Kalshi Politics — Mispricing Audit

Report date 2026-06-03 · Category Politics · Horizon 45 days · Capital $1,000

Screened 26 active politics markets closing by 2026-07-18; ran full diligence on 14 and adversarially stress-tested each. Eleven died on contact with the verifier — the edge was already arbitraged into the price, the resolution rules hid a trap, or the thesis rested on the wrong settlement source. Two picks survive as deployable, plus a watchlist item that a live-orderbook check disqualified at market. Net deployed $859.60 of $1,000 (≈14% held in reserve), blended EV +16.1%, dollar edge ≈ +$139.

1 · How this was researched

Universe. Pulled every active market whose event category = 'politics' closing within 45 days from the live Kalshi mirror (join trading_markets → trading_events), with each market's latest market_snapshots row for price/volume.
Stage-1 screen. Cut anything with volume24h < 1500, bid–ask spread > 5¢, or already fully priced (yes bid ≥ 95¢ / yes ask ≤ 5¢). 26 candidates passed.
Stage-2 diligence (parallel agents). For each of 14 substantive candidates: pulled the exact rules text, a 14-day intraday price history, the live depth-of-book from the Kalshi public API, and primary-source news (congress.gov context, wire reports, official statements). Each produced a true-probability estimate with explicit edge math.
Stage-3 adversarial verify. A second, skeptical agent independently re-pulled the rules and the freshest news for every "keep," tasked with refuting the edge. It killed 11 of 14 — most often because the catalyst had already repriced the market in the prior 48 hours, or because a definitional trap could flip resolution.
Live-book reality check. Before sizing, I re-hit the orderbook for each survivor. This is what disqualified KXJOINTRUMPADMIN-…-KPAX: the snapshot read looked like a clean 4¢ edge, but the resting size at NO 92–94¢ is the bid side; the only way to buy NO is to lift the 97.5¢ ask, leaving ≈0.5¢ — not tradeable.

Subjective probabilities throughout. The aim is a defensible edge from public information only — no private-info edge, by design.

2 · Markets reviewed

Ticker	Question (short)	Px (yes)	Vol 24h	Verdict
KXNEWTARIFFS-26JUN-JUL01	Any new/increased tariff EO in June 2026	82/83	2,410	PICK · BUY YES
KXTRUMPKENNEDYNAME-26-26JUL01	Trump name physically removed from Kennedy Ctr	49/50	2,322	PICK · BUY NO
KXJOINTRUMPADMIN-…-KPAX	Ken Paxton joins the administration	3/6	2,000	WATCHLIST (no fill)
KXSENATEREC-26MAY-JUN06	Reconciliation bill passes Senate by Jun 6	86/87	48,097	WATCHLIST
KXUSAIRANAGREEMENT-27-26JUL	US agrees to new Iran nuclear deal	10/11	60,394	WATCHLIST
KXRECNCH-26-JUN12	House passes reconciliation bill	89/90	10,265	PASS — priced
KXICEERO-26MAR-JUL01	ICE ERO funding becomes law by Jul 1	88/90	2,020	PASS — priced
KXFISAEXTEND-26MAY-JUN12	FISA 702 reauth becomes law by Jun 12	21/25	6,511	PASS — rules trap
KXFISAEXTEND-26MAY-JUN13	FISA 702 reauth becomes law by Jun 13	51/56	4,641	PASS — priced
KXKASHOUT-26APR-JUL01	Kash Patel leaves FBI by Jul 1	8/8	4,616	PASS — priced
KXTARIFFRATEPRC-26JUL01-5	China tariff rate below 10% on Jul 1	9/10	1,772	PASS — priced
KXVOTEHUBTRUMPUPDOWN-26JUN04	Trump approval > 39.3% (VoteHub) Jun 4	84/85	40,731	PASS — wrong source
KXMAMDANIEO-26JUN06-T0	Mamdani signs >0 executive orders this week	14/15	10,247	PASS — rules trap
KXHORMUZNORM-26MAR17-B260701	Hormuz transit-calls 7d MA > 60 by Jul 1	11/12	223,643	PASS — priced
KXAPRPOTUS-26JUN05-* (8 legs)	RCP approval ladder (mutually exclusive)	—	~20k	PASS — no arb
8× attendance / count markets	NBA/UFC attendance, Mar-a-Lago, post counts	—	1.5–79k	PASS — noise

Price shown as yes bid / yes ask in cents at the latest snapshot. "Priced" = direction correct but edge already eaten by a recent volume-confirmed move. Full reasons in §5.

3 · Picks

Pick 1 — KXNEWTARIFFS-26JUN-JUL01 · BUY YES @ 0.85 limit High conviction

Will Trump issue any executive action imposing new or increased tariffs on any country in June 2026 (with a specified effective date)?

Live (yes)82 / 83¢

Limit85¢

True prob97%

Edge+14¢

EV / $+17%

CloseJul 1

Mispricing thesis

The rule resolves YES on any new-or-increased tariff action with a June effective date — it does not net reductions against increases. The resolving event has, in effect, already occurred, and even if a reviewer disputes it, Trump's tariff cadence makes a second qualifying action near-certain over the 27 remaining days of June. At an 83¢ ask the market implies ~83%; a defensible estimate is ~97%.

Evidence

The June 1, 2026 Section 232 proclamation newly subjects mobile industrial equipment (bulldozers, forklifts), aluminum lithographic plates and steel racks to 15%/25% derivative tariffs, effective June 8, 2026 — an unambiguous "new tariff" with a specified June effective date. (White House Fact Sheet / GHY trade alert, 2026-06-01; Washington Post, 2026-06-02.)
Verifier re-pulled the exact DB rule: "…has taken any executive action imposing new or increased tariffs … provided the action specifies an effective date (even if such date occurs after expiration) in June 2026." The threshold is any action, not net direction — refuting the "it's really a tariff cut" tail.
USTR's June 2 Section 301 proposal (10–12.5% on 60 economies) is in the comment/hearing phase (hearing July 7) and would not independently qualify — but it underscores the live tariff pipeline. (USTR press release, 2026-06-02.)

Tail risk (the single cleanest way this loses)

Kalshi administratively reads the June 1 proclamation as net "ending/reducing" tariffs (the White House did badge some rate cuts) and no other June action is judged unambiguously "new/increased." Low: the rule is explicit that any new tariff suffices, and Trump issues tariff actions almost weekly.

Numbers · liquidity · price history

Numbers. Buy 850 @ ~84¢ avg (limit 85). Max payout $850; cost ≈ $714; EV ≈ +$110.50 (+15.5% on cost). Edge ≈ +14¢ vs the 83¢ ask.
Liquidity. YES is deep below the quote: ~$367 @ 83¢, $204 @ 84¢, $200 @ 85¢, $500 @ 86¢ (live book). 850 contracts fill ≈ 83–85¢, well inside the 3¢ slippage bar.
Price history. Only ~3 days of data; avg traded 50¢ (6/1) → 68¢ (6/2) → 82¢ (6/3) on confirming volume as the Section 232 / Section 301 news landed. Most of the move is in; residual edge is the gap between ~83% implied and ~97% true.

Pick 2 — KXTRUMPKENNEDYNAME-26-26JUL01 · BUY NO @ 0.53 limit Medium conviction

Will Trump's name be physically removed from the exterior signage of the Kennedy Center before Jul 1, 2026?

Live (no)50 / 51¢

Limit (no)53¢

True P(no)≈62%

Edge+11¢

EV / $+22%

CloseJul 1

Mispricing thesis

Resolution requires the actual physical removal of facade letters — not a court ruling, a website change, or a legal status change. The parties who physically control the building are openly refusing to act, and an appeal is pending. The market prices YES at ~49–50¢, inflated by a headline court order and a viral (fake) "letters coming down" video; true probability of removal-by-July-1 is closer to ~36% (so NO ≈ 62–66%).

Evidence

On May 29, 2026 U.S. District Judge Christopher Cooper ordered the signage removed within 14 days (~June 12), ruling only Congress can rename the center. (NPR, 2026-05-29.)
The Kennedy Center signaled it will appeal; Interior Sec. Doug Burgum refused to commit to removal. The actors who'd have to physically strip the letters are stonewalling. (Spectrum News 1, 2026-06-01; NBC News, 2026-05-31.)
A viral video purporting to show workers removing the letters was debunked as AI-generated (Poynter / PolitiFact) — i.e. the trigger event has not fired; the name is still up as of June 3. This is exactly the kind of misinformation that pumps a YES bid.
Verifier confirmed the DB rule requires removal of exterior signage specifically — a trap that cuts for NO, not against it.

Tail risk (the single cleanest way this loses)

Judge Cooper's binding ~June 12 deadline sits inside the resolution window. If a stay is denied and the Center complies — or settles by stripping the 18 added letters — before July 1, the market resolves YES and NO loses ~51¢. This is a genuine coin-flip-grade risk, which is why conviction is Medium, not High.

Numbers · liquidity · price history

Numbers. Buy 280 NO @ ~52¢ avg (limit 53). Max payout $280; cost ≈ $145.60; EV ≈ +$28 (+19%). Edge ≈ +11¢ vs the 51¢ ask, on a true-NO of ~62%.
Liquidity. Thin on the side we want: only ~$100–150 of NO fills at 51¢ before pushing to 52–53¢. Position is deliberately capped at $146 for this reason; do not chase above 53¢.
Price history. ~2 days of data: a violent 24¢↔73¢ swing on 2026-06-02 (vol ~5,417) as the ruling propagated, settling to ~45–52¢ on 6/3. News is largely in the price; the edge is the market over-weighting a court order that nobody is yet obeying.

Why no High-EV "near-certain" pick was deployed. The most attractive-looking near-certainty — KXJOINTRUMPADMIN-…-KPAX (Paxton is the GOP Senate nominee, true P(join admin) ≈ 2%) — failed the live-orderbook test. Buying NO requires lifting the 97.5¢ ask (the deep 92–94¢ size is the bid, where you'd sell), leaving ≈0.5¢ of edge. It is a clean thesis with no clean fill, so it is a watchlist resting-bid only, not deployed capital. Honest beats padded.

4 · Recommended $1,000 portfolio

Pick	Side	Limit	Contracts	Cost	Max payout	P(win)	EV $	EV %
KXNEWTARIFFS-26JUN-JUL01	YES	85¢	850	$714.00	$850	97%	+$110.50	+15.5%
KXTRUMPKENNEDYNAME-26-26JUL01	NO	53¢	280	$145.60	$280	62%	+$28.00	+19.2%
Deployed (2 picks)			1,130	$859.60	$1,130		+$138.50	+16.1%
Cash reserve				$140.40				14.0%

Risk profile

Worst case (both lose): −$859.60. Requires Trump to issue zero new June tariff actions and the Kennedy Center letters to come down before July 1 — joint probability ≈ 1%, and the two are essentially independent (federal tariff policy vs a D.C. court timeline).
Best case (both win): payout $1,130 on $859.60 cost = +$270.40 (+31% on deployed).
Most-likely outcome (~62%): both win → +$270.40. The other meaningful branch — tariff wins, Kennedy loses (~35%) — is roughly break-even by design: the tariff leg's ~$136 win is sized to absorb the Kennedy leg's ~$146 max loss.
Concentration. 83% of deployed capital sits in the tariff leg. That is deliberate: it is the higher-conviction, deeper-liquidity, near-certain position, and the Kennedy leg cannot absorb more size without >3¢ slippage. The single shared risk is Kalshi administrative interpretation on the tariff leg — not market/event risk.

Execution notes

Tariffs (YES). Limit-buy at 85¢, scaling down to rest at 83¢ for the back half. Invalidation: a Kalshi clarification that the June 1 proclamation does not count and no further June action — extremely unlikely. This is mostly resolved already; hold to expiry.
Kennedy (NO). Limit-buy at 53¢, hard-cap at $146 (thin book). Add trigger: if a stay pending appeal is granted, NO strengthens — add from reserve. Exit trigger: if facade-removal work is reported as actually underway, the thesis is dead — sell immediately.
Reserve ($140). Dry powder for (a) KPAX NO if a resting bid at ≤95¢ fills, or (b) KXSENATEREC YES at mid-80s if the Senate vote-a-rama is confirmed to conclude with final passage on Thu 6/4–Fri 6/5 (collapsing its timing tail).
Hedges. None natural; the two legs are uncorrelated, so the portfolio is its own diversification.

5 · What I rejected and why

The screen's value is in the misses. Most rejections share one cause: a real catalyst had already repriced the market on confirming volume within the prior 48 hours, leaving no edge.

Ticker	Why rejected
KXSENATEREC-26MAY-JUN06	Watchlist. Reconciliation cleared a 53–46 motion-to-proceed on 6/3; price ran 25¢→86¢ in 48h on heavy volume. True ~88% vs 87¢ ask ≈ 1¢ edge, and a real vote-a-rama timing tail (final passage slipping past the Sat 6/6 close). Worth re-entering if passage is confirmed Thu/Fri.
KXRECNCH-26-JUN12	Priced. House passage true ~88% vs 90¢ ask = −2¢. The 6/2 anti-weaponization-fund drop and 6/3 Senate vote already moved it +40¢ on confirming volume. No side clears friction.
KXICEERO-26MAR-JUL01	Priced. ICE-ERO funding becomes-law true ~90.5% sits on top of the 89–90¢ ask (~1.5¢ edge). Same Senate catalyst already re-rated it +20¢.
KXFISAEXTEND-26MAY-JUN12	Rules trap. Rule says legislation that "reauthorizes" §702 — a fresh short stopgap (the likeliest YES path) could be ruled a mere "extension," not a reauthorization, flipping resolution. New Pulte/Warner DNI fight is net-negative. True ~25% ≈ price; ambiguity could flip it.
KXFISAEXTEND-26MAY-JUN13	Priced + ambiguous. True ~53% sits inside the wide 51/56 spread; today's volume-confirmed crash already repriced the Pulte news; same "extension vs reauthorization" ambiguity. No defensible positive-EV side.
KXKASHOUT-26APR-JUL01	Priced. NO direction is right, but the market already decayed 77%→8¢ as Patel survived. True P(out) ~8% ≈ the 8¢ ask; ~0–3¢ edge into ~95¢ downside on a single-decision-maker scandal. Not worth it.
KXTARIFFRATEPRC-26JUL01-5	Priced. NO is correct (baseline China rate is exactly 10%, Section 122 runs to Jul 24), but YES collapsed 22¢→8–10¢ today on 5–9× volume. Only ~2¢ net edge after spread — chasing a fresh move.
KXVOTEHUBTRUMPUPDOWN-26JUN04	Wrong source. The diligence NO case leaned on Silver Bulletin/USPollingData (38.6–38.8%), but the market settles off VoteHub specifically, which reads 39.5% — above the 39.3% line. A time-weighted average won't drop 0.2pt in a day. Market's ~84% YES is correct; the apparent edge was a settlement-source error.
KXMAMDANIEO-26JUN06-T0	Rules trap. Looks like an 80¢ steal (Mamdani signed a "bedtime" order 6/1), but an 80-point gap on a trivially-researchable event is the market pricing a nuance: the Comic-Sans publicity stunt is likely logged as a proclamation, not a numbered EO, and these series settle off the official register. If anything the asymmetry favors NO.
KXUSAIRANAGREEMENT-27-26JUL	Watchlist / reject. The rule's object ("nuclear deal") with no named settlement source is genuinely ambiguous — a signed ceasefire/Hormuz MOU containing a nuclear-non-pursuit clause might or might not count. Comparable Polymarket (~broader wording) trades ~28%; Kalshi 10–11¢ already reflects the narrower reading. Thin 7¢ edge, low conviction.
KXHORMUZNORM-26MAR17-B260701	Priced. NO is right (7-day MA ~6–10 must reach >60 by Jul 1), but the identical Polymarket contract trades ~20% YES vs Kalshi's 11–12% — Kalshi is already more bearish. A pending US-Iran ceasefire-and-reopening deal makes the 3% true-prob estimate indefensible; edge ≈ 0 at the 89¢ NO.
KXAPRPOTUS-26JUN05 (8-leg ladder)	No structural arb. Mutually-exclusive RCP approval ladder sums to ≈99¢ across the bids and ≈111¢ across the asks — ordinary vig, no risk-free leg-sum mispricing to harvest.
8× attendance / count markets	Noise. Trump/Netanyahu/Adin Ross UFC & NBA attendance, Mar-a-Lago trip counts, Truth-Social post counts, presidential-action counts — all unforecastable personal whims or base-rate variance, and already priced near coin-flip. No public signal confers an edge.

6 · Sources

Tariffs (Pick 1)

Kennedy Center (Pick 2)

Rejections & context

Data & method. Market universe, rules text, and price/volume history from the read-only Kalshi mirror (trading_markets, trading_events, market_snapshots); live depth-of-book and current prices from the Kalshi public API (api.elections.kalshi.com); news from the primary sources linked above. Diligence and adversarial verification were run as parallel research agents and then re-checked against the live orderbook before sizing.

Disclaimer. Not investment advice. All probabilities are subjective estimates from public information as of 2026-06-03 and may be wrong. Prediction-market contracts can and do resolve to zero — you can lose 100% of capital on any position. Prices move; limits and sizes assume fills near the quotes shown and are not guaranteed.