Attention Polymarket Data Geeks. I have built military grade orderbook capture and provided it for free. 🪖 Growing at 2.3 Billion rows per day. Check it out and pls share.

I see some were already answered, so here are the ones still open 🫡 July taker rebates: your rebate should be your tier % of the taker fees you paid. For the first five weeks (fee days Jun 1 – Jul 2), 0x520b undercounted fees on buys: each buy fee only counted at the share price. Buy at 20¢ and only 20% of that fee counted toward your rebate; sell fees counted in full. They fixed it from Jul 3. The Jul 8–9 payment from 0x71a7 is the top-up for that mistake. For each wallet it’s the rebate you should have got for Jun 1 – Jul 8, minus what 0x520b had already paid. Your tier each day is based on your previous 30 days of weighted volume, combos don’t count, and Jun 5–6 are skipped. For most wallets in the 3%, 8% and 18% lists, it matches to the cent. That also explains the gaps in the early payments. The Jun 23 payout also covered Jun 1–4 and Jun 7–18, and the Jul 1 payout covered Jun 25, 27 and 29. Rebates for May 28–31 and Jun 5–6 were never paid at all. The last $0.30 from the Sep 22 batch (0x53d1) is a refund on “Set 1 Winner: Dakotah Bobo vs Corey Craig”. Polymarket posted an on-chain clarification that evening resolving it to Bobo and refunding net losing positions, so all 54 payments are now explained. Bonus: 0x71a7’s round-dollar payouts on Sep 28 (72 wallets) and Oct 2 (22 wallets) look like payments from the Sep 19 trading-disruption.

On the 22 Sep batch (0xb1d7b542...8bd1), we reconstructed all 54 payments totaling 149.40 pUSD. The strongest explanation is mixed reimbursement of selected fees and purchase costs, rather than one market. 52 recipients received 146.97 pUSD. Across 107 selected fills, the recorded fee components total 146.985350 before per-wallet rounding, on four markets: • Russia military action against an EU country by 30 Sep and 31 Dec 2026. • EU country blames Russia for a new sabotage attack by 31 Oct and 31 Dec 2026. One combined builder charge is included, so this is not entirely platform fees. Archived Polymarket event pages show those exact conditions with fee coefficient 0.04 on 19 Sep and zero on 20 Sep. That is a fee-curve coefficient, not a flat 4% charge. The captures do not pin down the effective change or prove an incorrect charge. The remaining payments: • 2.13 pUSD matches 42.54736 Corey Craig match-winner shares × 0.05 = 2.127368, in Bobo/Craig. • 0.30 pUSD has two matching candidates: Craig first-set (5 × 0.06) or Maltsev vs Pytsko (10 × 0.03). The first-set market settled after the payment; Maltsev settled losing before it. The combined final ITF main and qualifying draws contain both Bobo and Craig but no pairing between them. That supports investigating the fixture, not declaring it canceled or wrongly resolved. Rounding matters: decimal half-up matches 53/54. One 0.495 basis was paid as 0.49. Tested half-down and binary-rounding variants match 54/54, but do not identify the payer's implementation. The 0.30 attribution and payment purpose remain unconfirmed. We have the 54-wallet comparison and transaction/log references prepared. This analysis is specific to the September batch; it does not explain the July or weekly payments. Military-market archives, before and after: web.archive.org/web/20260919… web.archive.org/web/20260920… ITF final main and qualifying draws: itftennis.com/en/tournament/… itftennis.com/en/tournament/…

Attention Polymarket Data Geeks. I have built military grade orderbook capture and provided it for free. 🪖 Growing at 2.3 Billion rows per day. Check it out and pls share.