College Football
Book Conviction SpreadsWhat Book Conviction measures — feeds, filters and scoring
A book that holds most of the money on one side should move its number toward that side. When it does the opposite — shading the price to make the popular side cheaper while already carrying that exposure — it is telling you its own number disagrees with its customers. That is what Book Conviction measures: how hard a book is leaning against its own money, priced in the margins that actually decide games.
- Action Network — consensus % of money and % of tickets, plus a timestamped history of both the number and the price, at seven books individually rather than only the consensus.
- DraftKings — DraftKings' own handle % and bet %, independent of the consensus above, which is what makes agreement between them worth anything.
- Pinnacle — the sharp reference price. It takes sharp action and moves on it instead of shading to balance a public book's exposure, so it answers a different question: not where the money is, but where the number should be.
Official picks — 2.5+
No official pick today. 1 spread cleared both filters but none reached 2.5 conviction, so none of them count towards the record. They are listed below.
Also flagged — below 2.5
Real signals that did not reach the official line. Listed and graded, but they do not count towards the headline record.
North Texas Mean Green at Navy Midshipmen
UNT holds 84% of the money (Action Network) and 83% at DraftKings, yet the number moved 0.5 pts the other way. The book is pricing UNT -3 cheaper while already holding most of that side's money.
Board changes
Plays that moved on or off the board between the first time they fired and kickoff. The board is re-read every hour, and a play that fell short of the filters by its last pre-kickoff reading is not a pick: it is shown here with how it would have done, but not counted in any record.
| Matchup | Timeline (ET) | Outcome |
|---|---|---|
| WVU @ TCU | 8:00 AM Flagged TCU -10.5 Off before kickoff (time not recorded) | TCU -10.5 loss not counted |
| BYU @ UCF | 8:00 AM Flagged UCF +7.5 Off before kickoff (time not recorded) | UCF +7.5 win not counted |
| OU @ MSST | 8:00 AM Flagged MSST +2.5 Off before kickoff (time not recorded) | MSST +2.5 loss not counted |
| KSU @ ASU | 11:15 AM Flagged ASU +1.5 Off before kickoff (time not recorded) | ASU +1.5 win not counted |
Results — spreads
| Matchup | Market | Pick | CLV | Final | Result |
|---|---|---|---|---|---|
| UNT @ NAVY | spread | NAVY +3 | +0.29 (+0 line / +0.29 price) | 49-39 | loss |
Graded against the number frozen before kickoff, not a later one. This slate: 0W - 1L (0%) · 1 of 1 beat the close, averaging +0.29 points.
By conviction
| Band | Conviction | Picks | Record | Beat the close |
|---|---|---|---|---|
| High | 5.5 + | 15 | 9W - 6L (60%) | 10/15 (+0.83) |
| Medium | 3.5 – 5.5 | 10 | 6W - 4L (60%) | 2/10 (+0.05) |
| Low | under 3.5 | 16 | 10W - 6L (62%) | 2/16 (+0.02) |
6 earlier picks predate the score and sit in no band. They were not weak, they were unscored — the score did not exist yet when they fired, and a pick cannot be sorted by a number nobody gave it.
These bands are the test, and they are not the board. Closing line value is readable within hours, so if conviction means anything the 5.5+ band should beat the close more often than the under 3.5 one — an answer available in weeks, where win/loss at this volume would take seasons. The cutoffs sit near the thirds of where the score actually lands, so the three groups are comparable in size; split at the publishing line instead, the top band held 37 picks against 1 and 2, and nothing can be read from a group of one.
Which picks get published is a separate question, and the answer has not changed: 2.5 and above is an official play. Everything below it is still listed and still graded.
It is measured in two halves, because a market moves two ways. Line is the number itself. Price is the juice on that number, converted to points at the same rate the screen uses to read a price move as a line move — roughly 4.5 points of implied probability to a point of spread. A pick taken at −3 −110 that closes −3 −125 cost the book a dime, and counting only the line would score that as nothing at all.
The close is DraftKings' actual closing number, and both ends of the comparison are read from the same book — the standing gap between two books is not closing line value. Entries marked * are measured against this screen's own last pre-kickoff reading instead, which is a proxy for the close rather than the close. Picks recorded before this was tracked show “—”.
Full board — spreads
| Matchup | ET | Bets | Open | Now | Money on | AN $% | AN T% | DK $% | DK T% | 7d | 3d | 1d | Key | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| MISS @ TEX | 7:30 | 71,040 | -8.5 | -3 | MISS | 80 | 82 | 49 | 65 | +3.5 | +0.5 | 0 | — | FINAL |
| TENN @ SC | 3:30 | 62,664 | 2.5 | -2.5 | TENN | 53 | 77 | 45 | 69 | -1 | 0 | 0 | — | FINAL |
| IU @ MICH | 12:00 | 50,684 | 6 | 7.5 | IU | 60 | 66 | 47 | 58 | -1.5 | +0.5 | 0 | — | FINAL |
| HOU @ UTAH | 10:15 | 49,744 | -6.5 | -6.5 | HOU | 75 | 74 | 47 | 40 | 0 | 0 | 0 | — | FINAL |
| BYU @ UCF | 3:30 | 49,118 | 9.5 | 7.5 | BYU | 64 | 78 | 47 | 64 | 0 | +1 | 0 | — | FINAL |
| LSU @ AUB | 12:00 | 41,813 | 6.5 | 7 | AUB | 55 | 54 | 26 | 46 | +1 | -0.5 | 0 | — | FINAL |
| USC @ WIS | 7:30 | 37,279 | 1.5 | 2.5 | WIS | 50 | 35 | 50 | 41 | +1 | -0.5 | -0.5 | — | FINAL |
| IOWA @ MINN | 3:30 | 34,810 | 2.5 | 7.5 | IOWA | 69 | 69 | 34 | 51 | +3.5 | +1 | 0 | — | FINAL |
| TTU @ CIN | 12:00 | 32,867 | 11.5 | 10.5 | CIN | 54 | 41 | 45 | 44 | -1 | -0.5 | -1 | — | FINAL |
| TA&M @ BAMA | 3:30 | 27,651 | -2.5 | -13.5 | BAMA | 69 | 64 | 88 | 79 | +2 | 0 | 0 | — | FINAL |
| VAN @ UK | 7:30 | 27,606 | -9.5 | -10.5 | UK | 66 | 68 | 53 | 71 | -1 | 0 | 0 | — | FINAL |
| MSU @ UCLA | 3:30 | 25,036 | -13.5 | -16.5 | UCLA | 55 | 56 | 82 | 78 | -3 | -2.5 | -2 | — | FINAL |
| BAY @ KU | 1:00 | 24,390 | 2.5 | -1.5 | BAY | 57 | 69 | 69 | 57 | -4 | -4 | -4 | 2 | FINAL |
| WVU @ TCU | 12:00 | 24,248 | -6.5 | -10 | WVU | 74 | 77 | 46 | 54 | -3.5 | -0.5 | +0.5 | 10 | FINAL |
| PITT @ MIA | 12:00 | 24,218 | -23.5 | -20.5 | MIA | 55 | 50 | 94 | 77 | +3 | -0.5 | -0.5 | — | FINAL |
| ORE @ ILL | 3:00 | 23,728 | 13.5 | 10.5 | ORE | 56 | 56 | 27 | 47 | -3 | +1 | +0.5 | — | FINAL |
| NDSU @ UNM | 10:00 | 21,744 | -3.5 | -4.5 | UNM | 56 | 41 | 44 | 68 | -0.5 | 0 | 0 | — | FINAL |
| OU @ MSST | 3:30 | 20,846 | 1.5 | 1.5 | OU | 56 | 56 | 67 | 60 | 0 | -3 | -2 | 2, 3, 4 | FINAL |
| VT @ CLEM | 7:00 | 20,139 | 1.5 | -3 | CLEM | 57 | 54 | 47 | 57 | -4.5 | -0.5 | -0.5 | — | FINAL |
| COLO @ OKST | 7:00 | 19,957 | -15.5 | -13.5 | OKST | 55 | 59 | 62 | 74 | +2 | 0 | 0 | — | FINAL |
| ISU @ ARI | 7:30 | 19,735 | -10 | -11.5 | ISU | 51 | 55 | 47 | 48 | -1.5 | -1.5 | -2 | 10 | FINAL |
| RICE @ FAU | 12:00 | 16,514 | -10.5 | -9.5 | FAU | 66 | 68 | 67 | 50 | +1 | -2 | -1.5 | — | FINAL |
| UNT @ NAVY | 3:30 | 15,560 | 1.5 | 3 | UNT | 84 | 74 | 83 | 75 | +1.5 | +0.5 | 0 | — | 1.6 FINAL |
| SYR @ UNC | 7:30 | 14,565 | -9.5 | -9.5 | SYR | 59 | 44 | 24 | 30 | 0 | -2 | 0 | 8 | FINAL |
| BOISE @ WSU | 6:00 | 13,828 | 16.5 | 14.5 | BOISE | 86 | 77 | 59 | 61 | -2 | +1.5 | +1 | — | FINAL |
| HAW @ NIU | 3:30 | 13,687 | 7.5 | 3 | NIU | 76 | 63 | 89 | 72 | -4.5 | -1 | -1 | — | FINAL |
| RUT @ NW | 12:00 | 13,464 | -16.5 | -15 | RUT | 60 | 45 | 32 | 36 | +1.5 | +1.5 | +1.5 | — | FINAL |
| KSU @ ASU | 3:45 | 12,515 | 2.5 | 1.5 | KSU | 64 | 65 | 65 | 72 | -1 | -2 | -1 | 2, 3 | FINAL |
| CAL @ SMU | 12:00 | 12,500 | -16.5 | -18.5 | CAL | 57 | 58 | 27 | 38 | -2 | +1 | 0 | — | FINAL |
| SDSU @ CSU | 6:30 | 12,000 | 1.5 | -1 | CSU | 52 | 50 | 98 | 72 | -2.5 | -2.5 | -2.5 | — | FINAL |
| USU @ TXST | 7:30 | 9,653 | -9.5 | -13.5 | USU | 55 | 50 | 10 | 13 | -4 | -0.5 | 0 | — | FINAL |
| WMU @ TOL | 12:00 | 8,608 | -1.5 | -2.5 | TOL | 60 | 59 | 65 | 53 | -1 | +1 | +1 | 3 | FINAL |
| GAST @ ARST | 7:00 | 8,446 | 1.5 | 2.5 | GAST | 77 | 70 | 90 | 59 | +1 | 0 | 0 | — | FINAL |
| BC @ GT | 3:30 | 8,289 | -6.5 | -6.5 | GT | 64 | 60 | 42 | 51 | 0 | -2.5 | -0.5 | — | FINAL |
| M-OH @ CMU | 1:00 | 8,274 | -1.5 | 3.5 | M-OH | 74 | 64 | 68 | 69 | +5 | +1 | 0 | — | FINAL |
Movement columns are the change in the home spread over each window. Teal = moved against the money (the signal); orange = moved with it. AN = Action Network consensus, DK = DraftKings. Percentages are for the side named in “Money on”.
How much to trust it
- Most days produce nothing, and that is the screen working. Thin markets are dropped before they are screened, both money feeds have to agree, and the line has to have moved against that money. A light board clears none of that: one recent Friday had four games and produced no pick at all. A screen that always finds something is not a screen.
- The edge is modest. The underlying move — reverse line movement — is a real but heavily degraded signal — books price it in now. This screen finds where a market is behaving oddly; it does not establish that the odd behaviour is wrong.
- Splits are a snapshot, not a history. Percentages are read at generation time and will drift before kickoff.
- Ticket shape can conflict. Where one feed shows money concentrated in few tickets and the other shows it spread across many, the sharp reading is weaker. The per-game cards show both so the disagreement stays visible.
- Totals move spreads. A falling total mechanically compresses a spread, so part of a spread move can be a total move wearing a disguise.
- Check the news. A line moving against the money is often just an injury the percentages have not caught up with yet.