Skip to content
CALIBRE
← Radar
Sealed verdict

Spread: Thunder (-3.5)

sharp-led Most of the money behind the move came from wallets with a positive learned edge.

THUNDER price
49% → 53%
2026-11-06 13:45 UTC to 2026-11-06 14:10 UTC
CALIBRE forecast
52.3%
-0.2 pts vs the 52.5% price
Traded by
27 wallets
31 fills in the move
Learned-edge share
99%
of the money pushing the move

Square money

Share of the money pushing this move that came from wallets with a proven losing record, read from the wallet index when the verdict was sealed. The six biggest pushers were judged; the rest count as not judged, which is why the coverage matters.

Square money 0% of the money pushing the move, calm. 99% of that money could be judged.0%CALM0100%

Judged 99% of the pushing money

0% of the money pushing the price came from wallets with a proven losing record; 99% of it belonged to wallets the index could judge. Caution signal. In a replay of 11,975 moves, loser-driven moves did 3.1 pts worse for followers (p = 0.07); early evidence, not conclusive.

Who traded it

The largest wallets in the move, with the record CALIBRE held for them when it sealed the verdict. Open a wallet for its full live audit.

Of the 5 largest wallets here, none has a proven edge by CALIBRE's rules (z of 2 or more over at least 100 bets). The label above describes edge CALIBRE had learned, not a claim that anyone behind this move is skilled.

WalletNet bet on THUNDEREdge heldzBetsX-ray
0xfe78…0319+$18,769+2.6 pts0.7106No proven edge
0x958b…f826+$1,102+5.5 pts1.222No proven edge
0x8314…0f1e+$500+8.0 pts2.252Proven winning record
0x5268…135d+$292+2.6 pts0.9218No proven edge
0xaa0e…2a0e-$240+0.0 pts-1.616No proven edge

"Edge held" is the wallet's edge after shrinkage for small samples and a cap of 8 points, in the NBA category. A positive number here is not evidence of skill unless z and the bet count clear the bar.

Receipt

Sealed in Walrus Memory
N00bbMMgSuk8AGdr7HC8ADwp_Ryx7evARQzdK5Psd_YWalruscan shows that this blob exists on Sui mainnet and the transactions that stored it, not the text inside it: Walrus Memory encrypts every memory. What it proves is that the record existed before the market resolved. The verdict above is that record, decrypted, and pnpm verify rebuilds the proof page from Walrus alone rather than from anything stored here.
Sui checkpoint
342300168 at 2026-11-06 14:13 UTCRead through Tatum's Sui gRPC gateway when the verdict was sealed.
Sealed at
2026-11-06 14:13 UTC
Judging model
v3Model v3 judges like v2 and also records the square-money share and each top wallet's X-ray label, read from the wallet index.
Memory key
0xd4cbaadfa69dce41c7399a0c67aabc29fac0898c9db6221359b557f0b1c1e781:1793974200

Data source

Polymarket public API (fallback). The market was found through Tatum's Prediction Markets API and the checkpoint was read through Tatum's Sui gRPC. The trades behind the move came from Polymarket's public Data API as a fallback (Tatum trades returned no fills for a market that just moved).

Graded status

Graded Resolved THUNDER. The forecast scored a Brier of 0.228 against 0.226 for the market price (lost to the price).

Outcome blob

Sealed summary: NBA move up on "Spread: Thunder (-3.5)" (THUNDER 49% to 53%) is sharp-led: 99% of the money behind it came from wallets with a learned edge. CALIBRE forecast 52% for THUNDER. Square money 0% of the money pushing the move (judged 99%). Data: Polymarket public API (Tatum trades returned no fills for a market that just moved)