Whose trades were followed by the biggest market-beating moves: how each stock moved after a member traded it, against the S&P 500.
Each figure is how a stock moved after a member traded it, against the S&P 500 (SPY) over the same trading days — the stock’s move, not the member’s profit. Not investment advice, and a disclosed trade is not evidence of wrongdoing.
What is measured. A decision is one member, one ticker, one direction, one filing: thirty rows for one stock in one report count once. For each, the stock’s close-to-close move over 30 or 90 trading days minus SPY’s over the same days, read the way the trade pointed — a sale followed by a fall scores positive. Closes are split-adjusted; dividends are ignored on both sides.
Two clocks. “After the trade” starts at the first close on or after the trade date. “After disclosure” starts at the first close after the report was filed — the earliest a member of the public could have copied it. Both are measured on the same decisions; the gap is the move that happened before anyone could see the trade.
What is unknown. When or whether the position was sold, and its size: amounts are disclosed brackets, so no dollar figure enters this and every decision counts once. Options, bonds, Treasuries, funds and exchanges are left out, as are rows read from scanned paper filings and rows whose filed dates contradict each other — the counts are under “Coverage”. Tickers with no price series (delisted, renamed or acquired companies) drop out, which flatters whoever held the survivors.
Small samples are pulled toward zero. A member’s average is shrunk toward zero by n / (n + k), with k = 21 estimated from how much members differ beyond chance: a handful of lucky picks cannot top the table. The bar is a 90% interval from re-drawing the member’s filings at random; when it crosses zero the data cannot tell the record from no edge at all. Members need at least 8 measured decisions, from at least 3 separate filings, to be ranked: one report is one moment in the market, however many tickers it lists.
The luck check. “Beats 87%” means the member’s average was higher than 87% of random portfolios with the same dates and buy/sell mix, each ticker swapped for one another member traded that week. It is a comparison, not a probability that the record is skill.
Recent trades are missing. A decision counts only once its horizon has passed on both clocks, so the most recent months are not in the table yet.
60 members with fewer than 8 decisions measured over 30 trading days, or fewer than 3 filings behind them. Listed, never ranked.
Rows without a ticker or a direction form no decision, so their decision count is blank. Data as of the pipeline’s last track-record pass (2026-10-24).