SHADOW Tracking only — not a betting record yet. The engine has not shown it beats the Kalshi price; every pick is logged (record_type=shadow) before any of it counts.

Prices: Kalshi taker, fee incl. · DK = DraftKings American via ESPN, vig incl. · Board quoted 9:02 pm ET · Injury report 9:00 pm ET

*Ours = the market-anchored probability (the no-vig market — the Kalshi mid where Kalshi quotes the line tight, else DraftKings de-vigged — nudged 25% toward the engine), not the raw engine.

NBA Labs · Shadow board · Friday, October 30, 2026

The Lab
Never Lies.

Every Kalshi and DraftKings player-prop line on the NBA slate, priced by the engine and anchored to the market. The price, the line, the number. Shadow season: every pick is tracked in the open before any of it counts.

Shadow plays today1
Games on the slate10
Track recordShadow · graded nightly
NBA · SHADOWDe'Anthony MeltonGSW
SHADOW
TOP EV
REBLAL @ GSW10:00 pm ET
De'Anthony Melton
OVER 3.5+195
GSW vs LAL · DK, vig incl.
EV per $1
+3¢
Ours* vs market
37% / 36%

Shadow top EV: De'Anthony Melton OVER 3.5. Ours* 37% vs the Kalshi mid 36% at DK +195, engine at 3.3 reb. Tracked, not sold — the engine has not yet shown it beats the Kalshi price.

Track recordShadow season · graded nightly

Shadow season — every pick is graded off the NBA's official game logs and tracked in the open on the Track Record: by night, by market, by tier and by price. Shadow, not a betting record — the official record starts only when a lane is promoted (checkpoint 2026-11-16).

Follow the Lab @CTBetz_

The Lab posts on X. NBA Labs runs in the open: every shadow pick is tracked before any of it is sold as an edge.

@CTBetz_

All picks

Every shadow pick on today’s slate, once, in tip order — PLAY first within a game. PLAY cleared the full gate; LEAN is the best test line on a game with no PLAY — a lower-confidence look, not an edge. Open one for the full read.

PLAY cleared the gateLEAN test line, lower confidence
Shadow picks2
CHA @ WAS7:00 pm ETNo shadow pick on this game — nothing here cleared the gate.
CHI @ BOS7:30 pm ETNo shadow pick on this game — nothing here cleared the gate.
DET @ BKN7:30 pm ETNo shadow pick on this game — nothing here cleared the gate.
NYK @ PHI7:30 pm ETNo shadow pick on this game — nothing here cleared the gate.
ORL @ TOR7:30 pm ETNo shadow pick on this game — nothing here cleared the gate.
HOU @ DAL8:00 pm ETNo shadow pick on this game — nothing here cleared the gate.
IND @ MIA8:00 pm ETNo shadow pick on this game — nothing here cleared the gate.
LAC @ MIN8:00 pm ETNo shadow pick on this game — nothing here cleared the gate.

The PLAY Board

Every pick that cleared the full shadow gate, ranked by EV per $1. Price = the American price at the venue (Kalshi also shows what a contract costs with the fee, DraftKings is vig incl.). Shadow: tracked, not a betting record yet.

Shadow plays1

Engine projects 3.3 reb in 22 min against the 3.5 line. Ours* 37% vs the Kalshi mid 36%.

EV per $1
+3¢
Ours*
37%
market-anchored
Market
36%
Kalshi mid
Price
+195
DK, vig incl.
Proj3.3 rebProj min22Also Kalshi39¢
The full read
  • Engine: 3.3 reb in 22 projected min vs the 3.5 line.
  • Taking OVER at DK +195 = 34% implied (vig incl.) · market 36% from the Kalshi mid · Ours* 37% → EV +3¢ per $1.
  • Out tonight: Kristaps Porziņģis (23 min) — minutes not redistributed yet.
  • Cleared every shadow gate check. Locked 9:02 pm ET off the DK board.

The LEAN Board

The best test line on each game with no PLAY, ranked by EV per $1 — held back by the gate, lower confidence, not an edge. Tracked in the open like every pick.

Test lines1

The Track Record

Units over time, every pick graded — the real number, win or lose, once there is an official record.

No official record yet.NBA Labs is in its shadow season: every pick is locked, graded and shown on the Track Record page (record_type=shadow) — tracking, not a betting record. The official W-L starts at the anchor and never includes a shadow night. First checkpoint 2026-11-16.The shadow season, night by night — tracking, not a betting record →