Fourth & Value · Weekly review
NBA weekly recap: 2026-10-15 to 2026-10-21
NBA saved picks and forecasts graded against official results, with market comparisons, unresolved outcomes and the next watchlist.
What this week can tell us
Partial full-board archive coverage: saved game snapshots begin 2026-10-20. The full-board sample has 14 games with saved pregame quotes; 14 have official results available. Separately, 0 published picks cover 0 games; 0 of those games have official results. These are archived samples, not a claim to cover every league game.
The strongest relative probability score was spread / run line (12 observations); the weakest was rebounds and assists (124 observations). These rankings describe this sample; even the strongest market may trail sportsbook probabilities.
Published picks
Only the final morning and afternoon editions for each Eastern date are counted; afternoon picks (policy afternoon-edition-1 or later) are new picks from the 4:30 p.m. prices and never repeat a morning pick. One unit risked at each published price; pushes return the stake. Unresolved picks are excluded from ROI. Policy versions stay separate.
| Policy | Market | W–L–push | Net units | ROI | Unresolved |
|---|
No eligible published picks for this sport were present in the archived editions. Full-board forecasts below are not published picks.
Closing-line value report card
Each published pick’s price compared with the closing market: the median no-vig chance across US books at the pick’s exact line, captured in the final minutes before the start. Value at the close is how much more the published price paid than that closing chance justified; beat the close is the share of picks where it was positive; market move is how far the closing chance moved toward our side (positive) or away from it (negative) after the pick. Over many picks, value at the close is the earliest reliable sign of a real edge, long before win–loss results settle down. It is a measure of price, not a guarantee.
| Edition | Picks | Closing prices | Value at the close | Beat the close | Market move |
|---|---|---|---|---|---|
| Morning picks | 0 | 0 | — | — | — |
| Afternoon picks | 0 | 0 | — | — | — |
| All picks this week | 0 | 0 | — | — | — |
| All picks to date | 0 | 0 | — | — | — |
No published NBA picks have been registered for closing prices yet.
Full-board market results
One most-offered paired line per game, player and market from the final saved pregame snapshot. O/U columns count over / under / push; team markets count home win or cover / away / push. These are market outcomes, not a strategy return.
| Market | Over or home / other / push | Unresolved | Brier model / market (n) | Mean error model / line (n) |
|---|---|---|---|---|
| Moneyline | 7 / 5 / 0 | 2 | 0.1773 / 0.2036 (12) | — |
| Assists | 58 / 81 / 0 | 30 | 0.2582 / 0.2491 (134) | 1.72 / 1.66 (134) |
| Blocks | 16 / 17 / 0 | 5 | 0.2566 / 0.2359 (32) | 0.94 / 0.97 (32) |
| Points | 72 / 90 / 0 | 34 | 0.2527 / 0.2496 (157) | 5.41 / 5.26 (157) |
| Points and assists | 55 / 75 / 0 | 25 | 0.2612 / 0.2509 (125) | 6.19 / 5.80 (125) |
| Points and rebounds | 65 / 90 / 0 | 33 | 0.2500 / 0.2507 (150) | 6.11 / 5.71 (150) |
| Points, rebounds and assists | 58 / 80 / 0 | 31 | 0.2591 / 0.2487 (133) | 6.55 / 6.00 (133) |
| Rebounds | 79 / 79 / 0 | 34 | 0.2703 / 0.2450 (153) | 1.83 / 1.74 (153) |
| Rebounds and assists | 58 / 71 / 0 | 25 | 0.2747 / 0.2447 (124) | 2.56 / 2.34 (124) |
| Steals | 35 / 36 / 0 | 8 | 0.2359 / 0.2273 (70) | 0.86 / 0.94 (70) |
| Three-pointers | 55 / 89 / 0 | 26 | 0.2426 / 0.2402 (139) | 1.16 / 1.18 (139) |
| Spread / run line | 7 / 5 / 0 | 2 | 0.2056 / 0.2467 (12) | — |
| Game total | 4 / 8 / 0 | 2 | 0.2252 / 0.2511 (12) | 10.51 / 11.62 (12) |
Forecast quality
Across 1253 matched non-push observations in 12 games, model Brier was 0.2554, versus market 0.2456. Lower is better. The game-block 95% interval for model minus market Brier is -0.0097 to +0.0271.
Probabilities are compared on the same non-push outcomes after conditioning the model on non-push. Missing forecasts stay missing. Error comparisons use matched observations only. Small samples and correlated markets limit what a strong week can prove.
Next week's watchlist
- Portland Trail Blazers @ Sacramento Kings: saved model 226.39, line 232.5, gap -6.11. Snapshot Oct 22, 2026 at 06:47 AM EDT.
- New York Knicks @ Boston Celtics: saved model 218.82, line 223.5, gap -4.68. Snapshot Oct 22, 2026 at 06:47 AM EDT.
- Dallas Mavericks @ Indiana Pacers: saved model 240.37, line 236.5, gap +3.87. Snapshot Oct 22, 2026 at 06:47 AM EDT.
- Chicago Bulls @ Milwaukee Bucks: saved model 235.25, line 238.5, gap -3.25. Snapshot Oct 22, 2026 at 06:47 AM EDT.
- Houston Rockets @ San Antonio Spurs: saved model 229.63, line 226.5, gap +3.13. Snapshot Oct 22, 2026 at 06:47 AM EDT.
These are dated model-market disagreements to research, not validated edges. Check lineups, availability and current prices.
Evidence and limitations
participation unconfirmed: 31, evidence at or after official tipoff: 226
Official nonparticipants, unsupported contracts and missing statistics are unresolved, never automatic zeroes or winning unders. Standard full-game settlement is assumed; sportsbook exceptions require separate review. MLB pitcher props require an official starting appearance. Team totals are not part of the supported archive. Results may be revised when official feeds are corrected.
Download every graded row and source snapshot · Current NBA board · How updates work · All recaps