First Half Betting Strategy

The game log bundled with this site records final scores only — no halftime scores. That means we cannot compute first-half cover rates, 1H key numbers, or scripted-drive success from our own data, and you should treat any site quoting precise 1H percentages without a named source skeptically. The drive-level figures in the table above are presented as a framework from public play-by-play research, not as numbers we verified.

What the bundled data can anchor is the full-game context every 1H bet derives from (7,276 completed games, 1999–2025):

Data: nflverse games.csv bundled with this site (7,276 completed games, 1999–2025), computed by the author.

The derivation rules quoted above (1H spread ≈ half the game spread, 1H total ≈ 46–48% of the game total) are market conventions you can check on any live board, not statistics — books skew the second half slightly higher-scoring because teams trailing late throw more and clock rules differ. Where your 1H opinion is really a full-game opinion at half scale, bet the full game; the 1H market's thinner limits and wider prices only earn their keep when you have a half-specific reason.

Illustrative current-format numbers — game line Chiefs −7, total 47:

• Convention says 1H spread ≈ −7/2 = −3.5. Books often shade to −3.5/−4 because favorites' edges are believed to concentrate early; that belief is exactly what you would need half-score data to test.
• 1H total at 47% of the game total ≈ 47 × 0.47 = 22.1, so boards will show 22 or 22.5. The missing 6% of scoring implicitly belongs to the second half.
• The implied 1H score is therefore about 13–9 — note how a 1H spread of −3.5 sits above the most common half-game margin of 3, the mirror of why full-game lines fight over the 3.

If you take the −3.5 at −110, your break-even is the usual 52.4% (110/210, exact formula math) — but against a market with lower limits, wider spreads between books, and faster repricing. Shop harder here than anywhere: on thin markets the difference between −110 and −120 (break-even 54.5%) is routinely available for the same bet.

• No half-time scores in the bundled data means every 1H-specific rate on this page is unverified framework; we have labeled which is which, and we would rather say so than invent a percentage.
• Scripted-play analysis is drive-level and needs play-by-play data; the logic (preparation shows up early) is sound, the exact percentages require sources we cannot check here.
• 1H markets are derivative. Their prices move with the full-game line, so stale 1H numbers after game-line moves are the realistic opportunity — and the hardest to catch manually.

Keep reading: key numbers (the full-game version of the 3 problem), win probability, and PROE & game script for how early-game tendencies are measured when play data is available.