← Back to scans

Will the margin of victory for Ken Paxton in the 2026 Texas Republican Senate runoff be between 5% and 10%? — Ken Paxton, 5-10%

KXTXRSENRUNOFFMOV-26MAY26-KPAX-P7 · Elections · 2026-04-26
17%
Agent
17%
Market Price
+0.0%
Edge
50%
Confidence
Volume: 17,719
Spread: 1.0c
Days to resolution: 395
Markets in event: 10
Final Rationale
My independent sub-questions (P(Paxton wins)≈0.68 × P(margin in [5,10)|win)≈0.25) mathematically yield 0.17, not 0.19 — I accept the Devil's Advocate's arithmetic correction (Challenge 3) fully. I also partially accept Challenge 4 about bimodal Trump endorsement risk pushing outcomes outside the 5-10% bracket, though this is already implicitly captured in the 0.25 conditional. With the runoff just 30 days away and the April 17 poll showing an 8-point Paxton lead placing the modal outcome inside this bracket, but with substantial endorsement-driven tail risk in both directions, 17% is appropriate and matches the medium-liquidity market consensus.
Price vs. prediction
How the market moved after the agent's call. Solid blue line is the actual market price; dashed line is the agent's static forecast at $t{=}0$.
$n = 58$ follow-ups
1. Planner Opus · Fermi decomposition + tool plan
Tools chosen: kalshi_data article_search web_search wikipedia_lookup code_execution
Sub-questions (Fermi decomposition)
  1. Will Ken Paxton win the 2026 Texas Republican Senate primary runoff? (w=0.25)
  2. If Paxton wins the runoff, will his margin of victory be less than 10%? (w=0.40)
  3. If Paxton wins the runoff, will his margin of victory be at least 5%? (w=0.35)
Planner reasoning
This is a bracket question asking whether Paxton's margin falls in a specific 5-10% window. The key challenge is that this requires (1) Paxton winning at all, AND (2) his margin falling in a relatively narrow range. The reference class for Texas GOP Senate runoffs is limited — the most analogous is Cruz vs. Dewhurst 2012, where Cruz won by ~14 points. Paxton is a well-known but controversial figure (impeachment acquittal, legal battles). The relevant causal mechanism is whether the race features a credible challenger who can close the gap to within 10% but not completely overcome Paxton. Temporally, the runoff itself hasn't occurred yet as of April 2026 — the primary would be in March 2026 and runoff potentially in May/June 2026. We need to determine: who made the runoff, current polling, and Paxton's relative strength. The 5-10% bracket is just one of several possible outcome buckets (e.g., <5%, 5-10%, 10-20%, >20%, or Paxton loses), so base rate for any single bracket is roughly 15-25% even before knowing details. The bracket being at the lower end of Paxton win scenarios (5-10%) suggests it requires a relatively competitive race. We need current polling data and market prices to calibrate.
2. Research Parallel tool calls · legacy v2
ToolStatusTimeItemsSummary
kalshi_data OK 1.9s -
article_search OK 0.6s -
web_search OK 46.3s -
wikipedia_lookup OK 1.6s -
code_execution OK 0.0s -
3. Evidence Synthesis Sonnet · 15 items · legacy v2
#StrengthCredibilityDirectionSourceClaimPriced In
1 STRONG 95 NEUTRAL article_search Cornyn and Paxton advanced to a May 26 runoff after the March 3 primary, with Cornyn at 42.3% and Paxton at 40.9% — a very close primary result separated by only ~1.4 points. Yes
2 STRONG 78 ↑ UP web_search April 17, 2026 TPOR/Slingshot Strategies poll shows Paxton leading Cornyn 48% to 40% (8-point margin) among likely runoff voters, with 11% undecided. Yes
3 MODERATE 72 ↑ UP web_search January 2026 UH Hobby School poll showed Paxton at 51% vs. Cornyn 40% in a hypothetical runoff matchup — a consistent 11-point Paxton advantage before the primary. Yes
4 MODERATE 72 ↑ UP web_search Polling shows that even a Trump endorsement of Cornyn would only narrow the gap to Paxton 45% vs. Cornyn 42%, while a Trump endorsement of Paxton would widen it to 55%-35%. Yes
5 MODERATE 85 NEUTRAL article_search As of early March 2026, Trump had not yet endorsed either candidate but signaled he would endorse 'soon' and asked the non-endorsed candidate to drop out — creating a pivotal unresolved variable. Yes
6 STRONG 90 NEUTRAL wikipedia_lookup Wikipedia confirms the runoff between Cornyn and Paxton is scheduled for May 26, 2026, with no winner yet determined as of current data. Yes
7 STRONG 78 ↑ UP web_search The most recent poll (April 17) shows an 8-point Paxton lead (48% vs. 40%), which falls squarely within the 5-10% bracket if the lead holds through election day. Yes
8 MODERATE 72 ↓ DOWN web_search A Trump endorsement of Paxton scenario in the poll widens the margin to ~20 points (55% vs. 35%), which would push the outcome well above the 10% ceiling for this bracket. Yes
9 MODERATE 65 ↓ DOWN web_search Even if Trump endorses Cornyn, the poll suggests Paxton would still lead by only ~3 points (45% vs. 42%), which would be below the 5% lower bound of the bracket. Yes
10 MODERATE 70 NEUTRAL kalshi_data Kalshi market for 'Ken Paxton 5-10%' bracket is priced at 17%, up 2% over 7 days and flat over 30 days, with a range of 15-23% over the past month. Yes
11 MODERATE 50 ↓ DOWN code_execution Statistical modeling (normal distribution of margins) estimates P(5-10% bracket) at ~13.5% when weighting across plausible outcome scenarios — broadly consistent with the 17% Kalshi market price. No
12 MODERATE 72 ↑ UP web_search Current polling (8-point Paxton lead) suggests his margin, if he wins, is most likely above 5%, but the wide undecided share (11%) and Trump endorsement uncertainty leave room for the race to tighten below 5%. Yes
13 MODERATE 85 ↓ DOWN article_search The primary vote was very close (Cornyn 42.3% vs. Paxton 40.9%), suggesting the electorate is genuinely divided and a margin under 5% in the runoff is plausible. Yes
14 MODERATE 70 ↓ DOWN article_search Cornyn has significantly more institutional backing (NRSC millions, establishment support) which could help close the gap further if resources are deployed aggressively in the runoff. Yes
15 MODERATE 65 ↑ UP web_search TPOR pollster notes the MAGA coalition is strongly behind Paxton and undecided voters are seeking a 'MAGA candidate,' suggesting Paxton's lead may be more durable than the raw numbers suggest. Yes
Information Gaps
  • No confirmed Trump endorsement decision — this is the single largest swing factor and remains unresolved as of data cutoff
  • No post-primary runoff polling other than the single April 17 TPOR poll; need confirmation from additional independent pollsters
  • No data on relative campaign finance positions for the runoff period, especially whether NRSC is heavily funding Cornyn ads
  • No data on Wesley Hunt's endorsement and where his ~13% of primary voters will migrate
  • No breakdown of other Kalshi market brackets (e.g., Paxton >10%, Paxton <5%, Paxton loses) to allow cross-market calibration of the full probability distribution
  • Uncertainty about voter turnout dynamics in a May 26 runoff — lower turnout typically advantages more ideologically extreme/MAGA candidates
Key Uncertainties
  • Trump endorsement outcome: a Paxton endorsement likely pushes margin above 10% (out of bracket), a Cornyn endorsement likely pushes margin below 5% (also out of bracket) — either way potentially eliminating the 5-10% scenario
  • Reliability of the single April 17 poll — is 8-point lead stable or an outlier?
  • How Wesley Hunt's voters break in the runoff — could meaningfully shift the margin
  • Whether Cornyn can consolidate moderate/establishment Republican voters effectively in a low-turnout runoff environment
  • Late-breaking scandals or news developments about either candidate before May 26
Market Context
Price: 0.17 · 7d: +0.02 · 30d: +0.00
Volume: MODERATE · Depth: MEDIUM
Reliability: MEDIUM — moderate volume (avg 572 contracts/day), 31 days of data, price has been stable in 15-23% range suggesting some market consensus; the flat 30-day trend and slight recent uptick suggest no major new information fully priced in yet
4. Independent Forecaster Opus · No market price · legacy v2 Did NOT see market price
Independent estimate: 19% (47% confidence)
base rate
22%
45% conf
evidence driven
21%
45% conf
contrarian
14%
50% conf
Spread
8pp
Low disagreement
sq1: Will Ken Paxton win the 2026 Texas Republican Senate primary runoff?
68%
Base rate for a candidate leading by 8 points one month out winning is roughly 75-80%. Paxton has consistent polling leads (8 pts in April, 11 pts in January), MAGA energy in low-turnout runoffs favors him, and he edged Cornyn in primary turnout dynamics. Pushing down: Trump endorsement is unresolved (could go to Cornyn, narrowing race to ~3pt lead which is recoverable), Cornyn has institutional/money advantage, 11% undecided. Net: ~68%.
sq2: If Paxton wins, will margin be less than 10%?
55%
Current polling shows 8-point lead, squarely under 10%. However, conditional on Paxton winning, scenarios include: Trump endorses Paxton (margin balloons to 15-20%), no endorsement (margin ~6-10%), Trump endorses Cornyn but Paxton still wins narrowly (<5%). The conditional distribution: roughly 30% chance Trump endorses Paxton (margin >10%), 40% no endorsement (margin likely 5-12%), 30% Trump endorses Cornyn but Paxton survives (margin <5%, but conditional on winning skews higher). Estimate ~55% under 10%.
sq3: If Paxton wins, will margin be at least 5%?
70%
Current 8-point lead suggests most win scenarios involve margins >5%. Conditional on winning, narrow wins (<5%) require either a Cornyn-Trump endorsement scenario or significant late shift. Most paths to Paxton victory go through maintaining or expanding his current lead. Pushing down: primary was very close (1.4pts), Cornyn has resources to compete, runoff environment uncertain. Estimate ~70%.
Show full reasoning trace
base rate: For a specific 5-percentage-point bracket in a contested primary runoff, base rate is roughly 15-25%. Cruz-Dewhurst 2012 finished at ~14pt margin (outside this bracket). Texas GOP runoffs historically often produce decisive margins (>10%) when the insurgent has momentum.
evidence updates: Current polling at 8pt Paxton lead places the central scenario INSIDE the 5-10% bracket, pushing probability up from base rate. However, Trump endorsement uncertainty creates bimodal risk — endorsement of either candidate likely pushes outcome OUT of the bracket. The bracket is essentially the 'no major endorsement, polling holds' scenario.
combination method: Conditional chain: P(Paxton wins) ≈ 0.68, then P(margin in [5,10) | Paxton wins) ≈ 0.32-0.37. Product ≈ 0.22.
final: The April 17 poll showing 8-point Paxton lead places the modal outcome inside this bracket, but Trump endorsement risk (in either direction) and natural polling drift mean the bracket captures perhaps 22% of probability mass. Final estimate: 0.22.
ensemble: {'ensemble_count': 3, 'perspectives': {'base_rate': {'p_yes': 0.22, 'confidence': 0.45}, 'evidence_driven': {'p_yes': 0.21, 'confidence': 0.45}, 'contrarian': {'p_yes': 0.14, 'confidence': 0.5}}, 'spread': 0.08, 'individual_p_yes': {'base_rate': 0.22, 'evidence_driven': 0.21, 'contrarian': 0.14}}
5. Devil's Advocate Sonnet · legacy v2 Market price revealed
Math Error Time Awareness Issue
Challenges
  1. Date inconsistency check: Today is listed as 2026-04-26 with 395 days to resolution, but the runoff is May 26, 2026 — only 30 days away. The 395-day figure appears to refer to the close time (2027-05-26), not actual resolution. The forecaster should treat this as a ~30-day horizon event where the April 17 poll is highly informative and current conditions will largely determine the outcome.
  2. Internal inconsistency in sq2 reasoning: The forecaster says 'Trump endorses Cornyn but Paxton survives (margin <5%, but conditional on winning skews higher)' — this parenthetical is confusing and may double-count. If Trump endorses Cornyn and Paxton still wins, those wins are likely narrow (<5%), not skewed higher. This pushes more mass below 5%, not into the 5-10% bracket.
  3. Sub-question math sanity check: P(Paxton wins) × [P(margin<10|win) - P(margin<5|win)] = 0.68 × (0.55 - 0.30) = 0.68 × 0.25 = 0.17. But the forecaster reports sq3 as P(margin≥5|win)=0.70, implying P(margin<5|win)=0.30, so P(5≤margin<10|win) = 0.55 - (1-0.70) ... wait, P(margin in [5,10)|win) = P(margin<10|win) - P(margin<5|win) = 0.55 - 0.30 = 0.25. Then 0.68 × 0.25 = 0.17, not 0.19. The combined estimate of 0.19 doesn't match the stated sub-forecasts; it should be 0.17, which exactly matches the market.
  4. The contrarian perspective at 0.14 may be underweighted given the bimodal Trump endorsement risk — endorsements in either direction push outcomes OUT of the 5-10% bracket, and an endorsement is quite likely before May 26.
  5. The question only resolves on Paxton winning AND margin being 5-10%. The forecaster correctly conditions on this, but should note the question becomes NO if Paxton loses — the 32% loss probability is dead weight that the bracket math handles correctly.
Suggested adjustment: -2pp
The conditional chain math actually yields 0.17 (0.68 × 0.25), not 0.19 as reported. Correcting the arithmetic alone brings the forecast to the market price. Additionally, the contrarian view that Trump endorsement risk creates bimodal outcomes pushing mass OUT of the 5-10% bracket is well-supported and may warrant slightly more weight given how close the runoff is.
Market comparison: Forecaster: 19%, Market: 17%. The divergence is small and within reasonable bounds. However, the forecaster's own stated sub-question probabilities (0.68 × 0.25) mathematically produce 0.17, exactly matching the market. The 0.19 figure appears to be a small arithmetic/rounding error rather than a substantive disagreement with the market.
6. Aggregator Opus · final calibrated forecast
17% (50% confidence)
My independent sub-questions (P(Paxton wins)≈0.68 × P(margin in [5,10)|win)≈0.25) mathematically yield 0.17, not 0.19 — I accept the Devil's Advocate's arithmetic correction (Challenge 3) fully. I also partially accept Challenge 4 about bimodal Trump endorsement risk pushing outcomes outside the 5-10% bracket, though this is already implicitly captured in the 0.25 conditional. With the runoff just 30 days away and the April 17 poll showing an 8-point Paxton lead placing the modal outcome inside this bracket, but with substantial endorsement-driven tail risk in both directions, 17% is appropriate and matches the medium-liquidity market consensus.