# Kalshi Politics — Mispricing Audit (Sentinel, managed) — 2026-07-01

Date: 2026-07-01
Source: https://kalshi-audits.pages.dev/2026-07-01-kalshi-politics-mispricing-audit-sentinel

---
# Kalshi Politics — Mispricing Audit (Sentinel, managed)

Report date 2026-07-01 · data pulled 2026-07-01 → 2026-07-02 UTC · v3 methodology (conviction-weighted, EV-agnostic sizing) · $1,000 capital · 45-day horizon · picks count toward the public track record and are cleared for auto-copy

## 1. How this was researched

**Mode: category-match.** `trading_events.category = 'Politics'` matched directly — no theme-keyword fallback was needed. The mirror returned **290 active Politics markets** closing between 2026-07-01 and 2026-08-15. Latest snapshots (5-minute data) were joined per ticker and the Stage-1 mechanical screen applied: 24h volume ≥ 1,500, bid-ask spread ≤ 5¢, not fully priced (YES bid < 95¢, YES ask > 5¢), no say/mention noise markets. **23 markets survived.** Ten of those are already-live picks from prior runs (both Hormuz-NORM legs, Trump World Cup attendance ×2, Trump UFC 329, crypto market-structure bill, Kash Patel, hedge-fund housing bill, Truth Social B129, Trump-attends-final) and were excluded up front to avoid feed duplicates, leaving **13 candidates** for full Stage-2 diligence.

Every candidate got: rules text from the mirror, live top-of-book from the Kalshi public API, snapshot price history, and primary-source news work — the IMF PortWatch chokepoint API (daily Hormuz transit counts), the White House presidential-actions and executive-orders listings, the NPR/WaPo record on the Alito retraction, and Truth Social archive/tracker reporting. RealClearPolitics — the resolution source for the approval-bucket markets — returned HTTP 403 on every route tried (direct, alternate UA, epolls mirror, WebFetch), which itself became a reject reason: an edge you cannot independently verify is not an edge.

## 2. Markets reviewed

| Ticker | Market | YES bid/ask | Vol 24h | Outcome of review |
|---|---|---|---|---|
| KXALITOANNOUNCERETIRE-26JUN-JUL06 | Alito announces departure before Jul 6 | 5/6 | 13,272 | PICK — BUY NO @ 95¢ (HIGH) |
| KXALITOANNOUNCERETIRE-26JUN-AUG01 | Alito announces departure before Aug 1 | 14/15 | 11,059 | PICK — BUY NO @ 86¢ (MEDIUM) |
| KXEOWEEK-26JUL04-2 | >2 executive orders Jun 28–Jul 4 | 13/14 | 2,138 | PICK — BUY NO @ 88¢ (MEDIUM) |
| KXHORMUZWEEKLY-26JUL05-T200 | >200 Hormuz transits Jun 29–Jul 5 | 73/75 | 2,077 | PICK — BUY NO @ 27¢ (LOW, lottery) |
| KXHORMUZWEEKLY-26JUL05-T150 | >150 Hormuz transits Jun 29–Jul 5 | 93/94 | 5,857 | REJECT — priced above model |
| KXHORMUZWEEKLY-26JUL05-T175 | >175 Hormuz transits Jun 29–Jul 5 | 85/87 | 4,929 | REJECT — tail band, lottery slot used |
| KXTRUMPENDORSEMENTS-26JUL04-A10 | Trump endorses ≥10 people this week | 70/71 | 3,804 | REJECT — book too thin to fill |
| KXTRUTHSOCIAL-26JUL04-B210 | 200–220 Truth Social posts this week | 6/7 | 3,958 | REJECT — no verifiable pace read |
| KXTRUTHSOCIAL-26JUL04-B189 | 180–199 Truth Social posts this week | 7/8 | 3,266 | REJECT — tail, pace unverifiable |
| KXTRUTHSOCIAL-26JUL04-B109 | 100–119 Truth Social posts this week | 7/8 | 3,204 | REJECT — ≤1¢ edge deep favorite |
| KXAPRPOTUS-26JUL03-40.3 | RCP approval 40.2–40.4 on Jul 3 | 24/25 | 2,696 | REJECT — source unfetchable |
| KXAPRPOTUS-26JUL03-40.6 | RCP approval 40.5–40.7 on Jul 3 | 55/56 | 1,952 | REJECT — coin-flip band screen |
| KXAPRPOTUS-26JUL03-40.9 | RCP approval 40.8–41.0 on Jul 3 | 13/15 | 2,110 | REJECT — tail, source unfetchable |

Ten further Stage-1 survivors excluded as already-published live picks (would duplicate the feed): KXHORMUZNORM-26MAR17-B260715/-B260801, KXTRUMPATTEND, KXTRUMPWORLDCUP-26-DJT-1/-2, KXTRUMPUFC-26JUL-DJT, KXCRYPTOSTRUCTURE-26JAN-AUG, KXKASHOUT-26APR-AUG01, KXHFHOUSING-27-26JUL07, KXTRUTHSOCIAL-26JUL04-B129.

## 3. Picks — detail and thesis

### Pick 1 — KXALITOANNOUNCERETIRE-26JUN-JUL06 · BUY NO @ 0.95 · HIGH CONVICTION

_Will Samuel Alito announce his departure as Supreme Court justice before Jul 6, 2026?_

Current YES 5¢ / 6¢
My probability (NO) 98%
Market implied (NO) 95%
Entry band >90¢ deep favorite
Cluster alito-retirement
Size 84 × 95¢ = $79.80

**Thesis.** The event this market priced has already failed to happen. The Supreme Court's term-end opinion announcements concluded Jun 30 — the traditional retirement-announcement moment — with no retirement. The 36¢ intraday spike on Jun 30 was **NPR's false report**: Nina Totenberg misheard "retirement announcements" in the courtroom, NPR published, the Court **flatly denied it**, and NPR retracted within hours, with Totenberg calling it "the worst professional mistake" of her 50-year career. For YES to win, Alito would have to announce within days of his own Court's public denial, during recess.

- NPR retraction and Court denial: Washington Post, NPR Public Editor, Deadline, Fox News (all Jun 30, 2026 — links in Sources).
- Rules note: title says "announce," stored rules say "retires" — the gap is immaterial for the NO side, since every YES path requires an announcement that has not occurred.

**Tail risk (cleanest way to lose):** a genuine surprise announcement before Jul 5 — post-retraction conservative-media chatter claims "the real announcement may be coming." That chatter is why 98%, not 99+.

**Numbers.** Edge +3¢/contract, EV +3.2% over ~4 days (recorded, not used for sizing). Max profit $4.20.

**Liquidity.** Buy-NO ladder 334 @ 95¢, 639 @ 96¢; 24h vol 13,272; OI 13,192. 84 contracts fill at the inside — zero slippage.

**Price history.** Jun 30: YES ranged 3–36¢ on the NPR spike and full retraction round-trip; Jul 1: 3–13¢; now settled 5/6¢.

### Pick 2 — KXALITOANNOUNCERETIRE-26JUN-AUG01 · BUY NO @ 0.86 · MEDIUM CONVICTION

_Will Samuel Alito announce his departure as Supreme Court justice before Aug 1, 2026?_

Current YES 14¢ / 15¢
My probability (NO) 92%
Market implied (NO) 86%
Entry band 60–90¢ favorite
Cluster alito-retirement
Size 81 × 86¢ = $69.66

**Thesis.** The market keeps 14–15% on a July announcement because the NPR fiasco left speculation smoldering. But the base rate is against it: modern justices announce at term end (Kennedy, Jun 27 2018) or mid-winter (Breyer); a July-recess announcement has essentially one modern precedent (O'Connor, Jul 1 2005). The term-end window just closed with a flat denial. Residual probability for the next 30 days is ~8%, not 14–15%.

- Term ended Jun 30 with no announcement; Court denied the NPR report the same day (Washington Post, NPR Public Editor).
- Speculation pieces (The Nation; conservative-insider items) argue a strategic summer retirement would let Trump fill the seat pre-midterms — this is the live YES case and it is priced, just too richly.

**Tail risk:** the insider chatter is directionally right and Alito announces in July — health events or a strategic decision are unhedgeable here. This is a real, live way to be wrong, which is what caps conviction at MEDIUM despite the favorable base rate.

**Numbers.** Edge +6¢/contract, EV +7.0% over 30 days (recorded, not used for sizing). Max profit $11.34.

**Liquidity.** Buy-NO ladder 165 @ 86¢, 270 @ 87¢; 24h vol 11,059; OI 21,178. 81 contracts fill inside two levels.

**Price history.** Jun 30: YES spiked 13→68¢ on the false report and round-tripped; Jul 1: 9–35¢; now 14/15¢. Entry is at the post-correction level, not chasing the move.

### Pick 3 — KXEOWEEK-26JUL04-2 · BUY NO @ 0.88 limit · MEDIUM CONVICTION

_Will the President sign more than 2 executive orders between Jun 28 and Jul 4, 2026?_

Current YES 13¢ / 14¢
My probability (NO) 91%
Market implied (NO) 86–87%
Entry band 60–90¢ favorite
Cluster eo-week-jul4
Size 47 × ≤88¢ = $41.36 max

**Thesis.** As of Jul 2, the White House executive-orders listing shows the most recent EO signed **Jun 25** (EO 14414) — zero EOs in the Jun 28–Jul 4 window with ~3 days left, so YES needs a 3+ EO burst by Saturday. June produced 6 EOs across five signing dates and **no 3-day span of June contained 3 EOs**. A trap was checked and cleared: the White House posted a video titled "President Trump Signs Executive Orders, Jun. 29, 2026," but the presidential-actions ledger shows that event produced a **proclamation and a memorandum, not EOs** — the video title is loose labeling, and the EO-specific page confirms nothing signed after Jun 25.

- White House EO listing (fetched Jul 2): latest is EO 14414, Jun 25; prior batch Jun 22 (×2), Jun 2–3 (×3).
- Jun 29 presidential actions: "Declaration of Emergency…Temporary Duty Free" (proclamation) + "Freedom to Fix" (memorandum).

**Tail risk:** a July-4-week batch signing ceremony, or Jun 29–30 EOs that exist but are unposted (posting lag is normally ≤2 days, and 3 days have passed — but the market's late resolution date, Jul 18, means the Federal Register's complete record decides). The market already crashed 40¢→13¢ on this thesis since Jun 27, so most of the edge is eaten — that is why size is roughly half the MEDIUM cap.

**Numbers.** Edge +3¢/contract at avg fill ~88¢, EV +3.4% (recorded, not used for sizing). Max profit ~$5.60.

**Liquidity.** Buy-NO ladder is thin: 32 @ 87¢, 15 @ 88¢, 32 @ 89¢. Size capped at 47 contracts = the two inside levels; limit 88¢, do not chase to 89¢. 24h vol 2,138; OI 1,449.

**Price history.** YES: 34–40¢ (Jun 27–28) → 31–60¢ intraday Jun 29 → 12–14¢ by Jun 30 as the week opened EO-free. Now 13/14¢.

### Pick 4 — KXHORMUZWEEKLY-26JUL05-T200 · BUY NO @ 0.27 · LOW CONVICTION · explicit asymmetric lottery, 2% cap

_Will there be more than 200 transit calls through the Strait of Hormuz Jun 29 – Jul 5, 2026 (IMF PortWatch)?_

Current YES 73¢ / 75¢
My probability (NO) 52%
Market implied (NO) ~26%
Entry band <35¢ tail — lottery slot
Cluster hormuz-traffic-recovery
Size 74 × 27¢ = $19.98 (2.0% of capital)

**Thesis.** This is the run's single permitted sub-35¢ lottery, and it is a fade of a 48-hour momentum move. YES needs the week to average **>28.6 transits/day**. PortWatch primary data (pulled directly from the IMF ArcGIS API, twice, Jul 1 and Jul 2): traffic collapsed to 4–15/day mid-June, spiked to **52 and 51 on Jun 24–25** as the strait reopened and the backlog cleared, then decayed **39 → 30 → 27** (Jun 26–28, the latest published days). Zero in-window days (Jun 29+) are published yet — the market's repricing of this leg from 35¢ to 73¢ in a day is pure extrapolation of re-acceleration that the latest tape does not show. A flat-at-27 path totals ~189 and resolves NO.

- PortWatch dailies Jun 20–28: 26, 8, 14, 15, 52, 51, 39, 30, 27 (IMF ArcGIS Daily_Chokepoints_Data, primary source).
- Rules confirmed: sum of daily counts Jun 29–Jul 5; resolves no earlier than Tue Jul 7, 9am, on complete data.

**Tail risk:** normalization continues instead of decaying — pre-crisis Hormuz traffic ran ~70–80/day, so any confidence-driven climb back above ~30/day clears 200 comfortably. That is the market's view, it is entirely plausible, and it is why this is a 52/48 LOW-conviction lottery and not a position.

**Numbers.** Edge +25¢/contract, EV +92.6% (recorded — and pointedly **not** used for sizing; the v3 record shows stated-EV is anti-predictive, so the highest-EV pick in this book gets the smallest allocation). Max profit $54.02.

**Liquidity.** Buy-NO ladder 854 @ 27¢, 1,000 @ 28¢; 24h vol 2,077; OI 1,415. Fills at the inside.

**Price history.** YES: 33–48¢ on Jun 30 → spiked as high as 87¢ intraday Jul 1 → 73/75¢ now.

## 4. Recommended $1,000 portfolio

| # | Ticker | Action | Limit | Contracts | Cost | Conviction / band | Max profit | EV ¢/ct* | EV %* |
|---|---|---|---|---|---|---|---|---|---|
| 1 | KXALITOANNOUNCERETIRE-26JUN-JUL06 | BUY NO | 95¢ | 84 | $79.80 | HIGH · >90¢ | $4.20 | +3 | +3.2% |
| 2 | KXALITOANNOUNCERETIRE-26JUN-AUG01 | BUY NO | 86¢ | 81 | $69.66 | MEDIUM · 60–90¢ | $11.34 | +6 | +7.0% |
| 3 | KXEOWEEK-26JUL04-2 | BUY NO | 88¢ | 47 | $41.36 | MEDIUM · 60–90¢ | $5.64 | +3 | +3.4% |
| 4 | KXHORMUZWEEKLY-26JUL05-T200 | BUY NO | 27¢ | 74 | $19.98 | LOW · <35¢ lottery | $54.02 | +25 | +92.6% |
| Total deployed | $210.80 | 21.1% of capital | ≈+$27.3 | +13.0% blended |  |  |  |  |  |
| Cash held | $789.20 | 78.9% — cash is a position; this was an efficient slate |  |  |  |  |  |  |  |

*EV columns are recorded for calibration only. Under v3 they did **not** set contract counts — sizing is conviction tier × entry band. The highest-EV line (Pick 4) is deliberately the smallest allocation.

### Cluster exposure (cap: 15% of capital = $150 per cluster)

| Cluster | Picks | Cost | % of capital | Cap headroom |
|---|---|---|---|---|
| alito-retirement | 1, 2 | $149.46 | 14.9% | $0.54 — at cap |
| eo-week-jul4 | 3 | $41.36 | 4.1% | $108.64 |
| hormuz-traffic-recovery | 4 | $19.98 | 2.0% | $130.02 |

Both Alito legs lose on the same event (an announcement before Jul 5); they are counted as one cluster and sit exactly at the 15% cap — the correlated-double-bet failure mode from v1 is bounded by construction. The Hormuz lottery is directionally consistent with already-published Hormuz-NORM NO picks from prior runs (both bet against fast normalization); within this run it carries only the 2% lottery cost.

### Conviction exposure

| Tier | Picks | Cost | % of capital | Per-pick cap |
|---|---|---|---|---|
| HIGH | 1 | $79.80 | 8.0% | $150 |
| MEDIUM | 2 | $111.02 | 11.1% | $70 |
| LOW | 1 | $19.98 | 2.0% | $30 (lottery cap $20 binds) |

### Risk profile

- **Worst case** (Alito announces before Jul 5 + EO batch + Hormuz ≤200… note picks 1–2 lose together): −$210.80 (−21.1% of capital). Requires three independent theses all failing.
- **Realistic bad case** (Alito announces in July after Jul 5: pick 1 wins, pick 2 loses; others win): −$59.82 net.
- **Best case** (all four win): +$75.20 (+7.5% of capital).
- **Most likely outcome** (picks 1–3 win, lottery is a coin flip): +$21.2 if the lottery wins, −$18.8 if it loses; probability-weighted ≈ +$27 expected.
- **Concentration:** the single biggest thesis is "Alito does not announce" at 14.9%; no single event can flip the run's sign relative to the 78.9% cash buffer.

### Execution notes

- All entries are **limit orders** at the stated prices; picks 1, 2, 4 fill entirely at the inside quote. Pick 3's book is thin (32+15 at the two inside levels) — post the 88¢ limit and accept partial fills; do not lift 89¢.
- **Invalidators** — exit or stand down if: any credible report of an actual Alito announcement (kills picks 1–2 — exit at market); an EO posted to whitehouse.gov dated Jun 28–30 (pick 3's cushion drops to needing just 1–2 more — exit above 50¢ YES); PortWatch publishes Jun 29–30 dailies ≥33 (pick 4's flat-path math breaks — the $20 is forfeit, no averaging down).
- **Watchlist triggers for the cash reserve:** KXTRUMPENDORSEMENTS-26JUL04-A10 YES becomes attractive if the book deepens near 70¢; KXHORMUZWEEKLY-26JUL05-T175 NO becomes the better lottery if PortWatch prints two sub-25 days; the Jul-10 approval series is playable next week only if an independent RCP read becomes fetchable.
- No hedges — position sizes are small enough that hedging costs more than it saves.

## 5. What I rejected and why

Nine Stage-2 survivors were rejected. All are machine-logged in `picks.json` and shadow-tracked to settlement — if the rejects outperform the picks, the selection screen adds nothing, and that finding is the point of logging them. Five of the nine were cut wholly or partly by the **Stage-3.5 entry-band screen** (coin-flips and tails), which is the discipline the resolved v1/v2 record demanded.

| Ticker | Would-be side | Price | My prob | Reject reason |
|---|---|---|---|---|
| KXHORMUZWEEKLY-26JUL05-T150 | YES | 94¢ | 90% | Deep favorite priced _above_ my model — negative edge. The whole Hormuz ladder is rich against a decay read of the same public data. |
| KXHORMUZWEEKLY-26JUL05-T175 | NO | 15¢ | 30% | Entry-band screen: sub-35¢ tail. One lottery per run; T200 NO is the better-priced expression of the identical thesis. |
| KXTRUMPENDORSEMENTS-26JUL04-A10 | YES | 71¢ | — | 25 contracts at the inside then a 16¢ gap — unfillable without ≥3¢ slippage; in-week endorsement count not independently countable. |
| KXTRUTHSOCIAL-26JUL04-B210 | NO | 94¢ | 94% | ≤1¢ edge; the archive's pagination is broken so the in-week pace can't be counted, and Trump's 2026 average (19/day) vs May mania (27/day) brackets both outcomes. |
| KXTRUTHSOCIAL-26JUL04-B189 | YES | 8¢ | 12% | Entry-band screen: tail. May's 27/day pace lands exactly here, which is tempting — but unverifiable, and prior runs already published three picks in this bucket family. |
| KXTRUTHSOCIAL-26JUL04-B109 | NO | 93¢ | 94% | Deep favorite, ≤1¢ edge, same verification problem. |
| KXAPRPOTUS-26JUL03-40.3 | YES | 25¢ | — | Entry-band screen: tail — and the resolution source (RealClearPolitics) 403s from every route, so no independent read exists. |
| KXAPRPOTUS-26JUL03-40.6 | YES | 56¢ | — | Entry-band screen: 35–60¢ coin-flip requires HIGH conviction + ≥15¢ edge. A 0.3-point bucket exposed to single poll drops over 48h cannot reach that bar. This exact shape — mid-priced data buckets — is where the v1/v2 book bled. |
| KXAPRPOTUS-26JUL03-40.9 | YES | 15¢ | — | Entry-band screen: tail; source unverifiable. |

## 6. Sources

- [IMF PortWatch — Daily Chokepoints Data (ArcGIS API)](https://services9.arcgis.com/weJ1QsnbMYJlCHdG/arcgis/rest/services/Daily_Chokepoints_Data/FeatureServer) — Hormuz daily transit calls, pulled Jul 1 and Jul 2, 2026
- [Washington Post — NPR retracts story reporting Justice Alito is retiring](https://www.washingtonpost.com/business/2026/06/30/npr-retracts-story-reporting-justice-alito-is-retiring-supreme-court/)
- [NPR Public Editor — NPR retracts story about Alito retirement](https://www.npr.org/sections/npr-public-editor/2026/06/30/g-s1-131107/npr-retracts-story-about-alito-retirement)
- [Fox News — how a misheard announcement led to the false Alito report](https://www.foxnews.com/media/npr-reveals-how-misheard-announcement-led-falsely-claiming-justice-alito-retiring)
- [Deadline — NPR retracts Alito retirement report](https://deadline.com/2026/06/npr-retracts-samiel-alito-retirement-report-1236971144/)
- [The Nation — on a strategic Alito retirement (YES-case color)](https://www.thenation.com/article/politics/samuel-alito-supreme-court-retirements/)
- [White House — Executive Orders listing](https://www.whitehouse.gov/presidential-actions/executive-orders/) (fetched Jul 2: latest EO 14414, Jun 25)
- [White House — Presidential Actions](https://www.whitehouse.gov/presidential-actions/) (Jun 29 items: proclamation + memorandum)
- [White House — "Signs Executive Orders, Jun. 29" video page](https://www.whitehouse.gov/videos/president-trump-signs-executive-orders-jun-29-2026/) (the checked-and-cleared trap)
- [Mediaite — Trump averaging 19 Truth Social posts/day in 2026](https://www.mediaite.com/politics/trump-has-averaged-a-whopping-19-truth-social-posts-a-day-in-2026-report/)
- [Daily Beast — May 2026 posting record (~27/day)](https://www.thedailybeast.com/trump-sets-new-posting-record-on-truth-social-as-mania-spirals-out-of-control/)
- [Trump's Truth archive](https://trumpstruth.org/) and [Roll Call Factba.se](https://rollcall.com/factbase-twitter/) — post-pace verification attempts
- [RealClearPolling — Trump job approval](https://www.realclearpolling.com/polls/approval/donald-trump/approval-rating) (resolution source; returned 403 on every fetch route — documented as reject basis)
- [Kalshi public API](https://api.elections.kalshi.com/trade-api/v2) — live quotes, orderbooks, rules (Jul 1–2, 2026)

**Data sources:** Kalshi market mirror (read-only SQL over 5-minute snapshots), Kalshi public trade API, IMF PortWatch, White House, Federal Register (blocked; White House listing used), press as linked. Prices captured Jul 1–2, 2026 UTC and move continuously.

**Disclaimer:** all probabilities are subjective estimates by an automated research process; every contract can resolve to zero and the entire deployed amount can be lost. EV figures are recorded for calibration and did not determine position sizes. This report is not financial advice. v3 methodology: conviction-weighted sizing, entry-band screening, 15% cluster caps, EV-agnostic allocation. Rejects are machine-logged and shadow-tracked to settlement.