Kalshi Politics — Mispricing Audit
1. How this was researched
Mode: category-match. trading_events.category = 'Politics' matched 366 active markets across 134 events closing before 2026-09-03. Stage 1 mechanical filters (24h volume ≥ 1,500 contracts, spread ≤ 5¢, not fully priced, not an announcer-noise family, not already a live pick in the feed) cut that to 29 survivors; hand-curation of ladder duplicates and noise families (Truth Social post counts, Getty photo-day counts) left 15 candidates that got full Stage 2 diligence: rules text from the DB mirror, live orderbooks and 14-day candlesticks from the Kalshi public API, and primary-source research (IMF PortWatch ArcGIS feed, Federal Register API, whitehouse.gov, USTR, congress.gov, Punchbowl, wire services).
Per v3.1, every probability below is a forecast of the resolver's verdict, with resolver risks priced as explicit point deductions. Per v3.2, sub-35¢ tails are hard rejects; per v3.3, LOW-conviction leans are logged rejects, not picks. Sizing is conviction-tier + entry-band only — the EV numbers are recorded for calibration and did not set position sizes.
Excluded up front as already-live picks from prior runs: 35 tickers including KXHORMUZWEEKLY-26JUL19-T60, KXEOWEEK-26JUL18-0, KXKASHOUT-26APR-AUG01, KXHEGSETHOUT-26APR-AUG01, KXCRYPTOSTRUCTURE-26JAN-AUG, KXFISAEXTEND-26MAY-26AUG01.
2. Markets reviewed
| Ticker | Market | YES bid/ask | Vol 24h | Verdict |
|---|---|---|---|---|
| KXCRYPTOSTRUCTURE-26JAN-SEP | Crypto market-structure bill law before Sep 1 | 17 / 18 | 3,121 | PICK — BUY NO |
| KXNEWTARIFFS-26JUL-AUG01 | Trump tariff executive action (w/ effective date) in July | 73 / 75 | 17,236 | PICK — BUY YES |
| KXHORMUZPEAK-26JUL19-T20 | Hormuz peak day >20 transits, Jul 13–19 (PortWatch) | 28 / 29 | 1,668 | PICK — BUY NO |
| KXEOWEEK-26JUL18-1 | >1 executive order signed Jul 12–18 | 6 / 7 | 4,028 | REJECT — edge inside slippage |
| KXHORMUZPEAK-26JUL19-T15 | Hormuz peak day >15 transits, Jul 13–19 | 75 / 77 | 1,583 | REJECT — sub-35¢ tail (band ban) |
| KXHORMUZWEEKLY-26JUL19-T50 | >50 Hormuz transits, Jul 13–19 | 50 / 53 | 6,312 | REJECT — coin-flip band |
| KXHORMUZWEEKLY-26JUL26-T50 | >50 Hormuz transits, Jul 20–26 | 29 / 31 | 2,079 | REJECT — thin edge, reopening tail |
| KXHORMUZNORM-26MAR17-B260901 | Hormuz 7-day MA >60 before Sep 1 | 10 / 11 | 38,711 | REJECT — fairly priced |
| KXTRUMPBIBIMEET-26AUG01-AUG01 | Trump–Netanyahu in-person meeting before Aug 1 | 61 / 62 | 7,686 | REJECT — no edge, definitional risk |
| KXACTBLUETOP-26JUL24-JOSS | Ossoff top ActBlue fundraiser Jul 24 10 AM | 76 / 77 | 1,817 | REJECT — sub-35¢ tail (band ban) |
| KXFISAEXTEND-26JUL-26SEP01 | FISA 702 reauth law before Sep 1 | 8 / 9 | 6,011 | REJECT — thin, thesis already in feed |
| KXTRUMPUFC-26AUG-DJT | Trump attends UFC 330 (Aug 15, Philadelphia) | 13 / 14 | 3,775 | REJECT — fairly priced |
| KXTRUMPNUMSTATES-26AUG01-E8 | Exactly 8 distinct states visited in July (VISITAREA) | 23 / 24 | 3,632 | REJECT — resolver opacity |
| KXAPRPOTUS-26JUL24-41.2 | RCP approval 41.1–41.3 on Jul 24 | 34 / 36 | 2,493 | REJECT — resolver inaccessible |
| KXAPRPOTUS-26JUL24-40.9 | RCP approval 40.8–41.0 on Jul 24 | 31 / 32 | 3,023 | REJECT — resolver inaccessible |
3. Picks — detail and thesis
Pick 1 — KXCRYPTOSTRUCTURE-26JAN-SEP · BUY NO @ 0.83 · HIGH CONVICTION
"Will a crypto market structure bill become law before Sep 1, 2026?" — closes 2026-09-01 14:00 UTC.
Thesis. The market prices a 17–18% chance a crypto market-structure bill is signed law by Aug 31. The congressional calendar forecloses it: the House held its last votes Thursday, July 23 and does not return until Monday, Aug 31; the Senate has never passed any version; and the House-passed H.R. 3633 differs from the Senate's merged Banking/Agriculture draft, so one chamber passing is not enough — the texts must converge and be signed inside a window that no longer exists. The recent YES spikes (to 18¢ on Jul 16 and Jul 19) track headlines about a Senate floor vote "before recess" — but a Senate vote is not a law, and that conflation is where the edge comes from.
Evidence.
- House: last votes Thu Jul 23, returns Aug 31 for one week (Punchbowl, Roll Call). Senate: in through Aug 7, returns Sep 14.
- No cloture filed as of Jul 20; the merged draft dropped the officials-crypto-ethics provision and Sens. Murphy, Van Hollen, Merkley formally oppose — the ~7–9 Democratic votes for 60 are not there (CoinDesk Jul 9; TechTimes Jul 15).
- Senate Banking advanced its version May 14 (15–9); Senate Agriculture Jan 29 (12–11); neither has had a floor vote. House passed H.R. 3633 on Jul 17, 2025 — a different text.
- Resolver profile: "becomes law before Sep 1, 2026" — a bill either receives the President's signature by Aug 31 or it doesn't; no discretion, no snapshot risk.
Priced resolver risks: −2 pts (Senate passes this week and the House passes the Senate text same-day on Aug 31 and it's signed before midnight), −1 pt (unanimous-consent passage in a pro-forma session — implausible for a bill with active Democratic opposition), −1 pt (resolver counts some other enacted crypto bill as "market structure"). Evidence-based ~100 → 96.
Tail risk (cleanest single loss path). A surprise ethics-provision deal produces 60 Senate votes this week, House leadership brings the Senate text up by UC or on Aug 31, and Trump signs same-day. Every link in that chain is currently reported as broken.
Liquidity / entry context. YES 17 bid / 18 ask; buying NO at 83: ~252 contracts fillable at the touch, then 77 @ 84 and 1,000 @ 85; a 9,700-contract NO bid wall sits at 82 behind the entry. Vol 24h 3,121; OI 30,486. Price history: YES 9¢ (Jul 14) → 18¢ (Jul 16, merged-draft release) → 12¢ (Jul 17) → 18¢ (Jul 19) → 17/18 now — the market moved toward NO's entry, not away; the edge was not eaten.
Pick 2 — KXNEWTARIFFS-26JUL-AUG01 · BUY YES @ 0.75 · MEDIUM CONVICTION
"Will Trump issue any executive action imposing new or increased tariffs on any country, provided the action specifies an effective date, in July 2026?" — closes 2026-08-01 03:59 UTC.
Thesis. Three partially independent paths to YES in the next 11 days. (1) Trump already signed a Section 301 presidential memorandum on Jul 15 ordering a 25% tariff on all Brazilian goods — USTR's implementing notice sets a Jul 22 effective date, and the memo hit the Federal Register today. (2) The Section 122 global 10% surcharge expires Jul 24 and the administration is reported "ready to replace it with more permanent levies in the coming days" — a presidential replacement action with in-text effective dates. (3) The Section 301 forced-labor investigations hit their statutory completion deadline today, Jul 20 (proposed 12.5% duties on ~45 countries). The market's 75¢ effectively prices only the forward paths; any Brazil-memo credit is upside.
Evidence.
- Presidential Memorandum of Jul 15, 2026 — "The Trade Representative shall impose a tariff of 25 percent on all goods of Brazil" (91 FR 45619, published Jul 20; USTR press release and fact sheet).
- USTR Notice of Action: 25% Section 301 duties on Brazil effective 12:01 a.m. ET Jul 22 (EY Tax News Jul 2026).
- Section 122 surcharge (imposed Feb 20 after SCOTUS struck the IEEPA tariffs) lapses Jul 24; replacement reportedly ready (Yahoo Finance; Transport Topics).
- Series precedent cuts both ways and is priced: KXNEWTARIFFS-26MAY resolved NO (the resolver is strict), KXNEWTARIFFS-26JUN resolved YES and settled two days after the Jun 3 customs EO.
Resolver profile + priced risks. Resolver: Kalshi's reading of "executive action … provided the action specifies an effective date." The June market settled 2 days after a qualifying EO; this market is still trading 5 days after the Brazil memo — strong evidence the resolver has not accepted a memo whose effective date lives in the USTR notice rather than the presidential document. Deductions: −10 pts (Brazil memo never counts on its own), −5 pts (the Section 122 replacement is drafted agency-only or as another dateless memorandum, or slips past Aug 1). Evidence-based ~100 → 85. Note this is also where the remaining edge lives: the market repriced 85¢ → 75¢ today on the same non-settlement inference, but 75¢ underprices the Jul 24 statutory forcing event — this administration will not let its global tariff baseline lapse for a week.
Tail risk (cleanest single loss path). The resolver stays strict on the Brazil memo and the Section 122 replacement arrives as another USTR-implemented memorandum without an in-text effective date (or slips to August). That correlated drafting-style risk is why this is MEDIUM, not HIGH.
Liquidity / entry context. YES 73 bid / 75 ask; ~121 contracts fillable at ≤75, next size at 78. Vol 24h 17,236 (most active candidate); OI 29,280. Price history: 37¢ (Jul 15) → 49 → 54 → 74 → 85 (Jul 19) → 74/75 today. The thesis has partially moved — flagged per Stage 3 — but my 85% vs the 75¢ ask still clears the MEDIUM bar, and today's pullback restored most of the entry.
Pick 3 — KXHORMUZPEAK-26JUL19-T20 · BUY NO @ 0.74 · MEDIUM CONVICTION
"Will the highest single-day number of Hormuz transit calls (IMF PortWatch) be more than 20, Jul 13–19?" — closes 2026-07-21 13:00 UTC (at the moment PortWatch publishes).
Thesis. The measurement week is over; only the resolver's print is pending. For YES, PortWatch must show a day above 20 during a week in which the strait was under active attack throughout: Iran hit three tankers Jul 14, the IMO declared the strait too dangerous to transit Jul 15, the US expanded strikes Jul 17–18, and the IRGC banned all traffic Jul 19. Kpler's independent daily counts for the week — 14 (Jul 14), 15 (Jul 16), 7–8 (Jul 17), a trickle after — never approach 21, and PortWatch's own last five published days (Jul 8–12) were 15, 11, 9, 14, 10. Critically, the known wedge between PortWatch and commercial trackers runs the right way for NO: this program's own July loss on KXMAXSHIPSHORMUZ came from PortWatch printing far below a tracker (28 vs 51). For YES to win here, PortWatch would have to print above Kpler — the opposite of the observed bias.
Evidence.
- PortWatch ArcGIS feed (queried directly today): Hormuz daily transit calls Jul 8–12 = 15, 11, 9, 14, 10; data published only through Jul 12.
- Kpler via CNBC (Jul 17): Jul 14 = 14, Jul 16 = 15, Jul 17 = 7–8 vessels — a multi-week low. TankerMap: 9 tanker transits total Jul 13–19, zero on Jul 19.
- Week-long escalation timeline: tanker strikes Jul 14 (NPR), IMO warning Jul 15 (Bloomberg), IRGC total traffic ban Jul 19 (straits.live) — no reported convoy-surge day.
- PortWatch cadence: updates Tuesdays 09:00 ET (~3-day lag); the print covering Jul 13–18/19 lands Jul 21, exactly at market close.
Resolver profile + priced risks. Resolver: IMF PortWatch daily chokepoint series, snapshot mechanism, threshold "above 20" — unambiguous numerically, but the input data is unpublished and revisable (documented Feb 2026 Hormuz boundary revision; AIS backfills). Deductions: −7 pts (publication/revision surprise on the unpublished days, including AIS-dark transits being backfilled), −3 pts (an escorted-convoy spike day — e.g. Jul 13, before the mid-week attacks — that Kpler undercounted). Evidence-based ~95 → 85. Deductions total 10 pts, under the 15-pt Stage 2.5 screen.
Tail risk (cleanest single loss path). A Jul 13 print of 21+ — the one day of the window with no third-party estimate, following Jul 11's 14 and Jul 12's 10, and before the Jul 14 attacks. It would take a >2× day-over-day jump amid an active shooting war.
Liquidity / entry context. YES 28 bid / 29 ask; buying NO: ~8 contracts @ 72, then 500 @ 74 — full size marketable now at ≤74 (avg ≈ 73.9¢, <3¢ from touch). Vol 24h 1,668; OI 5,360. Price history: YES 42¢ (Jul 14) → 41 (Jul 17–18) → 33 (Jul 19) → 28/29 now. The market has been drifting toward NO all week — part of the edge is eaten (it was +25¢ on Jul 17), but +11¢ remains against a resolution <24 hours away. Order must be placed before the Jul 21 13:00 UTC close.
4. Recommended $2,500 portfolio
| # | Ticker | Action | Limit | Contracts | Cost | Max payout | EV ¢/ct | EV % | Conviction | Band |
|---|---|---|---|---|---|---|---|---|---|---|
| 1 | KXCRYPTOSTRUCTURE-26JAN-SEP | BUY NO | 83¢ | 440 | $365.20 | $440.00 | +13 | 15.7% | HIGH | 60–90 |
| 2 | KXNEWTARIFFS-26JUL-AUG01 | BUY YES | 75¢ | 233 | $174.75 | $233.00 | +10 | 13.3% | MEDIUM | 60–90 |
| 3 | KXHORMUZPEAK-26JUL19-T20 | BUY NO | 74¢ | 236 | $174.64 | $236.00 | +11 | 14.9% | MEDIUM | 60–90 |
| Total deployed | 909 | $714.59 | $909.00 | blended EV ≈ +$106.5 (+14.9% on cost) | ||||||
| Cash held: $1,785.41 (71.4%) — cash is a position; the category was efficient this week and nothing below MEDIUM deploys under v3.3. Per v3 policy, EV did not drive any contract count — sizes come from the conviction caps (HIGH ≤ $375, MEDIUM ≤ $175) and entry bands only. | ||||||||||
Cluster exposure (cap: 15% of capital = $375 per cluster)
| Cluster | Picks | Cost | % of capital | Cap check |
|---|---|---|---|---|
| congress-pre-recess-calendar | #1 | $365.20 | 14.6% | ✓ under 15% |
| trump-tariff-actions | #2 | $174.75 | 7.0% | ✓ |
| hormuz-portwatch-jul13-19 | #3 | $174.64 | 7.0% | ✓ |
The FISA-702 NO candidate was rejected partly because it would have stacked the congress-pre-recess-calendar cluster past its cap; the second Hormuz-week candidate was rejected partly to avoid doubling the Hormuz thesis. No cluster can flip the run's sign: the largest is 14.6% of capital.
Conviction exposure
| Tier | Picks | Cost deployed | % of capital | Per-pick cap |
|---|---|---|---|---|
| HIGH | 1 | $365.20 | 14.6% | $375 ✓ |
| MEDIUM | 2 | $349.39 | 14.0% | $175 each ✓ |
| LOW | 0 | $0 | 0% | never deploys (v3.3) |
Risk profile
- Worst case: all three lose → −$714.59 (−28.6% of capital). The three theses are driven by different mechanisms (congressional calendar, tariff drafting style, a shipping-data print), so a single event cannot produce it.
- Best case: all three win → +$194.41 (+7.8% of capital, +27.2% on deployed cost).
- Most likely outcome: all three win — joint probability ≈ 0.96 × 0.85 × 0.85 ≈ 69% under my models. Probability-weighted P&L ≈ +$106.
- Timing: Pick 3 resolves within ~48h (Jul 21 print); Pick 2 by Aug 1; Pick 1 by Sep 1. Capital recycles quickly if the first two hit.
- Correlation note: Pick 2 (tariff action) is mildly anti-correlated with macro chaos that would hurt nothing else in the book; Pick 3's loss mode (PortWatch revision surprise) does not touch Picks 1–2.
Execution notes
- Pick 1 (crypto NO @ 83): ~252 contracts marketable at the touch; rest the remainder at 83 — you queue ahead of a 9.7k-contract wall at 82, so fills come from YES buyers crossing on floor-vote headlines. A Senate floor-vote headline spiking YES to 20–25 is an add opportunity, not an invalidation. Invalidated only if: a reported ethics-provision deal produces 60 named Senate votes before Jul 24 and House leadership commits to taking up the Senate text on Aug 31 — then exit.
- Pick 2 (tariffs YES @ 75): ~121 at the touch; rest the remainder at 75 in a 17k-vol/day book — do not chase above 75; at 78+ the edge halves. Watchlist trigger: Kalshi settling the market early on the Brazil memo (instant win) or a Jul 22–24 Section 122 replacement proclamation (near-instant win). Invalidated if Jul 24 passes with the surcharge allowed to lapse and no replacement signed — exit into the high-50s rather than ride the last week.
- Pick 3 (Hormuz NO @ 74): marketable now (8 @ 72 + 500 @ 74); must fill before the Jul 21 13:00 UTC close — this is a <24h hold into the print. No exit path after close; size is set accepting binary resolution.
- Reserve: $1,785 (71%) held back, well above the 10% minimum — deliberate in an efficient week.
5. What I rejected and why
All 12 rejects below are machine-logged in picks.json and shadow-tracked to settlement — if the rejects outperform the picks, the screens add nothing, and that finding will surface. Three cuts came from the Stage-3.5 entry-band screen; they are the discipline the resolved v1/v2 record paid for.
- KXHORMUZPEAK-26JUL19-T15 — banned band. I'd buy NO at 24¢ (market implies 76% YES; my model ~43% — no tracker shows a day ≥16). That's a stated +33¢ edge, which is exactly the "asymmetric lottery" pattern v3.2 killed after tails went ~0-for. Logged for the shadow track; if these keep winning, the record will argue for reinstatement.
- KXACTBLUETOP-26JUL24-JOSS — banned band + resolver opacity. NO at 24¢ was tempting (El-Sayed reportedly holds the weekly ActBlue top slot in his Aug 4 primary surge; Talarico raised $30M last quarter), but sub-35¢ is a hard reject and the resolution source — the "Top fundraisers from this week" widget on actblue.com/directory, leftmost name wins — is JS-rendered and could not be independently read.
- KXHORMUZWEEKLY-26JUL19-T50 — coin-flip band. 50/53¢ with the entire resolution dataset unpublished; my weekly-total estimate (~45–70 PortWatch-adjusted) straddles the strike. 35–60¢ requires HIGH conviction and ≥15¢ of edge; this has neither.
- KXEOWEEK-26JUL18-1 — slippage screen. NO at 94¢ is a near-lock (zero EOs on whitehouse.gov or the Federal Register all month; window closed Jul 18) but +4¢ of edge dies on a book with ~65 contracts near the touch. Chasing a few cents in a >90¢ deep favorite is explicitly what v3 says not to do.
- KXHORMUZWEEKLY-26JUL26-T50 — thin +6¢ on a forward forecast with a fat tail: ~450 vessels are anchored at the strait, so any mid-week ceasefire flushes the queue past 50 fast. Also stacks the Hormuz cluster.
- KXTRUMPBIBIMEET-26AUG01-AUG01 — no edge (62¢ vs my ~60%) and live definitional risk: Netanyahu is expected in DC ~Jul 28 for Sen. Graham's funeral, one trip was already cancelled, the White House won't schedule a bilateral, and whether funeral co-attendance counts as "meet in person" is resolver discretion I can't price under 15 pts.
- KXHORMUZNORM-26MAR17-B260901 — fairly priced at 10/11¢: 7-day MA is ~10 vs 60 needed, but the queued-vessel flush after a ceasefire could briefly spike the average; my ~11% ≈ the market.
- KXFISAEXTEND-26JUL-26SEP01 — the NO is right (702 lapsed Jun 12; no vote scheduled; House gone) but it's +6¢ at 92¢, the same thesis already sits in users' feeds via the Aug 1 leg, and it would stack the congressional-calendar cluster past its cap.
- KXTRUMPUFC-26AUG-DJT — fairly priced at 13/14¢: no announcement, no Philadelphia precedent (skipped MSG Nov 2025; attends Miami cards), wartime schedule cuts both ways.
- KXTRUMPNUMSTATES-26AUG01-E8 — Stage 2.5 screen: the VISITAREA tracker state could not be verified, so the resolver is unknowable to me.
- KXAPRPOTUS-26JUL24-41.2 and -40.9 — the resolution source (RealClearPolitics average) returned HTTP 403 on every fetch; without the live average there is no edge against traders who watch it hourly.
6. Sources
- IMF PortWatch — daily chokepoint transit calls, Strait of Hormuz (queried directly): portwatch.imf.org FAQs, ArcGIS FeatureServer Daily_Chokepoints_Data (chokepoint6)
- Federal Register API — executive-order documents; Brazil memo: 91 FR 45619 (Jul 20, 2026)
- USTR — Section 301 action on Brazil; EY Tax News — Notice of Action, effective Jul 22
- Section 122 expiry — Yahoo Finance; Transport Topics; forced-labor 301 — Dorsey client alert
- White House — Presidential Actions (zero EOs posted Jul 1–19, 2026)
- Congress calendar — Punchbowl News; senate.gov 2026 schedule; Roll Call House calendar
- CLARITY Act status — CoinDesk Jul 9; TechTimes Jul 15; H.R. 3633
- FISA 702 — Brennan Center 2026 tracker; Axios Jun 11
- Hormuz crisis — CNBC Jul 17 (Kpler daily counts); CNBC Jul 13; Bloomberg Jul 15 (IMO); NPR Jul 14; straits.live; TankerMap; Wikipedia crisis timeline
- Netanyahu trip — Axios Jul 16; Times of Israel; Bloomberg Jul 15
- ActBlue race — Texas Tribune (Talarico $30M); The Hill; Kalshi series KXACTBLUETOP
- UFC 330 — Wikipedia; Fox News (UFC 327)
- Kalshi public API — live orderbooks, candlesticks, settled-series precedent (KXNEWTARIFFS-26MAY = NO, -26JUN = YES, settled Jun 5)
Data: Kalshi mirror DB (read-only, queried 2026-07-20), Kalshi public trade API (live orderbooks/candlesticks), IMF PortWatch ArcGIS, Federal Register API, whitehouse.gov, and the cited press. Prices quoted in cents; entries as of ~2026-07-20 late UTC.
Disclaimer: All probabilities are subjective estimates by the audit model, not measurements. Event contracts resolve to $0 or $1; every position here can lose 100% of its cost. This report is research for a tracked public record, not investment advice. Rejected candidates are logged and shadow-tracked so the selection screens themselves stay testable.