# Kalshi Politics — Mispricing Audit (Anchor, NO-side structural) — 2026-10-20

Date: 2026-10-20
Source: https://kalshi-audits.pages.dev/2026-10-20-kalshi-politics-mispricing-audit-anchor

---
# Kalshi Politics — Mispricing Audit (Anchor, NO-side structural)

Report date 2026-10-20 · strategy=Anchor · audit_version=anchor · horizon 45d · capital $2,500 · category focus Politics (advisory) · universe = targets.json (120 allowlisted series, generated 2026-10-20T16:36Z)

## Summary

Picks1NO · status-quo · 85–89¢ band (half size)
Deployed$1255.0% of capital
Cash held$237595.0% — cash is a position
Logged rejects214 researched, 17 mechanical/band

**One leg survives.** After the Stage-0 mechanical cuts (live vol24 ≥ 500, spread ≤ 5¢, no close-after-expiration, no sub-7-day, no count/level markets) only eight of 878 open markets in the allowlist were candidates, and after the Anchor side rule and bands only one is deployable: **NO on “Hegseth out as Secretary of Defense before Dec 1, 2026”** (status quo; wartime defense chief; the only dated pressure vehicle — Dec 11 defense funding — lands after the rung closes; an announced-but-not-vacated resignation pays NO). It sits in the 85–89¢ thin-favorite band because the visible book on the NO side is thin below 86¢, so it is sized at half (5% of capital). The Schumer step-down rung is the cleaner thesis but is only fillable at 90¢ and is rejected on band. Both Bank of Canada legs are yes-side theses in substance (they forecast the decision's content), and the Anthropic IPO leg is below band on the NO side with no public S-1 on EDGAR to permit a YES.

Note: yesterday's Anchor run (Oct 19) also signaled the Hegseth Dec 1 leg NO at 85¢. It is not in `$EXISTING_PICKS`, so it is re-issued here; the bot should dedupe on ticker rather than add a second position.

## Picks

NO · BUYMechanism: status-quoConviction: mediumdeparture/ouster · Tier 2 (multi-news)

### KXHEGSETHOUT-26AUG-26DEC01 — Hegseth out as Secretary of Defense before Dec 1, 2026?

Limit NO 86¢ (fillable 83–86¢, walks the YES bids at 17/16/15/14)Contracts 145 (cost ≈ $125, max loss $125, max win $20)
Prob (YES) 10% pre-market → 12% finalEdge / EV fair NO 88¢ vs 86¢ limit → +2¢ · EV +2.3% (recorded, not used for sizing)
Entry book YES 17 / 20 · last 19 · vol24 706 · OI 88,179 · 808 NO contracts visible at ≤86¢ (cap 269)Close 2026-12-01 04:59Z (41 days) · expected expiration 2029-01-01 (closes early on departure)
Cluster hegseth-sod-exit48h move YES 21 → 19 (−2¢); Oct 15 spike to 36¢ on the Fort Hood livestream story fully reversed

**YES condition.** For YES to pay, Hegseth must actually vacate the Secretary of Defense role before Dec 1, 2026 04:59Z, resolved by Kalshi on the vacancy. `rules_secondary`: “must have an actual departure date and vacate the role within the time period”; temporary leaves, suspensions and recusals do not count. An announced resignation with a December effective date pays NO.

**Thesis.** Status quo: Hegseth is still Secretary of Defense on Dec 1 — no removal vehicle closes inside the window, the Dec 11 funding leverage lands after the rung closes, an announced-but-not-vacated resignation pays NO, and the Iran lull is slated to end after Nov 3 with Hegseth in Israel today planning it.

**Vehicle and calendar (primary sources).**

- **No removal vehicle closes inside the window.** Senate Democrats are floating Hegseth's ouster as the price of lame-duck cooperation on defense bills, tied to Pentagon funding that expires **Dec 11** (The Hill, “Democrats float Pete Hegseth's ouster in lame-duck defense battle”, Oct 19). That deadline is ten days after this rung closes, and a GOP senator called pushing him out “a bridge too far”. The House impeachment articles (Rep. Ansari, April) sit in a Republican House that does not return until Nov 9 and whose Speaker has said a privileged resolution would be tabled.
- **Wartime defense chief.** The US has said it will not strike Iran before Nov 3 but Trump said on Oct 13 the pause “might be running out in a few weeks”; Hegseth made an unannounced visit to Israel on Oct 19–20 to meet Netanyahu about a possible restart of operations (CBS News; Washington Times, Oct 20). Precedent `KXHEGSETHOUT-26APR-JUL01`: “administrations almost never change that office mid-conflict”.
- **Trump's stated position.** White House: “utmost confidence” (Aug); on Oct 15 the Pentagon said Trump and Hegseth are “completely aligned” after the livestream reversal. No successor is being floated — the Lutnick precedent's tell for a real exit.
- **Series record.** Four consecutive rungs (JUN01, JUL01, SEP01, OCT01) resolved NO through Signalgate, the USS Lincoln deployment, the Army Secretary's Sep 1 resignation and Sen. Tillis's Sep 2 replacement call.

**Precedents consulted.** `KXHEGSETHOUT-26APR-JUL01` (NO — active-war structural NO); `KXHEGSETHOUT-26APR-JUN01` (NO — “require an affirmative White House signal before paying”); `KXKASHOUT-26APR-JUN01` (NO — anonymous ‘matter of time’ sourcing is not a resolution event); `KXLUTNICKOUT-26JUL01` (NO — absent a named successor, price a single month near base rate); `KXTRUMPADMINLEAVE-26DEC31-TGAB` (YES — the series settles on the actual last day in office, which is the cushion here); `KXHEGSETHOUT-26AUG-26OCT01` and `-26APR-SEP01` verified settled NO via the public API this run.

**Resolver profile and risk pricing.** Resolver: Kalshi on official announcement of vacancy. Risks: (a) acting-secretary designation while Hegseth is nominally on leave — rules say leaves do not count, −1 pt; (b) death clause settles at last fair price — negligible; (c) a resignation announced in November with an effective date before Nov 30 — this is the real path and is inside the probability, not a resolver risk. Total deductions ≈ 1 pt (well under the 15-pt reject line).

**Cleanest tail.** A new discrete event: Trump fires him in a post-midterm reshuffle (the 2018 Sessions precedent, the day after the election) and the vacancy occurs before Nov 30, or Hegseth resigns at once if Democrats win the House and promise investigations. Both require Trump to reverse a stated position inside a planned restart of combat operations; this tail is ~10 of the 12 points. It is not a continuation of anything already underway, so the continuation-tail rule is satisfied. Why medium, not high: the Oct 13 Cook pick in this same strategy lost to exactly this kind of abrupt presidential action.

**Why half size.** Fillable NO depth is 100 contracts at ≤85¢ and 808 at ≤86¢; a 86¢ effective entry is in the 85–89¢ thin-favorite band → 5% of capital, 145 contracts, under the 269-contract depth cap.

## Position sizes and limits

| Ticker | Side | Limit | Contracts | Cost = max loss | Max win | Depth at ≤ limit | Depth cap (1/3) | Band |
|---|---|---|---|---|---|---|---|---|
| `KXHEGSETHOUT-26AUG-26DEC01` | NO | 86¢ | 145 | $124.70 | $20.30 | 808 | 269 | 85–89¢ → half size (5%) |
| Cash | — | — | — | $2375.30 | — | — | — | reserve ≥10% satisfied |

## EV for the record

| Ticker | Side | Market mid (YES) | Model P(YES) | Fair NO | Entry | Edge | EV % | Conviction |
|---|---|---|---|---|---|---|---|---|
| `KXHEGSETHOUT-26AUG-26DEC01` | NO | 18.5¢ | 12% | 88¢ | 86¢ | +2¢ | +2.3% | medium |

Sizing is flat and EV-agnostic by design; EV is recorded for the scoreboard only.

## Cluster table

| Cluster | Legs considered | Deployed | Runner-ups / rejects |
|---|---|---|---|
| hegseth-sod-exit | DEC01, NOV01 (FEB01/JUL01 rungs are beyond the 45d horizon) | DEC01 NO | NOV01: deep-favorite band |
| schumer-leader-call | NOV03 | — | deep-favorite band (fillable NO 90¢) |
| boc-oct28-decision | H0, H25, C25 | — | yes-side thesis ×2, deep-favorite band |
| anthropic-ipo-announce | DEC01, NOV01 | — | below band; deep-favorite band |
| trump-impeach / ca-wealth-tax / alberta-refs | 6 legs | — | close after expiration (Stage-0 mechanical) |
| tesla-spacex-merger / us-iran-deal / mcconnell-resign / mamdani-eo-count / diesel-export-ban | 7 legs | — | deep-favorite band, continuous-number, active pick |

## Risk profile

- **Capital at risk:** $125 (5.0%) in a single NO leg; worst case is −$125. Cash $2375 (95.0%).
- **Concentration:** one cluster, one resolver mechanism (Kalshi vacancy call on a cabinet office). No correlation with the existing diesel-export-ban position.
- **Event risk inside the window:** Nov 3 midterms (post-election reshuffle is the tail), Nov 9 House return (impeachment resolution tabling), Nov 26-ish possible Iran restart. None of these is a vehicle that removes him by itself.
- **Scoreboard context:** Anchor's last resolved picks: KXIPOANTHROPIC-DATE-26OCT17 NO won; KXTRYFIRECOOK NOV01 NO lost to abrupt presidential action — the same tail shape as this pick's.

## Execution notes

- Enter as a NO limit at 86¢ for 145 contracts; the book will fill 1 @83, 34 @84, 65 @85 and the remainder @86 for an average near 85.1¢. Do not chase above 86¢; above 87¢ the leg leaves the band's edge.
- If the YES side is swept by a pressure headline (as on Oct 15), a NO entry at 80–84¢ becomes available and would move the leg into the core band at full size — that is the only reason to revisit before Dec 1.
- Exit condition: an official vacancy announcement with an effective date before Nov 30, or a named successor being announced — otherwise hold to settlement.
- Dedupe: the Oct 19 Anchor signal on the same ticker (NO @85, 147 contracts) should be merged, not stacked.

## Universe and Stage-0 cuts

878 open markets across the 120 allowlisted series were pulled live from the public API. Mechanical pass: 8 markets (`KXMAMDANIEO-26OCT24-T0`, `KXDEMSCHUMER-27-NOV03`, `KXCBDECISIONCANADA-26OCT-H0/H25`, `KXIPOANTHROPIC-DATE-26DEC01`, `KXCOMPANYACTIONMERGER-27-26DEC01`, `KXUSAIRANAGREEMENT-27-26DEC`, `KXHEGSETHOUT-26AUG-26DEC01`). Most of the Tier-1 slate (KXTRUMPADMINLEAVE, KXFEDERALCHARGE, KXTRUMPOUT27, KXSAVEACT, KXLEAVEWALZ, KXTRUMPPARDON…) closes Dec 31 – Jan 4, outside the 45-day horizon, and is not in scope this run. The ballot-measure and impeachment ladders carry an expected expiration of Nov 3, 2026 against a close a year or more later and are cut as “close after expiration”. The Suez weekly ladder closes in 6 days (inside 7d) and is a count market. No Tier-3 series exist in today's file.

## Rejects log (all Stage-2 survivors and mechanical cuts, with the side we would have taken)

| Ticker | Archetype | Side @ price | P(YES) | Cluster | Reason |
|---|---|---|---|---|---|
| `KXDEMSCHUMER-27-NOV03` Dem Senate caucus member publicly says Schumer should step down before Nov 3, 2026? | departure/ouster | NO @ 90¢ | 7% | schumer-leader-call | deep-favorite band: fillable NO is 90¢ (best YES bid 10¢, 250 contracts) and there is zero visible depth at ≤89¢, so the only honest entry is ≥90¢. Thesis itself is clean status-quo/calendar (close is the night before Election Day; WaPo/PBS Oct 12: 'not one of the Democrats calling for Schumer to step down as leader is a current member of his caucus'; Sanders/Warren/Slotkin deferred to post-midterm), prob 7% pre-market -> 7% final, but max win ≤10% of risk fails the Anchor band. Oct 19 print at 61¢ on 701 contracts reversed same day with no sourced news. |
| `KXCBDECISIONCANADA-26OCT-H0` Bank of Canada maintains rate at Oct 28, 2026 meeting? | scheduled-process | NO @ 70¢ | 28% | boc-oct28-decision | yes-side thesis: NO on the hold leg is YES on a hike forecast — it predicts the decision's content, not a dated process step, and it is anti-status-quo (no Anchor mechanism). Sep CPI 3.1% (Oct 19) pushed OIS to ~80% hike vs Kalshi 66-68; specialist book (vol24 42k). Precedent KXFEDDECISION-26JUL-H25: a hawkish repricing into a live meeting still resolved to hold. |
| `KXCBDECISIONCANADA-26OCT-H25` Bank of Canada hikes 25bps at Oct 28, 2026 meeting? | scheduled-process | YES @ 68¢ | 72% | boc-oct28-decision | yes-side thesis: a hike-YES forecasts the content of a scheduled decision, not a process step the resolver pays on; neither already-happened nor process-in-motion applies. NO side (33¢) is below band. Prob 70% pre-market -> 72% final vs 66/68. |
| `KXIPOANTHROPIC-DATE-26DEC01` Anthropic officially announces an IPO before Dec 1, 2026? | corporate-event | NO @ 39¢ | 58% | anthropic-ipo-announce | below band: NO costs 36-39¢. YES (61-64¢) fails the side rule — no public S-1 is on EDGAR as of Oct 20 (EDGAR full-text search, S-1 forms Oct 1-20: zero Anthropic hits; company search: no match), and the mid-November listing target is Bloomberg 'people familiar' sourcing, not an official calendar. Prob 55% pre-market -> 58% final. Disclosure: this audit is written by an Anthropic model; the reject is mechanical (band + EDGAR fact). |
| `KXMAMDANIEO-26OCT24-T0` Mamdani signs above 0 executive orders Oct 18-24, 2026? | confirmation/enact | NO @ 84¢ | 25% | mamdani-eo-count | continuous-number market: weekly executive-order count ladder that slipped past the classifier; also closes in 5 days (inside 7d). Not researched past Stage 0. |
| `KXCOMPANYACTIONMERGER-27-26DEC01` Tesla/SpaceX definitive merger agreement before Dec 1, 2026? | corporate-event | NO @ 93¢ | 5% | tesla-spacex-merger | deep-favorite band: NO costs 92-93¢ (YES 7/8). Stage-0 mechanical; not researched. |
| `KXUSAIRANAGREEMENT-27-26DEC` US agrees to a new Iran nuclear deal before Dec 1, 2026? | scheduled-process | NO @ 94¢ | 4% | us-iran-deal | deep-favorite band: NO costs 93-94¢ (YES 6/7). Precedent KXUSAIRANAGREEMENT-27-26JUN/26JUL: rules require an executed instrument with both prongs; US says no strikes before Nov 3 but Trump says the pause 'might be running out' — a structural NO with no payout left. |
| `KXCAWEALTHTAX-26` California one-time billionaire tax initiative passes? | confirmation/enact | NO @ 77¢ | 28% | ca-wealth-tax | close after expiration: close_time 2027-11-03T04:59Z is a year after expected_expiration_time 2026-11-03T15:00Z (Stage-0 mechanical). |
| `KXIMPEACH-27-FEB01` President impeached before Feb 1, 2027? | legal-action | NO @ 83¢ | 8% | trump-impeach | close after expiration: expected_expiration 2026-11-03 vs close 2027-02-01 (Stage-0 mechanical). Would otherwise be a calendar NO (new House seated Jan 3, 2027). |
| `KXIMPEACH-28-JAN01` President impeached before Jan 1, 2028? | legal-action | NO @ 36¢ | 55% | trump-impeach | close after expiration (exp 2026-11-03 vs close 2028-01-01); also below band (NO 35-36¢) and beyond horizon. |
| `KXIMPEACH-27-JUL01` President impeached before Jul 1, 2027? | legal-action | NO @ 58¢ | 35% | trump-impeach | close after expiration (exp 2026-11-03 vs close 2027-07-01); also below band (NO 57-58¢) and beyond horizon. |
| `KXRETIREMM-26` Mitch McConnell resigns his office early? | departure/ouster | NO @ 99¢ | 1% | mcconnell-resign | deep-favorite band: YES ask ≤5¢ (mechanical cut), NO would cost ≥95¢. |
| `KXUSAIRANAGREEMENT-27-26NOV` US agrees to a new Iran nuclear deal before Nov 1, 2026? | scheduled-process | NO @ 99¢ | 1% | us-iran-deal | deep-favorite band: YES ask ≤5¢ (mechanical cut), NO would cost ≥95¢. |
| `KXIPOANTHROPIC-DATE-26NOV01` Anthropic officially announces an IPO before Nov 1, 2026? | corporate-event | NO @ 100¢ | 1% | anthropic-ipo-announce | deep-favorite band: YES ask ≤5¢ (mechanical cut), NO would cost ≥95¢. |
| `KXCOMPANYACTIONMERGER-27-26NOV01` Tesla/SpaceX definitive merger agreement before Nov 1, 2026? | corporate-event | NO @ 97¢ | 2% | tesla-spacex-merger | deep-favorite band: YES ask ≤5¢ (mechanical cut), NO would cost ≥95¢. |
| `KXHEGSETHOUT-26AUG-26NOV01` Hegseth out as Secretary of Defense before Nov 1, 2026? | departure/ouster | NO @ 99¢ | 1% | hegseth-sod-exit | deep-favorite band: YES ask ≤5¢ (mechanical cut), NO would cost ≥95¢; also cluster runner-up to the Dec 1 rung. |
| `KXCBDECISIONCANADA-26OCT-C25` Bank of Canada cuts 25bps at Oct 28, 2026 meeting? | scheduled-process | NO @ 100¢ | 1% | boc-oct28-decision | deep-favorite band: YES ask ≤5¢ (mechanical cut), NO would cost ≥95¢. |
| `KXALBERTAREFS-26OCT19-Q09` Alberta referendum Question 9 passes? | confirmation/enact | NO @ 100¢ | 40% | alberta-refs | inside spread / Stage-0 mechanical: YES 0/82 (spread 82¢) and close after expiration (exp 2026-11-19 vs close 2027-10-19). Not researched. |
| `KXALBERTAREFS-26OCT19-Q07` Alberta referendum Question 7 passes? | confirmation/enact | NO @ 100¢ | 40% | alberta-refs | inside spread / Stage-0 mechanical: YES 0/99 (spread 99¢) and close after expiration. Not researched. |
| `KXALBERTAREFS-26OCT19-Q08` Alberta referendum Question 8 passes? | confirmation/enact | NO @ 100¢ | 40% | alberta-refs | inside spread / Stage-0 mechanical: YES 0/54 (spread 54¢) and close after expiration. Not researched. |
| `KXDIESELEXPORTBAN-26` US diesel export ban in effect before Nov 3, 2026? | multi-news-other | NO @ 98¢ | 3% | diesel-export-ban | already an active pick ($EXISTING_PICKS); also vol24 448 < 500 and YES ask ≤5¢. |

## Precedent writeback

Two rows written to `/tmp/out/precedents.json` for markets verified settled via the public API this run: `KXHEGSETHOUT-26AUG-26OCT01` (NO — a GOP senator's replacement call plus the Army Secretary's resignation did not move the office) and `KXDEMSCHUMER-27-SEP01` (NO — only an explicit statement by a sitting caucus member counts; House members, candidates and outside groups are out of scope).

## Data notes

- Query API `/schema` read first; tables used: `resolution_precedents` (166 rows), `signals`, `trading_markets`. No `market_snapshots` table exists; price history came from the public candlesticks endpoint (60-minute bars, 14 days).
- Public-API decimal fields (`*_dollars`, `*_fp`) were used throughout; bare-cent fields are null.
- Orderbook depth is read from `orderbook_fp`: NO fills are taken against resting YES bids (NO price = 100 − YES bid).
- Primary sources cited: The Hill (Oct 19), CBS News and Washington Times (Oct 20), PBS/WaPo (Oct 12), TIME (Sep 2), Washington Post (Sep 1), EDGAR full-text search and company search (Oct 20), BNN Bloomberg (Oct 16).
- Disclosure: this audit is produced by an Anthropic model. The Anthropic IPO leg's treatment is mechanical (NO below band; YES blocked by the absence of a public S-1 on EDGAR).

Generated 2026-10-20 · Anchor methodology · 0–3 picks per run is the expected range; 1 pick, 21 logged rejects, $125 of $2,500 allocated.