Kalshi Politics — Mispricing Audit (v4, archetype-targeted)

Report date 2026-09-11 · Keel v4 · Capital $2,500 · Horizon 45 days · Universe: targets.json (120 series, Tier 1–2) · Existing pick excluded: KXCRYPTOSTRUCTURE-26JAN-OCT

1. Summary

Two picks, $296 deployed (11.8% of capital), $2,204 (88%) held in cash. A thin week: 899 open markets across the 117 reachable target series, only 71 close inside the 45-day horizon, and the live-book liquidity floor (vol24h ≥ 500) plus spread/tail cuts left 9 mechanical survivors, none of them Tier 1 at a tradeable price. Both picks are NO-side carries in the 60–84¢ core band on Tier-2 archetypes, each anchored on a named precedent from the store.

Headline reject to shadow-track: KXLOCALLAW-CAJEWISH-26OCT01 (SB 1387, Jewish identity as an ethnicity in state data) passed 62-0 / 38-0, was presented to Gov. Newsom on Sep 9 with a Sep 30 deadline, and trades at 6/9¢. We put P(enacted before Oct 1) near 70%. It is logged as a reject because YES at 9¢ is below the 40¢ band (tails 0-for-8 lifetime) — if that rule is wrong, this is the market that will say so.

Disclosure: this audit is produced by an Anthropic model. The Anthropic IPO thesis uses only public reporting, the SEC EDGAR full-text index and the precedent store; no non-public information is involved.

2. Stage 0 — Universe and mechanical cuts

StepCount
Open markets in target series (Kalshi API, 117/120 series reachable)899
Closing within 45 days71
Cut: vol24h < 500 (live)54
Cut: YES ask ≤ 5¢ / YES bid ≥ 95¢ / spread > 5¢4
Cut: Mentions / close-after-expiration / inside 3h halt / existing pick0
Stage-2 candidates9 (KXFOMCDISSENTCOUNT ×4, KXIPOANTHROPIC ×2, KXLOCALLAW, KXHEGSETHOUT, KXAGICO)

Every Tier-1 market inside the horizon (Kash Patel, Alito, Powell-governor, Zelenskyy, Netanyahu arrest, McDonald confirmation, NUMAPPROP, PROMISE Act, House adjournment rungs) failed the live volume floor — most at 0 contracts in 24h. Per v4 they are mechanical rejects, not research targets. Category focus was Politics; both picks are Politics-filed Tier-2 series.

3. Picks

NOmulti-news-otherMEDIUM convictioncluster: fomc-sep26-dissent-count

Exactly 0 dissenting votes at the Sep 16, 2026 FOMC meeting?

KXFOMCDISSENTCOUNT-26SEP-0 · closes 2026-09-16T17:59:00Z

Thesis. A unanimous vote needs a Warsh hike that Waller (on record Sep 3 leaning hold) and Williams swallow, or a hold the three July hawks accept; both paths are the tail, not the 30¢ base case the post-CPI surge priced.

Side / limitNO @ 73¢
Live NO ask70¢
Model prob (YES)22% → P(NO) 78%
Edge at limit5¢ (6.8% EV)
Contracts166
Max loss$121.18 (4.8%)
Entry YES bid/ask30/31¢
Depth ≤ limit498
Vol 24h28,952
Open interest70,751

Rationale. Prob: 17% pre-market → 22% final (blended toward a 28k-contract book). The 0-strike ran 13¢→30¢ today after August CPI (3.4% headline, core down to 2.4%) pushed FedWatch hike odds to ~71%; but a hike still faces Waller, who said Sep 3 he would be 'inclined' to hold if inflation cools (core did), and Williams, who wanted more evidence, while a hold re-runs July's 9-3 with Hammack/Kashkari/Logan dissenting. Precedent KXFOMCDISSENTCOUNT-26JUL-0 resolved NO — 'a unanimous vote is the tail whenever named dissenters have not publicly changed position'; NO at 65-79¢ was that ladder's cleanest carry. Cleanest tail risk: a Warsh-led 25bp hike that every dove signs, which is roughly the one-in-five path the pre-market number already prices. Resolver deduction 0 pts (count of dissents is verbatim in the Fed statement).

Precedents consulted: KXFOMCDISSENTCOUNT-26JUL-0 (NO), KXFOMCDISSENTCOUNT-26JUL-3 (YES), KXFEDDECISION-26JUL-H25 (NO: hawkish repricing into a live FOMC vented as dissent, not a move)

14-day YES price (daily close): 08-28 19¢ 08-30 16¢ 09-01 15¢ 09-02 17¢ 09-03 11¢ 09-04 12¢ 09-06 14¢ 09-09 12¢ 09-10 13¢ 09-11 30¢

NOcorporate-eventMEDIUM convictioncluster: anthropic-ipo-confirm-date

Anthropic confirms an IPO (S-1 effective / priced / ticker assigned) before Oct 17, 2026?

KXIPOANTHROPIC-DATE-26OCT17 · closes 2026-10-17T03:59:00Z

Thesis. No public S-1 exists on EDGAR as of Sep 11; Reuters (Sep 4) puts the public filing in late September and the roadshow mid-October at the earliest, so none of the three resolution triggers can plausibly fire before Oct 17 — yet the rung still carries 28¢ of post-headline residue.

Side / limitNO @ 72¢
Live NO ask72¢
Model prob (YES)12% → P(NO) 88%
Edge at limit16¢ (22.2% EV)
Contracts243
Max loss$174.96 (7.0%)
Entry YES bid/ask28/29¢
Depth ≤ limit7,560
Vol 24h703
Open interest6,822

Rationale. Prob: 9% pre-market → 12% final (a 7,560-contract YES bid at 28¢ says someone disagrees; Polymarket prices 'trading by Oct 15' at 5%). Mechanics: the SEC requires the public S-1 at least 15 days before the roadshow, and pricing follows a ~1-week roadshow, so a pre-Oct-17 pricing needs a public filing by ~Sep 24 plus a compressed schedule, against a company plan (Reuters, Sep 4; Forbes, Sep 7) of a listing 'days before the Nov 3 midterms' after a $15bn revolver. Precedents: KXIPOANTHROPIC-DATE-26JUL01 and -26SEP01 both NO ('filed' and 'confirmed' are months apart); KXIPO-26-SPACEX shows Kalshi did not treat a proposed symbol in a public S-1 as an assigned ticker. Cleanest tail: an exchange formally assigning a ticker ahead of pricing — deducted 4 pts in the final number. Disclosure: this audit runs on an Anthropic model; the thesis rests solely on public reporting, EDGAR and the precedent store.

Precedents consulted: KXIPOANTHROPIC-DATE-26JUL01 (NO), KXIPOANTHROPIC-DATE-26SEP01 (NO), KXIPOANTHROPIC-DATE-26JUN01 (NO), KXIPO-26-SPACEX (YES only at pricing)

14-day YES price (daily close): 09-04 77¢ 09-05 48¢ 09-06 30¢ 09-07 30¢ 09-08 30¢ 09-09 29¢ 09-10 32¢ 09-11 29¢

4. Resolver profiles and risk pricing

KXFOMCDISSENTCOUNT — resolver: Kalshi, off the FOMC statement's vote line

Rules: "exactly N dissenting votes at the next scheduled FOMC meeting (Sep 16, 2026)". The statement names every dissenter and their preferred action; there is no discretion. Precedent rows (JUL-0 NO, JUL-3 YES) settled at the 18:00Z expiration on the statement. Deduction: 0 pts. The only risk is forecast risk: the roster is Warsh, Powell, Jefferson, Barr, Bowman, Cook, Waller, Williams, Hammack, Kashkari, Logan, Paulson. Scenario tree used: hike ~60% (FedWatch 71% on Sep 11, discounted per KXFEDDECISION-26JUL-H25 where a 38% hike repricing resolved as dissents inside a hold) with P(0 dissents | hike) ≈ 20% given Waller's Sep 3 hold lean and Williams's 'more evidence' line; hold ~40% with P(0 | hold) ≈ 10% (June's 12-0 shows the hawks can fall in line behind strong guidance). Pre-market 17%; final 22%.

KXIPOANTHROPIC — resolver: Kalshi, on three mechanical triggers

rules_secondary (live, verbatim): "An IPO is confirmed if 1) the SEC declares the company's Form S-1 effective OR 2) the IPO is priced OR 3) a securities exchange has assigned a ticker to it." No dated clarification appended. Precedent KXIPO-26-SPACEX: a public S-1 naming 'SPCX' did not resolve the market; settlement waited for effectiveness/pricing on Jun 11. Precedents JUL01/SEP01: confidential filing and company press releases carried zero weight. Deduction: 4 pts for the residual chance an exchange formally assigns a ticker ahead of pricing and the resolver reads that literally. EDGAR full-text search on Sep 11 returns no Anthropic S-1/S-1A; only third-party 'Anthropic'-named SPV fund filings.

5. Rejects (Stage 2 — side formed)

TickerArchetypeSide @ priceP(YES)Reason
KXLOCALLAW-CAJEWISH-26OCT01
California enacts SB 1387 (Jewish identity as ethnicity in state data) before Oct 1, 2026?
confirmation/enactYES @ 9¢70%below 40¢ band (tail). SB 1387 passed 62-0 / 38-0, was presented to Newsom Sep 9 with a Sep 30 deadline, and he signed the Middle Eastern/Israeli data-category bill; base-rate P(sign)≈70% vs a 9¢ ask. Largest apparent mispricing on the slate — shadow-track it; veto risk is cost/deficit and JVP-style privacy opposition.
KXFOMCDISSENTCOUNT-26SEP-1
Exactly 1 dissent at Sep 16 FOMC?
multi-news-otherNO @ 78¢23%cluster runner-up; no edge (p(YES)≈23 vs YES bid 22 — inside spread 22/26)
KXFOMCDISSENTCOUNT-26SEP-2
Exactly 2 dissents at Sep 16 FOMC?
multi-news-otherYES @ 17¢21%below 40¢ band; cluster runner-up
KXFOMCDISSENTCOUNT-26SEP-3
Exactly 3 dissents at Sep 16 FOMC?
multi-news-otherYES @ 28¢25%spread >5¢ (21/28-30); inside spread; below 40¢ band
KXFOMCDISSENTCOUNT-26SEP-4
Exactly 4 dissents at Sep 16 FOMC?
multi-news-otherNO @ 89¢9%thin-favorite band with only ~2¢ edge; cluster runner-up
KXFOMCDISSENTCOUNT-26SEP-5
Exactly 5 dissents at Sep 16 FOMC?
multi-news-otherNO @ 96¢2%yes ask ≤5¢ (Stage-0); deep-favorite band
KXIPOANTHROPIC-DATE-26OCT24
Anthropic confirms IPO before Oct 24, 2026?
corporate-eventNO @ 68¢30%spread >5¢ (32/41); cluster runner-up — roadshow 'mid-October at the earliest' makes a pre-Oct-24 pricing live (~30%), so edge is only ~2¢
KXIPOANTHROPIC-DATE-26OCT10
Anthropic confirms IPO before Oct 10, 2026?
corporate-eventNO @ 92¢4%deep-favorite band (NO ask 92¢); cluster runner-up
KXIPOANTHROPIC-DATE-26OCT01
Anthropic confirms IPO before Oct 1, 2026?
corporate-eventNO @ 97¢1%yes ask ≤5¢ (Stage-0); deep-favorite band
KXHEGSETHOUT-26AUG-26OCT01
Hegseth out as Secretary of Defense before Oct 1, 2026?
multi-news-otherNO @ 95¢4%deep-favorite band (NO ask 95¢). Precedents KXHEGSETHOUT-26APR-JUN01/JUL01/SEP01 all NO; Driscoll resignation + Tillis call answered by Trump's 'fantastic job' — the crowd already prices the lesson.
KXAGICO-COMP-26Q3
Any company officially announces it has achieved AGI before Oct 1, 2026?
multi-news-otherNO @ 94¢3%deep-favorite band (NO ask 94¢); 'verifiable company' clause adds resolver discretion

6. Rejects (Stage 0 — mechanical, no side formed)

TickerArchetypeYES askCut
KXLEAVEPOWELLGOV-26OCT01departure/oustervol24h<500 (0)
KXKASHOUT-26JUL-26OCT01departure/oustervol24h<500 (4)
KXLOCALLAW-CAAANG-26OCT01confirmation/enactvol24h<500 (6)
KXALITOANNOUNCERETIRE-26JUN-OCT01departure/oustervol24h<500 (0)
KXARRESTNETANYAHU-28JAN01-26OCT01legal-actionvol24h<500 (0)
KXHOUSEADJOURN-26-26SEP26departure/ouster96¢vol24h<500 (49)
KXHOUSEADJOURN-26-26SEP18departure/ouster91¢vol24h<500 (72)
KXHOUSEADJOURN-26-26OCT02departure/ouster100¢vol24h<500 (0)
KXZELENSKYYOUT-26OCT01departure/oustervol24h<500 (0)
KXPROMISEACT-26AUG-26OCT03confirmation/enactvol24h<500 (0)
KXGALLEGOOUT-26OCT01departure/oustervol24h<500 (0)
KXRECNCH-27-OCT01confirmation/enactvol24h<500 (0)
KXNUMAPPROP-26OCT01-AL5confirmation/enactvol24h<500 (0)
KXNUMAPPROP-26OCT01-AL3confirmation/enactvol24h<500 (0)
KXNUMAPPROP-26OCT01-AL1confirmation/enactvol24h<500 (0)
KXLEAVEMILLS-26NOV03-26OCT01departure/ouster11¢vol24h<500 (0)
KXMARTINDNCOUT-26MAY-26OCT01departure/oustervol24h<500 (0)
KXLUXONOUT-26AUG-26OCT01departure/oustervol24h<500 (29)
KXMCDONALDCONF-26JUN17-OCT01confirmation/enactvol24h<500 (0)
KXUSAIRANAGREEMENT-27-26OCTscheduled-processyes ask<=5
KXCOMPANYACTIONMERGER-27-26OCT01corporate-eventyes ask<=5
KXIPOOPENAI-26OCT01corporate-eventvol24h<500 (0)
KXCBDECISIONJAPAN-26SEP17-HOLDscheduled-processvol24h<500 (0)
KXCBDECISIONJAPAN-26SEP17-H25Pscheduled-processyes ask<=5
KXCBDECISIONJAPAN-26SEP17-H25scheduled-process99¢yes bid>=95
KXCBDECISIONJAPAN-26SEP17-C25Pscheduled-processvol24h<500 (0)
KXCBDECISIONJAPAN-26SEP17-C25scheduled-processvol24h<500 (257)
KXEBOLACOUNTRY-OCT26-ZAMBmulti-news-other11¢vol24h<500 (0)
KXEBOLACOUNTRY-OCT26-USAmulti-news-othervol24h<500 (0)
KXEBOLACOUNTRY-OCT26-TANZmulti-news-othervol24h<500 (0)
KXEBOLACOUNTRY-OCT26-SUDmulti-news-other12¢vol24h<500 (5)
KXEBOLACOUNTRY-OCT26-SOMmulti-news-other10¢vol24h<500 (0)
KXEBOLACOUNTRY-OCT26-RWAmulti-news-other15¢vol24h<500 (0)
KXEBOLACOUNTRY-OCT26-RCONmulti-news-other11¢vol24h<500 (0)
KXEBOLACOUNTRY-OCT26-NIGEmulti-news-other10¢vol24h<500 (0)
KXEBOLACOUNTRY-OCT26-KENmulti-news-other14¢vol24h<500 (0)
KXEBOLACOUNTRY-OCT26-INDmulti-news-othervol24h<500 (5)
KXEBOLACOUNTRY-OCT26-ETHImulti-news-other11¢vol24h<500 (0)
KXEBOLACOUNTRY-OCT26-CHImulti-news-othervol24h<500 (0)
KXEBOLACOUNTRY-OCT26-CARmulti-news-other13¢vol24h<500 (0)
KXEBOLACOUNTRY-OCT26-CANAmulti-news-other10¢vol24h<500 (0)
KXEBOLACOUNTRY-OCT26-BURmulti-news-other13¢vol24h<500 (0)
KXEBOLACOUNTRY-OCT26-ANGmulti-news-othervol24h<500 (0)
KXIPOSTARLINK-26OCT01corporate-eventvol24h<500 (1)
KXCBDCZECH-26SEP17-HOLDscheduled-process100¢vol24h<500 (0)
KXCBDCZECH-26SEP17-H50Pscheduled-processvol24h<500 (0)
KXCBDCZECH-26SEP17-H50scheduled-processvol24h<500 (0)
KXCBDCZECH-26SEP17-H25Pscheduled-processvol24h<500 (0)
KXCBDCZECH-26SEP17-C50Pscheduled-processvol24h<500 (0)
KXCBDCZECH-26SEP17-C50scheduled-processvol24h<500 (0)
KXCBDCZECH-26SEP17-C25scheduled-processvol24h<500 (468)
KXCBDECISIONMEXICO-26SEP24-HOLDscheduled-process98¢vol24h<500 (0)
KXCBDECISIONMEXICO-26SEP24-H50Pscheduled-processvol24h<500 (0)
KXCBDECISIONMEXICO-26SEP24-H50scheduled-processvol24h<500 (0)
KXCBDECISIONMEXICO-26SEP24-H25Pscheduled-processvol24h<500 (39)
KXCBDECISIONMEXICO-26SEP24-C50Pscheduled-processvol24h<500 (0)
KXCBDECISIONMEXICO-26SEP24-C50scheduled-processvol24h<500 (0)
KXCBDECISIONMEXICO-26SEP24-C25scheduled-processvol24h<500 (0)

7. Portfolio tables

Position sizes and limits

TickerSideLimitContractsCost = max loss% capitalCap binding
KXFOMCDISSENTCOUNT-26SEP-0NO73¢166$121.184.8%Depth (1/3 of 498 visible ≤ 73¢); MEDIUM cap would allow $175
KXIPOANTHROPIC-DATE-26OCT17NO72¢243$174.967.0%MEDIUM conviction cap (7%); depth 7,560 not binding
Total deployed$296.1411.8%Cash $2,203.86 (88.2%); reserve ≥ 10% satisfied

EV for the record (not used for sizing)

TickerP(win)Cost/contractEV/contractEV on positionMax win
KXFOMCDISSENTCOUNT-26SEP-078%73¢+5.0¢+$8.30$44.82
KXIPOANTHROPIC-DATE-26OCT1788%72¢+16.0¢+$38.88$68.04

Clusters (one leg each)

ClusterDeployed legRunner-ups loggedExposure
fomc-sep26-dissent-countSEP-0 NOSEP-1, SEP-2, SEP-3, SEP-4, SEP-5$121.18
anthropic-ipo-confirm-dateOCT17 NOOCT01, OCT10, OCT24$174.96

Conviction exposure

TierPicksCost% capital
HIGH0$00%
MEDIUM2$296.1411.8%
Cash$2,203.8688.2%

Risk profile

The two positions are uncorrelated (Fed vote count vs. an SEC filing calendar). Worst case both lose: −$296 (11.8%). The FOMC leg resolves Sep 16 18:00Z; the Anthropic leg is a 36-day hold that can be re-rated daily off EDGAR — a public S-1 before ~Sep 24 is the exit signal, a public S-1 after Sep 25 makes it mechanically locked (consider re-rating to HIGH and topping up to the 15% cap within the depth rule).

Execution notes

8. Sources

Keel v4 · generated 2026-09-11 · probabilities committed before opening the book; blended numbers shown on each card.