BRADFORD POLYMARKET TELEMETRY AUDIT — 2 NOVEMBER 2026 Article: article-telemetry.txt (also copied to article-final.txt). Earlier draft preserved as article-public-narrative.txt. Scope and account selection The buyer cohort is every wallet with a direct Bradford Yes purchase from 2026-10-21 06:00:00 through just before 20:00:00 America/Toronto. There are 31 wallets and 73 purchase activity records. Their combined purchase value, price multiplied by shares, is $3,870.3320979977248 before fees. The five largest purchasers account for 70.83139197311681% of this amount. The full TRADE activity history of each wallet was retrieved for September 23 through October 23 inclusive, Toronto time. Epoch bounds are 1790136000 and 1792814399. All pages were followed until has_more=false, retaining all original filters with every cursor. The 31 histories contain 745,655 rows. Mrworldwide and the earlier 0x0a9d buyer were also audited; neither belongs to the 31-wallet October 21 buyer cohort. Canada classification Gamma /events?tag_id=245 was paginated to completion: 322 events. Twelve sports, celebrity and unrelated US events were excluded. The exclusion list is included in telemetry-account-audit.json. Activity was matched by event slug or market condition ID. Canadian-team sports and Toronto weather were excluded from the Canadian political/economic numerator. They remain in the all-market turnover denominator. Six of the 31 wallets have more than 50% of their turnover in the defined Canadian category. Three have 100%: Skirk1123, Grapoorrrr and torontotest. Twenty have less than 2%. The remaining majority-Canada accounts are bernhard-excepturi (52.83%, total turnover $764.37), RajahSingh (50.45%, $807.53) and proxy wallet 0xa7c8cf3b171898a0a00488d197f376d70086ef3e (50.77%, $18.48). These are period-specific trading shares. They do not establish lifetime specialization, residence, nationality or real identity. Mrworldwide's Canadian turnover was $177,518.30523749048 across 91 events, 38.826718791264646% of its $457,206.5597194607 total. Toronto contributed $10,015.182300181665. The earlier 0x0a9d buyer's Canadian share was 98.72%. Units and volume The v2 API documentation explicitly defines bare volume/size as shares. For this event, Gamma's Bradford volume field exactly matches summed taker shares in the legacy trade feed, not price*shares. Do not call the raw 549,054.55 Bradford volume field $549,054.55 of cash turnover. All dollar figures in the revised article are price*size before fees. Full-event market totals use taker_only=true so counterparties are not double-counted. Account turnover adds that account's purchases and sales, including its maker and taker activity. It is gross trading turnover, not deposits, net exposure or capital committed. The complete event taker feed contains 7,955 records, $637,508.400411886 in trade value over the whole market, and $245,342.89119687653 before 6 a.m. October 21. On October 21 it contains $17,646.39569599394 in total trade value: Bradford $8,596.221297984546 and Chow $9,050.174398009392. The separate $3,870 buyer figure is direct Bradford Yes purchases during 06:00–20:00. It must not be called the entire market's turnover or the cost of moving the chart from 35.5% to 55%. Execution reconstruction At timestamp 1792778196 (October 23, 13:56:36 EDT), Mrworldwide sold 5,777.88 Yes shares for $2,594.7451. Counterparty fills reconcile exactly: 265.75 shares at $0.46, 4,716.29 at $0.45, and 795.84 at $0.44. Skirk bought 3,454.29 shares for $1,554.4305 in this transaction. At the same timestamp, 0x11eaaf41ec9bf94cc33e3936242f17b7aa9e1676 sold 327.78 Yes shares for $121.5597. Reconciled fills were 104.93 at $0.44, 5 at $0.43, 15 at $0.42, and 202.85 at $0.33. There were no fills between $0.41 and $0.34 in this transaction. The first three levels total $54.6192 and the $0.33 fills total $66.9405. The two receipts are in Polygon block 96048550 at transaction indices 9 and 38. Receipt order establishes settlement order. Matching occurs off chain, so the article describes them as occurring in the same second and does not equate block order with a proven matching-engine sequence. Some counterparties sold No tokens, which can merge with the taker's Yes sale. Their Yes-equivalent execution price is 1 minus the No price. Ignoring these rows would fail to reconcile the seller's quantity and cash. Identical-looking rows are retained; separate fills can have identical wallet, price, size, timestamp and transaction hash. Price samples Bradford: Oct 21 06:59:23 = 0.355; Oct 21 19:07:18 = 0.55; Oct 23 13:56:35 = 0.47; Oct 23 13:57:24 = 0.38. Chow: Oct 23 14:05:32 = 0.62. These are displayed price history samples. A fill at 33 cents does not mean the displayed chart also reached 33% in that minute. Historical full order-book snapshots were unavailable. Executed price levels establish the size and range of the trades. They do not disclose all cancelled orders, reconstruct every intervening quote, or establish an individual trader's reason for placing an order. Reproduction Run python3 analyse-telemetry.py from this directory or by absolute path. It produces telemetry-account-audit.json, telemetry-volume.json and telemetry-executions.json. Its fill reconciliation assertions compare counterparty shares and Yes-equivalent cash with the taker's totals. fetch-account-audit.py and fetch-buyer-cohort.py retrieve the activity histories. fetch-event-takers.py retrieves all one-sided event trades. Raw complete histories, original cursor pages, receipts, metadata and price history are retained under raw/. Primary endpoints https://data-api.polymarket.com/v2/activity https://data-api.polymarket.com/v2/trades https://gamma-api.polymarket.com/events?tag_id=245 https://gamma-api.polymarket.com/events/slug/toronto-mayoral-election-winner-545 https://docs.polymarket.com/concepts/prices-orderbook https://data-api.polymarket.com/openapi.json The article links individual account histories, transaction receipts, price history and the public news sources beside the relevant passages. 4 NOVEMBER UPDATE — 50% CROSSING AND COMMENT REACTIONS The first October 21 minute sample above 50% is 12:39:24 EDT, at 51%. The preceding samples are 47.5% at 12:37:18 and 50% at 12:38:17. Skirk's two Bradford Yes purchases executed at 13:44:12, totaling $795. Grapoor's purchases executed at 14:25:00 ($130.48) and 14:56:44 ($99.96). Their combined purchases were 26.5% of the cohort's 06:00–20:00 total. These purchases came after the first crossing above 50%. Nearby trades and the four public comment reactions are preserved in crossing-and-interactions.json, with transaction and API source links. The four comment responses were fetched again on November 4. Reaction IDs and reacting proxy wallets were matched to the earlier saved records. Skirk hearted Grapoor comments 3309574, 3314008 and 3314218. Grapoor hearted Skirk comment 3307832. The comments were created October 24–26; the API does not timestamp reactions. These interactions do not establish coordinated trading.