Methodology

Insider trades: SEC Form 4 filings from EDGAR. U.S. corporate insiders (officers, directors, 10%+ owners) must disclose trades in their own company's stock within two business days. We ingest every Form 4 daily. Consequence: what you see here can lag the actual trade by up to two business days — that lag is a legal reality, not a data error.

Market beliefs: Polymarket's public APIs. Once a company is linked to a related prediction market, we snapshot that market's probability once per day. The probability history starts the day the link is made.

The link between them: software suggests company↔market pairs by matching names and tickers against market questions, but every pair shown on this site was approved by a human. Nothing is auto-published.

Only open-market purchases and sales (SEC transaction codes P and S) count toward the score — trades where an insider voluntarily spent or received their own money at market prices. Everything else is stored and shown, but excluded from scoring as noise:

• Compensation mechanics — grants (A), option exercises (M), tax withholding (F): the employer's schedule, not the insider's conviction.
• Gifts (G), conversions (C), other codes — not market bets.
• Derivative-table rows — the reported "price" on these is an exercise price, not money at risk, so their dollar values would mislead.
• Rule 10b5-1 plan trades — trades executed under pre-arranged plans. We exclude a filing when its 10b5-1 checkbox is set or when its footnotes disclose a plan (the checkbox alone under-reports).
• Implausible values — single transactions over $1B are treated as probable filing typos (they occur!) and excluded from scoring.

For each approved company↔market pair, over a trailing 14-day window:

insider_signal = Σ value of open-market buys − Σ value of open-market sells
market_move    = probability today − probability 14 days ago
score = log₁₀(1 + |insider_signal|) × (1 + |market_move| × 5) × cluster_bonus

The cluster bonus (×1.5) applies when three or more distinct insiders traded the same direction in the window — several people acting alike is historically more informative than one. The log means a $10M signal doesn't drown a $1M signal; the market-move multiplier rewards pairs where the prediction market is actually moving. Pairs need at least $250K of net insider flow to appear on Trending.

Agree vs. diverge: the verdict compares what each side believes about the company, which depends on what a "yes" means for that question. For a bullish question ("Will X beat earnings?"), insiders buying while the probability rises = agree. For a bearish question ("Will X hit (LOW) $140?"), rising odds are bad news — insiders selling into that is agree: both negative, even though the raw directions look opposite. Opposite beliefs = diverge — the interesting case. Questions with no bullish/bearish reading (KPI ranges like "operating margin 34–36%?") are marked no direct read instead of being forced into a verdict. We label the pattern; we never claim to know why anyone traded.

The noise floor: Polymarket odds wiggle just from the gap between buyers and sellers — and that gap is different for every Polymarket market. So each one gets its own flat band: half its own bid-ask spread, never less than 1 percentage point (spread is smoothed over the last 7 nightly snapshots). An odds move smaller than the band counts as flat — everywhere. A flat Polymarket market gets "held steady" language instead of "rose" or "fell", and no agree/diverge verdict (shown as polymarket flat): a wiggle inside the spread is noise, not a change in belief, and we never let noise pick a side.

Polymarket depth (the dots): every night we also snapshot each Polymarket market's order book and measure the dollars sitting within 5 cents of the current odds — the amount you could actually bet without moving the odds. Cards and rows show it as ●●● deep ($1,000+) · ●●○ tradeable ($100+) · ●○○ thin (under $100). This describes only the Polymarket side — the stock trades normally on the stock market. Thin Polymarket markets can show dramatic-looking odds "moves" that are really one small order — weight them accordingly.

The score measures magnitude, not disagreement. Both inputs are absolute values, so a pair where insiders and the market agree can score just as high as one where they diverge — the score says "a lot is happening here," and the separate agree/diverge verdict says which way. (The score was originally labeled "divergence"; it was renamed "signal" because that is what it actually measures.)

What's a high score? The score is not out of 10 or 100 — it has no ceiling, and its only job is to rank. Because of the log, its first digit roughly equals the number of zeros in the insider dollars: $1M of net flow ≈ 6, $10M ≈ 7, $100M ≈ 8, before the market-move and cluster multipliers scale it up. The $250K entry floor makes ~5.4 the lowest score that can appear. In practice most entries land between 5 and 15; anything above 20 is exceptional — large money, a large market swing, and several insiders moving together at once.

⚡ Actionable now: most signals are history — the odds already moved, or the market closed. This badge marks the ones that are still live: the market is open, net insider flow clears the $250K floor, the question has a clear bullish/bearish reading, Polymarket's probability has moved 5 points or less since the most recent filing became public (measured from the price on the filing date; for markets we linked after the filing, from their first snapshot), and the market does not already agree with the insiders — when both sides believe the same thing, the information is already priced in and there is nothing left to act on. Badged pairs sort to the top of Trending and the dashboard, and carry two arrows showing which way the stock and the market's YES probability move if the filings prove right. The signal serves two audiences, so actionability is shown per venue with two badges. ⚡📈 STOCK appears on every live setup — the insider filing is fresh and Polymarket's odds haven't priced it in. The blue ⚡🎲 POLYMARKET badge appears only when that Polymarket market is also tradeable: spread of 5 points or tighter with at least $100 waiting to trade near the shown Polymarket odds. No blue badge means that particular Polymarket market is too thin to bet on near its shown odds — the stock-side setup still stands. Two honest limits: we do not yet check whether the stock already moved on the filing (that check is planned once we have measured how stocks typically react), and the badges do not mean the insiders are right. Not investment advice.

The Wallets page grades Polymarket accounts on resolved stock-linked markets, using the public trade tape. Every fill is scored against the market's final payout (a buy of 100 shares at 19¢ that resolves yes earns $81), and each fill carries an earliness weight: 1.0 at the moment the market opened, falling to 0 at close — so buying near-certainties seconds before resolution scores near zero, while a correct position taken weeks early keeps full credit. Each market is graded once, permanently: history only deepens, and the leaderboard can be viewed over 90 days, 1 year, or all time. Wallets qualify with at least 5 graded markets and $200 staked; fills under $10 are ignored, and a market only counts toward a wallet's record once they staked $25 in it.

The consistency grade answers a different question than profit: does the wallet's running-profit chart climb steadily, or is it spikes and luck? It combines the chart's slope relative to its wobble (the K-ratio from the trading literature), how deep and long its losing stretches were (the Ulcer Index — only drawdowns count, upside swings are free), and whether the wallet is still active (the grade fades with a 90-day half-life after its last resolved bet). Grades are relative to every other qualified wallet. Records under 30 graded markets get no grade at all, and records under 50 are pulled toward the middle — a short hot streak is usually luck, per the small-sample corrections of Bailey & López de Prado. The final ranking is weighted profit × a consistency factor (0.5–1.0); both parts are shown so the ranking is auditable.

The tape attributes the taker side of each trade only, so per-wallet volume is a floor. And the name matters: these are sharp wallets, not "insider" wallets — a repeated record of beating the odds is observable fact; why a wallet is sharp (better information, better models, or luck that hasn't run out) is not something a trade tape can prove. A high consistency grade raises the odds of skill; it never promises the next bet.

Who leans where (company pages, and the SMART MONEY ALIGNED chip on Trending) lines up three independent reads on the same stock: what SEC insiders did (net open-market Form 4 flow over the score window), what the top-ranked sharp wallets hold (the top-100 wallets from the Wallets leaderboard, counting open positions of $50+ in that Polymarket market, refreshed nightly with the radar), and which way the Polymarket crowd has moved the odds (with the same noise band the verdicts use). A wallet's Yes/No side is translated into a stock direction through the question's meaning — holding YES on a bankruptcy-style question is a bearish stance — and questions with no up-or-down meaning for the stock are shown as "no direct read", never guessed. The headline appears only when insiders and sharp wallets genuinely lean the same way: gold when the Polymarket odds haven't followed (or lean against), blue when all three already point the same way — which can also mean it's priced in.

Follow cost (the Wallets radar) answers the copy-trader's first question: can you still get the price this wallet got? Green means the Polymarket odds now sit within that Polymarket market's own noise band of the wallet's entry (or below it); amber shows the extra cents per share you'd pay; and "moved" means the Polymarket odds are 5+ points past their entry — the same sitewide drift line the ⚡ badge uses — so much of the move that wallet saw coming has already happened. It compares prices only; it never accounts for fees or for the market moving while you trade. Not investment advice.

The True P&L board (Wallets page) rebuilds each tracked wallet's entire Polymarket cash history — every buy, sell, redemption, split and merge back to its first fill — and sorts every dollar into one of two piles: trading (money won from or lost to other traders, plus what open positions are worth at today's Polymarket odds) and Polymarket program income (liquidity rewards, yield, fee rebates, referral bonuses — money the platform pays out, not money out-traded from anyone). Deposits and withdrawals never count as profit or loss in either pile.

Two rankings sit on the same rows. True traders ranks by trading profit alone — the pure "who wins by being right" list. Total gains adds program income back in, so wallets that earn mostly by collecting rewards stay visible instead of disappearing — knowing who they are is intel in its own right. A wallet appears only once its full history has been crawled; a partial history would show a wrong total, so the board grows as the nightly crawl finishes each wallet. Program income here means income paid by Polymarket itself — tokens airdropped to a wallet outside Polymarket are not visible in this data. Polymarket's own leaderboard number is shown alongside for comparison; its formula is undocumented and blends the two piles.

The Big money page shows the largest money that moved yesterday (US-Eastern day) in the stock-linked Polymarket markets this site tracks. One line is one wallet's whole day in one Polymarket market, in one direction — all of its fills added together, because a position built in forty small trades is one conviction, not forty. Each line is stamped with the wallet's leaderboard record as of that night: rank, markets won, earliness-weighted profit, consistency grade. The stamp is frozen at capture — it shows what was knowable that evening, and it never retro-updates.

"Big" is deliberately relative, not a dollar rule: the tape keeps the day's top convictions by size, whatever the day produced. The graded-money lean below the tape nets everything wallets with a graded record traded in each Polymarket market — buying YES and selling NO both push toward YES — across all of the day's trades, not just the lines big enough to make the tape. The public tape credits one side of each trade, so every number is a floor on true activity. A lean shows where proven wallets put money; it is not a prediction.

The Earnings page covers every tracked Polymarket market that resolves on an upcoming earnings report (its question turns on earnings, EPS or revenue, and it has an approved company pair). For each one, the full public tape since the market listed is netted per wallet into one direction — buying YES and selling NO both push toward YES — and the top convictions are shown with the same frozen leaderboard stamp the Big money tape uses. Two flags are derived, never invented: fresh wallet means the wallet's first activity anywhere on Polymarket came after this market listed (checked against Polymarket's public activity feed — no made-up "newer than N days" cutoff), and underdog entry means the wallet bought the side the Polymarket odds priced below 50¢ at their average entry.

The SEC filing strip on each company counts open-market insider buys and sells (Form 4) filed since the earliest of that company's earnings markets listed — the same window as the Polymarket flow, so the two timelines are directly comparable. Sales under a pre-set 10b5-1 plan are labeled, because they were scheduled long before the window. Everything on the page is a fact about money — size, direction, timing. Unusual conviction is not an accusation, most of these markets are thin, and the taker-side tape makes every figure a floor. Not investment advice.

• Form 4 lags trades by up to 2 business days; daily processing adds up to one more.
• A market's probability history begins when we link it — young links show "market flat / young" until 14 days accumulate.
• Insiders sell for many reasons (diversification, taxes, houses); buys are generally more informative than sells.
• Filings occasionally contain filer errors; we flag the implausible ones but cannot catch everything. Every number links to its source so you can check.
• This is a data-journalism project — not investment advice, and nothing here implies wrongdoing by anyone named.