Viktor Eko | OrcaLayer

Over $100,000 has already left wallets Polymarket is investigating over Israel's election. Bloomberg found 600+ brand-new wallets backing Netanyahu or Likud right before polls closed. Netanyahu was 18.5 cents on election day. He's 47 now. I doubt that's 600 different people.

Karen Bass was the underdog for LA mayor on Polymarket at 3 a.m. Los Angeles time. Twenty minutes later she was the favorite. It's been that way all day. What did Polymarket see before LA woke up?

This wallet made $2.19M since June. Here's how. It belongs to a trader called gmpm2, and 97% of its markets on Polymarket are sports. The last seven days added $408K of it.

$1.93M up on Polymarket since February, and this wallet has put $89K on the Knicks to beat the Celtics tonight. The trader, called Talvez10, bought at 45 cents. Tip-off is 7pm ET. Biggest win so far: $291K, on Knicks vs 76ers. Win rate: 49.8%. Volume: $45.39M.

An Ohio metals distributor just hedged tungsten on Kalshi, with Susquehanna making the market. Tungsten has no listed futures. On Polymarket, sports did over 6x economics volume in the last day. If I ran either platform, I'd chase that Ohio buyer, not the next sports record.

GTA VI's launch trailer is at about 56.8 million YouTube views, six days in. Polymarket's odds it finishes week one between 60M and 80M went from under a third on Monday night to the clear favorite today. What does Polymarket know that the view counter doesn't?

This wallet made $6.98M this month. Here's how. It belongs to joseph00677, a Polymarket trader who started in October. The money came from one kind of bet. 1. Football favorites to win. Nine big Yes bets on clubs and national teams: Arsenal, Real Madrid, Atlético, Spain, Sweden. 2. Size. $254K to $3M a bet, $13.8M in total. Sunday's Arsenal bet was $3M at 66 cents and made $1.59M. 3. Price. Every entry between 53 and 83 cents. No long shots. 4. Timing. The winners were held 2 to 6 hours each. Match-day money, and it goes in slowly: 12,001 trades across just 19 markets. 5. Record. 8 of 9 won. The one loss, Deportivo at 83 cents, cost $448K. Those nine bets made $7.3M. Everything else lost about $330K. Those side bets also drag the overall win rate on the wallet down to 55.6%. But I'm not ready to call it an edge. At those prices you'd expect about 6 wins from 9. It got 8. Two extra wins over two weeks looks the same to me as a good run, and I'd want another ten of these bets before I believed otherwise.

$1.39M up on Polymarket since September, and this wallet has put $40K on tonight's college football game. The trader, called mooseborzoii, took Appalachian State to beat Coastal Carolina at 52 cents. Kickoff is 8pm ET. Biggest single win so far: $1.15M on Ravens vs Falcons. Win rate: 50.8%. Profit factor: 1.13x. I wouldn't copy this one. A coin-flip hit rate with this much size looks like a volume player to me, not someone with an edge on App State.

Late in a level football match, a team's price on Polymarket jumps from 0.24 to 0.50 in a few seconds. Everyone watching thinks the same thing: goal. A goal that late usually decides the match, so you buy that team before the price catches up. We tested exactly that. We recorded the order book of every Polymarket football market for three weeks and found 837 jumps like this in 459 matches. Buying the jumping team 2 seconds later lost $2.85 per $10 on average. Football has a 1-second delay on taker orders, so 2 seconds is about as fast as any bot gets. And faster was worse: at 0 seconds the loss was $4.21, at 30 seconds it was $1.07. The first seconds after a jump are when the price is most wrong in the direction you are buying. The reason is in the spread, and it is visible the moment the jump happens. When a real goal goes in, both sides of the book move together. The ask goes up, the bid goes up, the gap between them stays small. When sellers simply pull their quotes, only the ask moves. The mid-price jumps, but the bid stays where it was. The chart is one of those. The mid went from 0.24 to 0.505. But the ask was 0.78 and the bid was still about 0.23, a 55-cent gap. A bot buying 2 seconds later paid 0.81. Ten seconds after the jump the ask was back at 0.25. The match ended in a draw. That was not a rare case. 627 of the 837 jumps had a spread of 20 cents or more, and the "scoring" team went on to win only 36% of them. So we kept only the clean-looking jumps: spread of 4 cents or less. 85 left. The jumping team won 76% of them, and buying still lost $0.51 per $10. By the time you can buy, the goal is already in the price. Even a tight spread is not proof. The "level score" before the jump is read from prices too. In 51 of those 85, the three match prices before the jump added up to more than 1.10, which a normal book does not do. In one of them the team had been leading for 20 minutes. Its bids vanished for a moment, the score looked level, and the bids coming back looked like a late goal. Then the obvious fix: buy on goal signals, but add a stop-loss so the false ones don't ride to zero. We tried 28 stop settings, from 3 to 30 cents, watched from 2 minutes to the end of the match. On 1,004 replayed signals every one of them did worse than just holding. On 463 later live signals, none was clearly better than holding. A 3-cent stop fired on 86 to 96% of bets. Live football prices swing a few cents within minutes on every attack that comes to nothing. A stop turns each of those swings into a realized loss, and Polymarket has no stop orders, so every exit is a market sell that pays the taker fee a second time. What we did not test: pre-match markets, other sports, and stops on strategies that already make money. Three weeks of data, mostly club leagues. If you build anything in-play on Polymarket: check the spread before you call a jump a goal, check that the three match prices add up to about 1 before you trust the score, and test any stop against simply holding on the same signals. All 837 jumps with prices 0 to 60 seconds after each one are public, link below.

Apple closed above $5 trillion for the first time on Wednesday, and Nvidia fell about 3% the same day. Since Wednesday morning, Polymarket's odds that Apple ends 2026 as the world's largest company by market cap have more than doubled. What changed for Apple, other than one good Wednesday?