Polymarket — Live Account & Playbook
Live positions for 0xD1eED2…42A8ae from the Polymarket Data API, plus the top recommended add-ons and sells derived from the market-cap race analysis on the Largest Company tab. Everything refreshes every 60 seconds.
Live Account Positions — 0xD1eE…A8ae
Polymarket Data API · refreshes every 60s · zero-value expired positions hidden| Market | Side ⓘ | Shares | Avg | Last | Value | P&L ($) | P&L (%) | P&L (24hr) | Date |
|---|---|---|---|---|---|---|---|---|---|
| Apple — largest co. Dec 31↗ Ends 12/31/26 | YES | 5,051 | 19.0¢ | 28.1¢ | $1,416.94 | +$457.16 | +47.6% | — | — |
| NVIDIA — largest co. Dec 31↗ Ends 12/31/26 | NO | 2,865 | 35.2¢ | 42.5¢ | $1,217.80 | +$208.66 | +20.7% | — | — |
| S&P 500 (SPY) hit $770 high in July↗ Ends 8/1/26 | NO | 1,211 | 94.1¢ | 97.4¢ | $1,178.90 | +$38.90 | +3.4% | — | — |
| Bitcoin reach $75,000 in July↗ Ends 8/1/26 | NO | 933 | 97.6¢ | 99.5¢ | $927.50 | +$27.28 | +2.9% | — | — |
| S&P 500 (SPY) hit $700 low in July↗ Ends 8/1/26 | NO | 740 | 94.6¢ | 95.1¢ | $703.53 | +$3.53 | +0.5% | — | — |
| Bitcoin dip to $57,500 in July↗ Ends 8/1/26 | NO | 521 | 96.0¢ | 97.8¢ | $509.13 | +$9.12 | +1.8% | — | — |
| Bitcoin dip to $55,000 in July↗ Ends 8/1/26 | NO | 511 | 97.9¢ | 98.6¢ | $503.32 | +$3.32 | +0.7% | — | — |
| Paul Skenes — MLB strikeout leader↗ 2026 season | NO | 522 | 95.8¢ | 95.7¢ | $499.51 | −$0.49 | −0.1% | — | — |
| NVIDIA — largest co. July 31↗ Ends 7/31/26 | NO | 1,997 | 21.0¢ | 21.5¢ | $429.37 | +$9.16 | +2.2% | — | — |
| NVIDIA — 2nd largest July 31↗ Ends 7/31/26 | YES | 1,381 | 9.8¢ | 22.0¢ | $303.74 | +$168.36 | +124.4% | — | — |
| Crude Oil new all-time high by Dec 31↗ Ends 12/31/26 | YES | 1,830 | 20.0¢ | 13.5¢ | $247.05 | −$118.95 | −32.5% | — | — |
| Apple — largest co. July 31↗ Ends 7/31/26 | YES | 1,018 | 5.9¢ | 21.9¢ | $222.90 | +$162.72 | +270.4% | — | — |
| Atlanta Braves win 2026 World Series↗ Ends 10/31/26 | YES | 1,110 | 9.4¢ | 7.4¢ | $81.59 | −$23.41 | −22.3% | — | — |
| Dylan Cease — MLB strikeout leader↗ 2026 season | YES | 520 | 4.0¢ | 11.4¢ | $59.01 | +$38.40 | +186.4% | — | — |
| WTI Crude hit $100 high in July↗ Ends 8/1/26 | YES | 614 | 42.0¢ | 7.7¢ | $46.99 | −$211.01 | −81.8% | — | — |
| Amazon — largest co. Dec 31↗ Ends 12/31/26 | YES | 2,533 | 0.8¢ | 0.4¢ | $8.87 | −$10.55 | −54.3% | — | — |
Positions Value = Data-API account value minus spendable cash; Positions Value comes from the Data API; Cash in Wallet reads the wallet's USDC balances (native + bridged) live from Polygon; Portfolio = the two combined. Polymarket's "Available to trade" runs LOWER than Cash in Wallet whenever resting limit orders are open — the difference is collateral reserved by those orders, which Polymarket's private order engine does not expose publicly. The account has been actively executing this page's tickets on Jul 26 — the live refresh recomputes every row and stat from the raw APIs on each sync. Zero-value expired positions hidden; snapshot Jul 26, 2026 ~22:00 UTC.
Top 4 Add-Ons — All Positions
Global conviction ranking across the whole book — the best trade tops every list it appears on · not financial advice#1 ADD · HIGHEST CONVICTION Buy Apple-August YES @ ~23.5¢ — the biggest modeled edge anywhere
The single best trade on this page, full stop — which is why it is also #1 on the Largest Company list and #1 among new positions. The account holds Apple-July and Apple-December but skips August, the cheapest leg of the term structure (model fair 37–41¢), with both earnings catalysts inside the window. Same trade, same rank, everywhere it appears.
#2 ADD · SAME THESIS, 2ND VENUE Add NVDA "2nd-largest July" YES @ ~22¢ (held: 1,381 @ 9.8¢, +124%)
If Apple takes #1 by Jul 31, NVIDIA becomes #2 — this contract is the Apple-flip thesis traded on a different book. It prices at 22¢ while the model puts the flip near 30¢: the same edge as AAPL-Jul YES, in a second venue with its own liquidity. Adding here also splits execution risk across books instead of piling into one thin ask.
#3 ADD · COMPLETE THE PAIR Buy Alphabet-Dec NO @ ~91¢ — the vol-inversion hedge
Same trade as Largest Company #2: the account owns the long-challenger half of the December vol-inversion (AAPL-Dec YES) but not the short-Alphabet half. December prices Alphabet at ~5× Apple's implied relative vol — indefensible — and this leg also nets out the "NVDA stumbles" scenario the existing NVDA-Dec NO leans on. Ranked below #2 only because the NO side is capital-heavy per point of edge.
#4 ADD · AUG 1 ROLL ⏳ NOT ACTIONABLE TODAY Roll the July carry ladder — when August books go live
The five near-expiry winners — SPY $770 NO, SPY $700-low NO, and the three Bitcoin NOs at 95–99¢ — hand back ≈$3,590 in cash on Aug 1, and the roll into next month's ladders is the natural add-on. But the liquidity check kills any early action: as of Jul 26 the August SPY event is an empty shell — $58 of total liquidity with every strike parked at a 50/50 placeholder — and the August Bitcoin event isn't listed at all yet. There is no real counterparty behind those 50¢ quotes. Monthly ladders get seeded with market-maker books around month-end.
Ranking is global and congruent with the other lists: #1 here is #1 on the Largest Company list and #1 among suggested new positions. Next in line, below the cut: averaging up AAPL-Dec (LC #3), averaging down NVDA-Aug NO (LC #4), topping up the SPY $770 July NO carry, and a small momentum add to Dylan Cease.
Top 4 New Positions Suggested — Not Currently Held
Sourced across ALL Polymarket books (ranking partitions, carry ladders, convexity) · ranked by conviction · not financial advice#1 NEW · HIGHEST CONVICTION Apple-August YES @ ~23.5¢
The account holds Apple-July and Apple-December but not August — making the best trade on the page also a brand-new position. Biggest modeled edge anywhere (fair 37–41¢ vs 23.5¢), two catalysts inside the window, real book behind it. One thesis, ranked #1 on every list it appears on.
#2 NEW · CROSS-BOOK PARTITION The "Second Place, August" basket — buy BOTH Apple-2nd YES + NVDA-2nd YES @ ~88.5¢ combined
A different book family entirely: the 2nd-largest company August market. Whoever wins the #1 race, the loser of the NVDA/Apple duel is almost certainly #2 — the nearest third wheel (Alphabet) sits 26% below Apple with five weeks left. So Apple-2nd YES (~65.5¢) plus NVDA-2nd YES (23¢) pays $1 in essentially every world, for ~88.5¢ today: ≈+13% in five weeks, thesis-free. This is the same "sum should be 100" logic the Mispricing Monitor applies within books, exploited across them. ⚠ The Apple leg's quote is from the slower gamma feed — verify it on the live book before entry (the twice-daily auto-run re-checks it).
#3 NEW · STRUCTURAL SHORT Alphabet-Dec NO @ ~91¢
Not held, deepest book on the page, and the cleanest way to monetize the December vol-inversion: the market prices Alphabet-vs-NVDA at ~5× the implied volatility of Apple-vs-NVDA. Pairs naturally with the existing AAPL-Dec YES holding to complete a hedged relative-value structure.
#4 NEW · CONVEXITY PUNT Alphabet-August YES @ ~2.6¢ — smallest size, honest label
Alphabet needs ~+29% on NVDA in 26 trading days. At base vol (2%/day) 2.6¢ is slightly rich (fair 0.7¢) — but at the realized-vol regime just printed (4.5% relative in one session, Jul 24) fair climbs to 2.4–5¢. Strictly a bet that the high-vol regime persists: defined-risk lottery, not a modeled edge at base assumptions. Book: bid 2.4¢ × 60, ask 2.8¢ × 56 then 2.9¢ × 490.
Sourcing methodology: candidates are screened across book families — ranking partitions (1st/2nd/3rd-place books, where cross-book sums that should equal ~100¢ often don't), next-month carry ladders (SPY/BTC/WTI strikes — currently unseeded, see the Aug 1 roll card), and convexity punts — then ranked on one conviction scale with the largest-company trades. The twice-daily auto-run rescans all families as books seed and verifies any gamma-sourced quote against the live CLOB.
Top 4 Add-Ons — Largest Company
Derived from the Largest Company analysis (other tab) · not financial advice#1 ADD · NEW LEG Buy Apple-August YES @ ~23.5¢ — the date you don't own
The book holds Apple-July (+290%) and Apple-December (+47.6%) but skips August entirely — the cheapest leg of the whole term structure. 23.5¢ vs model fair 37–41¢, and the window spans both catalysts (Apple earnings Jul 30, NVIDIA earnings ~Aug 26). This is the #1 conviction trade on the Largest Company tab and the only date missing from the account. A $500 add at 23.5¢ returns ≈$2,128 if Apple takes #1 by Aug 31.
#2 ADD · COMPLETE THE PAIR Buy Alphabet-Dec NO @ ~90¢
The account owns the long-challenger half of the December vol-inversion trade (AAPL-Dec YES) but not the short-Alphabet half. December prices Alphabet at ~5× Apple's implied relative vol — adding GOOGL-Dec NO completes the relative-value pair and nets out the generic "NVDA stumbles" scenario that the existing NVDA-Dec NO already leans on. Capital-heavy (90¢ risked to win 10¢), so size modestly.
#3 ADD · AVERAGE UP Add to Apple-Dec YES @ 28.1¢ (held: 5,051 @ 19.0¢)
The book's biggest winner is still ~17.6 pts below σ=2%/day fair value (45.6¢). Averaging up while still under fair is the disciplined version of letting a winner run — a moderate add keeps the blended entry near 22–24¢. Reassess after Thursday's Apple print, and treat a cap-gap blowout past ~5% as the invalidation signal.
#4 ADD · AVERAGE DOWN, CAPPED Add NVDA-Aug NO @ ~26¢ (held: 939 @ 34.1¢, −23.7%)
The model says NVDA-August YES is ~17 pts rich — meaning the NO already held is cheap at 26¢ (model-fair NO ≈ 43¢), so averaging down has genuine edge behind it, not just hope. One discipline: the account is short NVDA on all three dates, which is a single correlated bet on the same story — cap the combined NVDA-short book at a fixed share of the portfolio before adding.
Top 4 Sells — What to Cut From Current Positions
Ranked by salvageable value · not financial advice#1 SELL · THESIS BROKEN SpaceX-Dec YES — 10,645 sh @ 4.9¢, now 1.5¢ (−69%)
SpaceX needs ≈+162% in five months to take #1; the diffusion model rounds that to zero and the market agrees at 1.5¢. The IPO-pop scenario that justified paying 4.9¢ has come and gone. $159.68 of salvage value beats a near-certain donut — redeploy straight into Add-On #1.
#2 SELL · DAYS FROM ZERO WTI $95 & $100 July YES — $85.28 + $36.55 left
Both July oil contracts expire Friday and need WTI far above anywhere July has traded. These decay to zero in four sessions; any bid is a gift. Combined salvage ≈ $122. The account's July book is otherwise healthy — these are the only two legs with no path.
#3 SELL · SLOW BLEED Crude Oil all-time-high-by-Dec YES — 1,830 sh @ 20¢, now 13.5¢
Needs a new all-time high (>$147) within five months — the same oil-spike thesis that already zeroed the March and April strips at a cost of ~$4,600. At $247.05 this is the largest recoverable chunk in the losers column. If oil-shock exposure is still wanted, the Aramco longshot legs on the ranking books are a cheaper expression of the same tail.
#4 SELL · DEAD WEIGHT The sub-$10 scraps — Pirates, Padres, AMZN-Dec
Pirates NL Central (1.1¢), Padres NL West (1.0¢) and Amazon-Dec (0.35¢) total ≈$23 with no live path between them. Selling frees little capital but cleans the book so open P&L tracks real bets. Leave the near-expiry winners alone — the SPY and Bitcoin NO legs at 95–99¢ resolve Aug 1 and are done compounding; they're cash in a week.
Largest Company on Earth — Odds vs. Reality
Polymarket runs three books on which company finishes as the world's largest by market cap: July 31, August 31 and December 31, 2026. This page compares live Polymarket odds against a probability model driven by the actual market-cap gap, flags where the odds don't reflect reality, and measures how fast the odds move. Live prices pull from the Polymarket CLOB API and Finnhub every 60 seconds.
Top 4 Trades — Ranked by Conviction, Sized to the Account
Tickets sized to the live account ($10,660.69 value · ≈$1,991 cash) against REAL order-book depth, Jul 26 ~18:00 UTC · not financial advice#1 · LONG Buy Apple-August YES @ ~23.5¢ → model fair 37–41¢
The calendar-kink trade. You pay 0.4¢ over the July contract for 21 extra trading days that span both catalysts — Apple earnings (Jul 30) and NVIDIA earnings (~Aug 26). The volatility implied by July's own 23.1¢ prices August at ≈37¢; σ=2%/day says 41¢. Edge ≈ +14 to +18 pts. If Apple flips #1 this week, this likely trades 60¢+ — consider taking half off after the print. Main risk: sustained NVDA out-performance decays it toward 0 over 5 weeks; a ≥10-pt adverse week happens ~33% of the time, so size to hold through one. Invalidated if the cap gap widens past ~5%.
#2 · PAIR Long Apple-Dec YES @ 28.1¢ / Short Alphabet-Dec YES (buy its NO @ ~90¢)
The vol-inversion trade. December prices Alphabet-vs-NVDA at ~5× the implied volatility of Apple-vs-NVDA — indefensible when Apple sits 2.3% from #1 and Alphabet 29% away. The pair wins under any consistent vol: at σ=2%/d Apple is 17.6 pts cheap while Alphabet is fair; at July's implied 1.4%/d Apple is still ~16 pts cheap and Alphabet becomes ~6 pts rich. Being long one challenger and short the other also nets out generic "NVDA stumbles" risk. Size roughly equal notional. Risks: an Alphabet-specific melt-up hurts both legs at once; both legs lock capital ~5 months, and the NO side is capital-heavy (90¢ to win 10¢).
#3 · PROCESS Fade flow moves in the July/August books
The repeatable one. Three times in 30 days these odds moved 10–36 pts with no confirming stock move — Jul 17's 92¢ → 53¢ → 89¢ round trip being the flagship. Mechanic: rest limit bids ~10 pts below live model fair and offers ~10 pts above, on both NVDA and AAPL legs (this page recomputes fair on every refresh). The volatility table says ≥10-pt weeks hit ~40% of the time and most are noise on thin books, so fills come to you and mean-revert. Discipline: never fade through a catalyst date (Jul 30, Aug 26), and pull orders if the stock tape confirms the odds move — some moves are information.
#4 · EVENT Small Apple-July YES @ 23.1¢ into Thursday's print
The lottery ticket — smallest size. Apple needs ~+2.4% relative to NVDA by Friday's close, with an earnings reaction session as the finale. Typical Apple earnings-day moves run ±3–5%, and Jul 24 alone delivered +4.5% relative. Zero-drift fair is ~30¢, so you're buying a live binary at cheap-to-fair with all the convexity on your side of the print. But it is binary: a soft quarter and this is worth 0 by Friday 4pm. Cap it at lottery-ticket size (≤1–2% of bankroll). Prefer less variance? Skip the print and run trade #3 on the overreaction afterward instead.
Mispricing Monitor — Where Odds May Not Reflect Reality
Ranked by size of dislocation1 · CALENDAR KINK Apple's term structure is broken
July AAPL trades 23.1¢ with 5 trading days left; August trades 23.5¢ with 26 days. You pay just 0.4¢ more for 21 extra trading days of optionality — including NVDA's own earnings (~Aug 26). The volatility that reproduces July's 23.1¢ (1.40%/day) puts August at ≈37¢ and December at ≈44¢; any richer vol assumption pushes both higher still. The market is implicitly assuming relative volatility collapses after Friday (implied σ: Jul 1.40%/d → Aug 0.63 → Dec 0.38) — while realized was ~4.5% relative in one session on Jul 24. Expressions: long AAPL-Aug YES outright, or the calendar spread (buy Aug 23.5 / sell Jul 23.1) which is near-flat if Apple flips this week and pure optionality if it doesn't.
2 · VOL INVERSION Alphabet priced 5× more volatile than Apple
Backing implied relative vol out of December prices: AAPL-Dec 28.1¢ implies σrel ≈ 0.38%/day; GOOGL-Dec 10¢ implies ≈ 1.86%/day — the book treats Alphabet-vs-NVDA as ~5× more volatile than Apple-vs-NVDA. Both pairs share the same NVDA leg; that ratio is not defensible. Either GOOGL-Dec is rich, AAPL-Dec is cheap, or both. Relative-value pair: long AAPL-Dec YES / short GOOGL-Dec YES, sized so a generic "NVDA stumbles" scenario nets flat.
3 · MOMENTUM PREMIUM NVDA priced with heavy drift
Zero-drift fair for NVDA is ~69 / 57 / 39 (Jul / Aug / Dec) vs market 78 / 74 / 57.5. To justify 57.5¢ in December, NVDA must out-grow Apple by ≈ 2% per month for five straight months (at σ=2%/d). That's the market's actual bet. If you believe in AI-capex momentum, the book is fine and the "cheap" flags above shrink; if you believe big-cap relative performance is near a random walk, NVDA NO is systematically overpriced across all three dates. Decide which world you're in before trading anything on this page.
4 · NO FREE LUNCH The books themselves are tight
Sum of YES mids: July 100.4¢, August 101.35¢, December 99.2¢. Buying NO on all 8 August legs locks ~+0.19% at mids (before spreads); buying every December YES including longshots earns ~+0.8% over five months if the winner is listed. Both are below carry once you cross real spreads (July NVDA book: 77 bid / 79 ask, only $40 at top-of-book). There is no Dutch book here — the edge on this page lives in relative pricing (cards 1–3), not in the sums.
The Analysis Table — Market Odds vs. Model Fair Value
Model: zero-drift lognormal on the cap ratio, σrel = 2%/day| Company | Mkt Cap ($B) | Gap to #1 | Polymarket Odds | Model Fair | Edge (pts) | Verdict | P(±5pts / wk) |
|---|---|---|---|---|---|---|---|
| JULY 31 — resolves in 5 trading days · book sum 100.4¢ · Apple earnings land Jul 30 after close, reaction session is resolution day | |||||||
| NVIDIA | 5,005.5 | — | 77.0% | 69.4% | +7.6 | RICH | 73% |
| Apple | 4,891.2 | 2.34% | 23.1% | 30.3% | −7.2 | CHEAP | 62% |
| Alphabet | 3,892.7 | 28.6% | 0.05% | 0.0% | +0.1 | FAIR | <1% |
| Microsoft | 2,835.4 | 76.5% | 0.05% | 0.0% | +0.1 | FAIR | <1% |
| Amazon | 2,496.8 | 100.5% | 0.05% | 0.0% | +0.1 | FAIR | <1% |
| AUGUST 31 — 26 trading days · book sum 101.35¢ · window contains NVDA earnings ≈ Aug 26 | |||||||
| NVIDIA | 5,005.5 | — | 74.0% | 57.1% | +16.9 | RICH | 67% |
| Apple | 4,891.2 | 2.34% | 23.5% | 41.0% | −17.5 | CHEAP | 67% |
| Alphabet | 3,892.7 | 28.6% | 2.6% | 0.7% | +1.9 | FAIR | ~10% |
| Amazon | 2,496.8 | 100.5% | 0.30% | 0.0% | +0.3 | FAIR | <1% |
| Broadcom | 1,817.0 | 175.5% | 0.30% | 0.0% | +0.3 | FAIR | <1% |
| DECEMBER 31 — 111 trading days · book sum 99.2¢ · Oct + Nov earnings rounds for all three leaders inside the window | |||||||
| NVIDIA | 5,005.5 | — | 57.5% | 39.1% | +18.4 | RICH | 69% |
| Apple | 4,891.2 | 2.34% | 28.1% | 45.6% | −17.6 | CHEAP | 50% |
| Alphabet | 3,892.7 | 28.6% | 10.0% | 11.6% | −1.6 | FAIR | 15% |
| SpaceX (SPCX, IPO'd Jun 12) | ≈1,914 | 161.5% | 1.50% | n/a† | — | WATCH | <1% |
| Tesla | 1,175.7 | 325.8% | 0.70% | 0.0% | +0.7 | FAIR | <1% |
† SpaceX needs ≈ +162% in 5 months to take #1 — the diffusion model isn't meaningful there; its 1.5¢ price is a pure momentum/squeeze lottery ticket. Caps refresh live from Finnhub (price × shares outstanding); Polymarket odds are CLOB midpoints. Snapshot baked Jul 26, 2026 17:15 UTC.
The July Race — Odds Since Listing
Daily CLOB closes, Jun 27 → Jul 26The December Race — Odds Since Jun 27
Daily CLOB closesHow Fast Do These Odds Move? (1-Day & 1-Week)
Empirical, daily closes since Jun 27 (Aug book since Jul 17)| Contract | σ 1-day (pts) | Avg |1-day| | Max 1-day | P(|1d| ≥ 5pts) | P(|1wk| ≥ 10pts) | Max 1-week | Model: pts per ±1% rel. move |
|---|---|---|---|---|---|---|---|
| NVDA · July | 6.9 | 4.7 | 17.5 | 43% | 46% | 37.0 | ±8.2 |
| AAPL · July | 6.7 | 4.2 | 18.3 | 27% | 31% | 35.1 | ±8.2 |
| NVDA · August | 9.1 | 7.4 | 17.5 | 50% | 33% | 25.5 | ±3.8 |
| AAPL · August | 9.8 | 7.3 | 19.5 | 40% | 50% | 24.5 | ±3.8 |
| NVDA · December | 4.0 | 3.1 | 10.5 | 23% | 23% | 18.5 | ±1.9 |
| AAPL · December | 3.9 | 2.6 | 11.7 | 13% | 19% | 21.5 | ±1.9 |
| GOOGL · December | 1.3 | 0.9 | 4.0 | 0% | 0% | 6.0 | ±0.9 |
| SPCX · December | 0.3 | 0.2 | 1.3 | 0% | 0% | 1.3 | — |
Read: the July NVDA contract moved an average of 4.7 points per day this month, with a 43% chance of a ≥5-point day and a 46% chance of a ≥10-point week. It swung 37 points inside one 5-day window (Jul 13→18: 92¢ → 55¢), with single-day jumps of 17.5 and 16.5 points. A typical Apple earnings-day stock move (±4%) maps to roughly ±30 points on the July book — the July market is, functionally, an option on Thursday's Apple print.
Resolution — Who Actually Decides, and Where It Can Go Wrong
Risk: LOW if finish >2% apart · ELEVATED if <1.5%The rule, verbatim
"This market will resolve to the largest company in the world by market cap on July 31, 2026, as of market close. The resolution source for this market will be a consensus of credible reporting." — identical wording across all three books. No named source; disputes go to the UMA optimistic oracle. Precedent: the June 2026 edition closed with NVIDIA the runaway consensus #1 — yet Polymarket's gamma API still served it as unresolved weeks later, and its cached odds for these books ran up to 2 days stale vs the CLOB during this analysis. Trust the CLOB feed, not cached ones.
"Credible reporting" is measurably fuzzy
The de facto scoreboard traders cite is companiesmarketcap.com — and on Jul 26 it was internally inconsistent: its front page showed Apple $4.63T while price × shares from the Jul 24 close gives $4.89T; its Apple detail page still said $4.44T "as of May 19," its Alphabet page $4.77T "as of May 17" (Alphabet is actually ~$3.89T). If the race finishes inside ~1–2%, which tally "credible reporting" converges on becomes the whole game. Compute price × latest 10-Q share count yourself; don't trust aggregators near the wire.
Known quirks in these specific books
• "Market close July 31" — Aramco's exchange (Tadawul) trades Sun–Thu, so its number is Thursday Jul 30's close. Irrelevant at today's gap, decisive only in an oil-shock tail. • SpaceX IPO'd Jun 12 (SPCX, ~$1.9T) — the December SpaceX leg now has an observable market cap; pre-IPO ambiguity is gone. • Share counts: Apple retires ~2.5%/yr via buybacks between 10-Qs; a sub-0.5% finish can flip on which share count a source uses. • Alphabet = all classes (A+B+C); most tallies agree, but not all. • August book still carries 16 unopened "Company A–P" placeholder markets at 50¢ with zero liquidity — ignore them; they are not tradable and not part of the sum.
Live plumbing on this page
Odds: Polymarket CLOB GET /midpoint?token_id=… per contract (no auth, CORS-open) every 60s. Caps: Finnhub quotes × fixed share counts, refreshed on the same tick. Model fair values, edges and verdict badges recompute client-side on every refresh (zero-drift lognormal, σ=2%/day, trading-day countdown with US holiday calendar). If either API is unreachable the page falls back to the baked snapshot and the pill in the header flips from LIVE to SNAPSHOT.
Methodology & Honest Caveats
Fair values use a zero-drift lognormal diffusion on the challenger/leader market-cap ratio: P(flip) = Φ(ln(capchal/caplead) / σ√T), T in trading days (5 / 26 / 111), base σrel = 2%/day with a 1.5–2.5% sensitivity band. That model assumes no persistent relative drift — precisely the assumption the market rejects (card 3), so treat "edge" as model-vs-market disagreement, not guaranteed profit. Empirical odds-move stats use ~30 daily CLOB closes (10 for the August book) and include weekend sessions. Mid prices are not executable: July top-of-book was 2¢ wide and $40 deep at best bid. Odds snapshot Jul 26, 2026 17:15 UTC; caps from Jul 24 close. Nothing here is financial advice; prediction markets can resolve against you on oracle judgment alone.