The LINE Report Card
Every pick. Every result.
Every official NFL pick since the record began, wins and misses alike, plus the research record for college games and the limits of all of it.
Updated Nov 11, 11:39 AM ET from the official log.
81.3%. 13 of 16 official NFL winners called correctly
Oct 29 – Nov 9, 2026 · Weeks 8–9
- Record
- 13–3
- Model
- LINE 2.0 v1 (frozen), powered by CMOS
- Recorded
- All 16 locked before kickoff, fingerprinted
- Type
- Official live picks. No backtests.
Recent resultsFinal scores, away–home
- BUF at MINWeek 9 · LINE picked BUF23–6Correct
- TB at CHIWeek 9 · LINE picked CHI24–27Correct
- CHI at SEAWeek 8 · LINE picked SEA34–17Missed
- PHI at WASWeek 8 · LINE picked WAS10–27Correct
- NE at MIAWeek 8 · LINE picked NE24–19Correct
- KC at DENWeek 8 · LINE picked DEN24–28Correct
Small sample: 16 games. A few results can swing these numbers. Winner picks, not betting results: calling winners isn’t the same as beating the spread or making money. Backtests are kept separate (63.6% of 1,519 games, 2020–2025) and never counted here. Updated November 11, 2026.
Official NFL record · 2026 season
The record starts in Week 8, when LINE began locking official picks before kickoff. A pick is correct when LINE’s favored team (win probability of 50% or more) wins. Ties have no winner and aren’t scored. “Avg. margin miss” is how far the predicted point margin was from the real one, on average; lower is better. The official log’s integrity check passes for all 16 verified games. These are winner picks, not betting results.
Week by week
| Week | Locked | Correct | Missed | Accuracy | Favorite |
|---|---|---|---|---|---|
| Week 8 | 14 | 11 | 3 | 78.6% | 10–4 |
| Week 9 | 2 | 2 | 0 | 100.0% | 2–0 |
Every official pick
| Week | Game | LINE pick | Predicted | Final | Result | Week log |
|---|---|---|---|---|---|---|
| Week 9 | BUF at MIN | BUF 52.6% | 23–22 | 23–6 | Correct | 1546cab17d18 |
| Week 9 | TB at CHI | CHI 57.9% | 25–28 | 24–27 | Correct | 1546cab17d18 |
| Week 8 | CHI at SEA | SEA 58.2% | 24–27 | 34–17 | Missed | 2924e2c5b4a8 |
| Week 8 | PHI at WAS | WAS 63.4% | 18–23 | 10–27 | Correct | 2924e2c5b4a8 |
| Week 8 | KC at DEN | DEN 54.5% | 20–21 | 24–28 | Correct | 2924e2c5b4a8 |
| Week 8 | NE at MIA | NE 67.1% | 25–19 | 24–19 | Correct | 2924e2c5b4a8 |
| Week 8 | LAC at LAR | LAR 70.4% | 19–26 | 9–24 | Correct | 2924e2c5b4a8 |
| Week 8 | ARI at DAL | DAL 64.8% | 25–30 | 33–30 | Missed | 2924e2c5b4a8 |
| Week 8 | ATL at TB | TB 59.3% | 22–26 | 0–16 | Correct | 2924e2c5b4a8 |
| Week 8 | BAL at BUF | BUF 56.8% | 25–28 | 10–36 | Correct | 2924e2c5b4a8 |
| Week 8 | CLE at PIT | PIT 58.5% | 20–23 | 6–27 | Correct | 2924e2c5b4a8 |
| Week 8 | IND at JAX | JAX 67.9% | 22–28 | 7–42 | Correct | 2924e2c5b4a8 |
| Week 8 | LV at NYJ | NYJ 52.1% | 23–23 | 38–33 | Missed | 2924e2c5b4a8 |
| Week 8 | MIN at DET | DET 56.8% | 21–24 | 15–19 | Correct | 2924e2c5b4a8 |
| Week 8 | TEN at CIN | CIN 75.4% | 18–27 | 36–43 | Correct | 2924e2c5b4a8 |
| Week 8 | CAR at GB | GB 60.8% | 21–24 | 19–27 | Correct | 2924e2c5b4a8 |
Scores are away–home. “Week log” is the start of the SHA-256 fingerprint of the file the week’s official picks were saved in, before kickoff. Every pick in that file shares it; the full fingerprint and timestamps are in the app.
NCAA research record
Research forecasts saved at least 30 minutes before each game and never changed, scored against the data provider’s final. They’re not official LINE predictions and are never combined with the NFL record.
| Sport | Saved before start | Correct | Missed | Accuracy |
|---|---|---|---|---|
| NCAA Football 2026 | 266 | 0 | 0 | — |
| NCAA Basketball 2026-27 | 614 | 0 | 0 | — |
| NCAA Baseball | No licensed data provider yet, so no forecasts and nothing to score. | |||
Counts use finals confirmed in LINE’s data at the time of this update; the app updates as games finish.
Backtests, and where LINE falls short
Before the live record, LINE 2.0 was tested on 1,519 NFL games from 2020 to 2025 using closing lines. It picked the winner 63.6% of the time and was right against the spread 49.9% of the time. The sportsbook favorite picked the winner 66.9% of the time, and its average margin miss was 9.8 points against LINE’s 10.3.
In that test, the market was more accurate than LINE. LINE is useful as a second, transparent opinion and a way to see where the numbers disagree. It is not a proven way to beat the sportsbook.
- Small sample. The live record is 16 games so far. Short runs, good or bad, say little about the long run.
- Against the spread is roughly a coin flip. 49.9% in the backtest, below the roughly 52.4% needed to cover standard sportsbook fees. Picking winners isn’t the same as making money.
- Backtests flatter. The 2020–2025 test used closing lines, which know more than the lines available when a pick is locked, so live results are usually worse. Backtests are never counted in the official record.
- Late news. Injuries or quarterback changes after a pick is locked aren’t in that pick.
- College is research. NCAA models are newer and less tested, which is why they’re labeled research.