (reported, Newsweek/pollsmax). - 2026-09-13: CBS "races to watch" piece confirms matchup and framing as top GOP pickup opportunity (confirmed, CBS). - Pollsmax aggregate (as of research date): El-Sayed +1.4 (45.9% to 44.5%) across 27 polls (reported, pollsmax.com). # Event Will the Democrats win the 2026 Michigan U.S. Senate race (El-Sayed vs. Rogers), resolved by AP/Fox/NBC call or certification. # Outcomes to forecast - Yes (Democrat/El-Sayed wins) - No (Republican/Rogers or other wins) # Kalshi market anchor No Kalshi-direct price was returned for this specific ticker in the raw research (only related Kalshi markets on 2028 Senate/Presidency sweep, not this race). **Primary cross-market anchor is Polymarket**: current YES (Dem win) price **61.5%**, down 3pts over 7 days but up 3pts over 30 days; range 55.5%–74% over 91 days; volume ~$456K. Treat 61.5% as the closest available consensus anchor given absent Kalshi-direct data. # Sub-question answers 1. **Polymarket/Kalshi implied probability** — Polymarket: 61.5% Dem win, down from a high of 74% earlier in the cycle, recently declining (-3% 7d) but up slightly over 30d (+3%). No Kalshi-direct price found for this ticker. 2. **Leading Dem candidates / primary favorite** — Stevens was the establishment/Whitmer-backed moderate, El-Sayed the progressive (AOC/Sanders-backed); McMorrow withdrew in July. El-Sayed won Aug 4 primary 48.5%-to-Stevens. No solid data yet on whether nominee choice shifted general-election polling versus a Stevens matchup (CNN, CBS, Wikipedia). 3. **Mike Rogers as GOP nominee / head-to-head** — Rogers ran unopposed for the GOP nomination (Wikipedia). Head-to-head polls are mixed/close: WDIV/Detroit News has Rogers +1.4; MSU has El-Sayed +5; Susquehanna El-Sayed +7; AARP/Fabrizio-Ward El-Sayed +1; Trafalgar El-Sayed +1. Aggregate (Pollsmax): El-Sayed +1.4. 4. **National environment / historical base rate** — Trump approval reportedly slipping; Democrats face a tough overall map needing +4 seats nationally but Michigan is framed as GOP's best pickup shot (Cook). Code-execution model of open swing-seat races since 1994 shows the president's opposition party (Dems here, since Trump is president) wins ~70% of the time historically. 5. **Independent/third-party candidates** — No significant independent or third-party contender identified in the research; none mentioned as vote-splitting threat. 6. **Forecaster ratings / fundraising** — Cook Political Report rates it a Toss-up, calling it the GOP's most likely pickup and the only Dem-held seat so rated. Decision Desk HQ model gives El-Sayed 66% (up from 51% in mid-August). No specific fundraising totals reported in research. # Key facts (high-confidence, factual) 1. [Wikipedia] El-Sayed (D) vs. Rogers (R), open seat, Peters (D) retiring. 2. [Wikipedia] Republicans haven't won a MI Senate race since 1994; 2024 race decided by 0.3pts (Slotkin over Rogers). 3. [CBS/Wikipedia] El-Sayed won primary 48.5% Aug 4, 2026. 4. [Pollsmax] Aggregate: El-Sayed +1.4 (45.9-44.5) across 27 polls. 5. [Newsweek/The Hill] WDIV poll: Rogers 45.8%, El-Sayed 44.4%, 9% undecided. 6. [claude_news] Decision Desk HQ: El-Sayed 66% win probability (early Sept), up from 51% (Aug 13). 7. [claude_news] Cook Political Report: rated Toss-up, GOP's top pickup target. 8. [claude_news] El-Sayed favorability -8 net; Rogers -9 net; El-Sayed +14 among independents. # Cross-market signals - Kalshi related: No direct ticker price found; adjacent 2028 Dem Senate sweep market at 50% (down 10pts over 30d) — not directly comparable but signals softening Dem Senate optimism generally. - Polymarket: 61.5% Dem win, declined from 74% high, recent 7-day dip. - Sportsbook: none found. # Analyst opinions and speculation - CNN: El-Sayed's win seen as test of progressive-vs-establishment theory within Democratic Party; outcome could reshape narrative regardless of general result. - Rogers privately said he'd feel "in good shape" against El-Sayed vs. Stevens (CNN/reported, unverified). - Decision Desk HQ model shift (51%→66% for El-Sayed) suggests improving Dem trajectory into September, contrasting with Polymarket's recent 7-day dip and individual polls showing a closer race (Rogers ahead in WDIV poll). # Directional lean per outcome - **Yes (Dem win)**: Supported by polling aggregate lead (+1.4 to +7 in several polls), Decision Desk HQ model (66%), historical base rate for opposition party in swing-state open seats (~70%), independents favoring El-Sayed (+14), Polymarket at 61.5%. - **No (GOP win)**: Supported by WDIV poll showing Rogers ahead, Cook Political Report's Toss-up/GOP-pickup framing, Michigan being Trump-won state in 2024, El-Sayed's net-negative favorability, recent 7-day Polymarket decline from 74% peak. # Gaps / unknowns - No Kalshi-direct YES price retrieved for this specific ticker — cannot state Kalshi's own anchor price. - No fundraising totals compared between candidates. - No confirmed independent/third-party candidate data. - Uncertain how much of Polymarket's 74%→61.5% decline reflects real tightening vs. noise. # Calibration anchors - Polymarket price (best available cross-market anchor): 61.5% Yes. - Decision Desk HQ model: 66% El-Sayed win probability. - Historical base rate (opposition party, open swing Senate seats since 1994): ~70%. - Polling-error model (current ~1.4-3pt Dem lead, σ≈4.5): ~65-75% implied win probability. - Convergent estimate range across methods: ~62-75% Democratic win probability.

4. Forecasters Fable 5.1 + GPT-5.6 Sol · in parallel

claude-fable-5-1

52% conf

Yes	64%
No	36%

No Kalshi-direct price exists, so the Polymarket consensus of 61.5% Dem win is the primary anchor. Supporting evidence leans modestly Democratic: the Pollsmax aggregate has El-Sayed +1.4 across 27 polls with several quality polls at +5 to +7, Decision Desk HQ's model rose to 66%, midterm dynamics favor the out-party (~70% historical base rate in open swing seats), and Michigan hasn't elected a GOP senator since 1994. Countervailing signals — Cook's Toss-up/top-GOP-pickup framing, the WDIV poll showing Rogers narrowly ahead, El-Sayed's net-negative favorability, and the recent 7-day Polymarket dip — keep this a genuine coin-flip-plus rather than a comfortable Dem lean. I land slightly above the market at 64% Yes, reflecting the fundamentals tilt while respecting the tight polling.

gpt-5.6-sol

72% conf

Yes	64%
No	36%

The best market anchor is Polymarket at 61.5% Yes, since no direct Kalshi price was available. El-Sayed's 1.4-point polling-average lead, advantages in most listed polls, and Decision Desk HQ's 66% estimate point toward Yes, while the Rogers-leading WDIV poll and Cook's Toss-up rating keep the race competitive. The opposition-party historical advantage and absence of a significant third-party threat modestly strengthen the Democratic case. I therefore move slightly above the market anchor to 64%, but not as high as the historical model because Michigan's 2024 Senate result was extremely close and El-Sayed has negative favorability.

5. Devil's Advocate Sonnet 5 · red-team critique

Both forecasters converge on identical 64% despite claiming independent reasoning, and neither explains why they land exactly between Polymarket (61.5%) and Decision Desk HQ (66%) rather than weighting toward the historical base rate (~70%) or the polling-error model range (65-75%) cited in the brief's own "convergent estimate" section — this suggests anchoring bias rather than genuine synthesis.

Neither forecast grapples with the tension between Decision Desk HQ's sharp rise (51%→66% in three weeks) and Polymarket's simultaneous 7-day decline; this divergence in trend direction is a meaningful signal of disagreement among sophisticated aggregators that deserves more explicit discussion of which trend is more likely to be noise.

Both treat the 2024 Slotkin-Rogers race (decided by 0.3pts) as a minor caveat rather than the single most relevant reference class — a nearly identical matchup (same Republican candidate, same state) resolved as a razor-thin Dem win, which arguably should pull the estimate closer to a true coin-flip rather than 64%.

Neither forecast addresses the "9% undecided" figure in the WDIV poll or discusses undecided-voter breakout patterns in Trump-era Michigan races, which could materially swing the outcome given the race's closeness.

Both forecasts underweight Cook Political Report's explicit framing as "GOP's most likely pickup" and "only Dem-held seat so rated" as a toss-up — this is a credentialed nonpartisan assessment that arguably deserves more weight than a slight nod, especially since it directly contradicts the more bullish Decision Desk HQ model.

Neither forecaster explores the possibility that El-Sayed's progressive positioning (AOC/Sanders-backed) could underperform generic-Democrat polling in a socially moderate swing state, which is a specific downside scenario the brief flags via the "progressive-vs-establishment theory" framing but the forecasts never engage with candidate-quality/ideology risk directly.

6. Aggregator Fable 5.1 · final calibrated forecast

63% (62% confidence)

Yes: 63% No: 37%

The only available market anchor is Polymarket at 61.5% Dem win, and both forecasters landed at 64% — a modest premium justified by the midterm environment under a Republican president (out-party wins open swing seats ~70% historically), Michigan's 32-year streak of Democratic Senate wins, a small but consistent polling-aggregate lead (+1.4 across 27 polls), and Decision Desk HQ's rising 66% estimate. The critique's downside points are real — El-Sayed's progressive profile and net-negative favorability, Cook's Toss-up/top-GOP-pickup framing, the WDIV poll with Rogers narrowly ahead, and the 0.3-pt 2024 Slotkin–Rogers result — but the 2024 race occurred with Trump on the ballot in a presidential year, whereas 2026 is a midterm with Trump's approval slipping, which structurally favors Democrats relative to that reference class. The critique's upside points (base rate ~70%, polling-error model 65-75%) roughly offset the downside ones, so I see no reason to move far from consensus. I settle at 63% Yes, essentially at consensus but shaded a hair toward the market to reflect genuine uncertainty over how a progressive nominee performs with undecideds in a swing state.

Pipeline Timing

Total pipeline time: 156.7s

Per-tool research timings shown in the Research section above.