How the football grades have done · 2026-09-09 – 2026-11-16
The registered read (declared 2026-09-08, before any game settled). Per market, the A leans: flat-stake return at the price the card showed, the same result at the close, and beat-the-close on the leans whose line moved. First look at 30 settled leans in a market, season-end look after week 18. Verdicts: edge-shaped = return above zero and beat-the-close above 52%; noise = return within five points of zero and beat-the-close 48-52%; wrong = return under -5% or beat-the-close under 48%, and the band comes off on the owner's word; overconfident = the model's mean P(lean side) exceeds the hit rate by more than 8 points, and the cut is re-derived. No cut, feature or blend is tuned on 2026 rows before the first look. Break-even at -110 is 52.4%.
Columns. RETURN is the flat-stake return per $1 at the price the card showed. AT CLOSE is the same result priced at the captured close. BEAT is how often the lean's side got a better number than the close, counted over the leans whose line moved. MODEL is the model's mean chance on the lean's side, to hold against WIN %.
OVERALL
| SEGMENT | N | W | L | PUSH | WIN % | RETURN | AT CLOSE | BEAT | MODEL | NOTES |
|---|
| Every settled grade | 450 | 241 | 202 | 7 | 54.4% | +1.2% | +1.4% | 48% of 123 | 55.1% | |
| of which graded — a lean | 15 | 10 | 5 | 0 | 66.7% | +25.1% | +25.5% | 40% of 5 | 56.9% | insufficient sample |
| of which no lean — coverage | 435 | 231 | 197 | 7 | 54.0% | +0.4% | +0.6% | 48% of 118 | 55.0% | |
The two indented rows sum to the first. Graded is the pick record — every card where the model claimed an edge. No lean is every card where it agreed with the market; settlement still writes a W or an L against its nominal side, which is bookkeeping, not a call.
BY MARKET AND BAND· the pick record, crossed · the verdict reads the A leans
| SEGMENT | N | W | L | PUSH | WIN % | RETURN | AT CLOSE | BEAT | MODEL | NOTES |
|---|
| Spread · A | 7 | 6 | 1 | 0 | 85.7% | +65.9% | +66.4% | — | 54.7% | first look at 30 · 7 so far |
| Spread · all graded picks | 7 | 6 | 1 | 0 | 85.7% | +65.9% | +66.4% | — | 54.7% | insufficient sample |
| Moneyline · A | 8 | 4 | 4 | 0 | 50.0% | -10.6% | -10.2% | 40% of 5 | 58.7% | first look at 30 · 8 so far |
| Moneyline · all graded picks | 8 | 4 | 4 | 0 | 50.0% | -10.6% | -10.2% | 40% of 5 | 58.7% | insufficient sample |
Verdicts. Spread A: first look at 30 · 7 so far · Moneyline A: first look at 30 · 8 so far
NO LEAN — COVERAGE· the market's right; not a track record
| SEGMENT | N | W | L | PUSH | WIN % | RETURN | AT CLOSE | BEAT | MODEL | NOTES |
|---|
| Moneyline | 142 | 87 | 54 | 1 | 61.7% | +5.4% | +5.6% | 46% of 98 | 59.2% | |
| Spread | 143 | 68 | 71 | 4 | 48.9% | -4.5% | -4.1% | 67% of 6 | 51.8% | |
| Total | 150 | 76 | 72 | 2 | 51.3% | +0.4% | +0.2% | 57% of 14 | 54.1% | |
BY DAY· the graded leans per slate · open a day for its games
PLAYER PROPS· the projection ledger, settled from the play-by-play
Same discipline, declared the same day. First look at 100 settled leans overall and 30 STRONGEST leans, read separately: return at the lean price, hit rate, and the calibration of every priced projection (the model's mean P(over) against the over rate). Same verdict table. The prior is the one prop lane run before this: MLB strikeouts, 49,559 priced closes, lost to the close. A lean is the projection's leftover after the vig-free price clearing 8 points; STRONGEST is the card's own top three at the last pregame build — firm, one per driver, best expected return — frozen in the ledger so this table reads what the card showed. First look taken 2026-10-18 (weeks 1-6): return above zero, but overconfident — the leans said 67% and hit 50%. MODEL is the projection's mean chance on the lean's side; FAIR is the same chance once the model gets only the say it earned against the book this season (fitted on these rows, so it matches the hit rate by construction on them — the read is whether it keeps matching).
Reading the props tables. RETURN is what $1 on every bet in the row did, at the price taken; it is the number that matters, because a hit rate means nothing without the price. NEEDS is the hit rate those prices need to break even (a +600 touchdown needs about 14%, a -110 needs 52.4%), to hold against HIT %. ± LUCK is how far the return could swing on luck alone (one standard error): long prices swing far more than short ones on the same count, so a few big hits can make a row look better than it has shown. The split tables sort on the return less one swing, and a row turns green only when it is still in profit two swings down.
| SEGMENT | N | HITS | HIT % | NEEDS | RETURN | ± LUCK | MODEL | FAIR | NOTES |
|---|
| Every lean | 3701 | 1778 | 48.0% | 41.2% | -0.4% | ±2.2% | 65.8% | 49.0% | |
| of which STRONGEST — the card's top three | 431 | 228 | 52.9% | 51.3% | +2.9% | ±4.7% | 76.0% | 52.2% | |
| of which firm — no flag at the build | 2124 | 1061 | 50.0% | 52.0% | -4.2% | ±2.1% | 69.4% | 51.7% | |
| of which BEST OF THE BOARD — the sheet's top five on a day of two or more games | 50 | 26 | 52.0% | 51.1% | +0.2% | ±13.9% | 80.2% | 54.1% | |
| of which BEST BY POSITION — the sheet's top three running backs on a day of two or more games | 30 | 16 | 53.3% | 51.5% | +1.9% | ±17.8% | 72.6% | 51.1% | |
| of which BEST BY POSITION — the sheet's top three wide receivers on a day of two or more games | 30 | 19 | 63.3% | 50.3% | +24.2% | ±18.0% | 77.6% | 55.1% | |
| of which BEST BY POSITION — the sheet's top three tight ends on a day of two or more games | 30 | 16 | 53.3% | 52.0% | +1.9% | ±17.8% | 74.4% | 55.4% | |
| of which BEST BY POSITION — the sheet's top three quarterbacks on a day of two or more games | 30 | 14 | 46.7% | 52.8% | -11.8% | ±17.5% | 68.9% | 51.8% | |
| of which the RECEIVER READ — receiving yards since it went live | 167 | 90 | 53.9% | 52.7% | +2.2% | ±7.3% | 65.2% | 54.2% | |
| of which the RECEIVER READ — catches since it went live | 168 | 97 | 57.7% | 51.8% | +10.2% | ±7.5% | 66.7% | 57.5% | |
OVER VS UNDER· the model's leans by side, against every side blind
| SEGMENT | N | HITS | HIT % | NEEDS | RETURN | ± LUCK | MODEL | FAIR | NOTES |
|---|
| Every over lean | 1685 | 807 | 47.9% | 51.2% | -7.0% | ±2.4% | 67.9% | 50.9% | |
| of which STRONGEST | 167 | 77 | 46.1% | 50.9% | -9.9% | ±7.6% | 74.4% | 51.4% | |
| every priced over, no model | 5597 | 2707 | 48.4% | 51.7% | -7.0% | ±1.3% | 49.1% | 49.3% | |
| Every under lean | 1735 | 926 | 53.4% | 52.3% | +1.6% | ±2.3% | 69.9% | 52.6% | |
| of which STRONGEST | 263 | 150 | 57.0% | 51.7% | +10.4% | ±5.9% | 77.2% | 52.8% | |
| every priced under, no model | 5597 | 2890 | 51.6% | 52.4% | -2.2% | ±1.3% | 50.8% | 50.7% | |
| Anytime TD (yes only, kept apart) | 281 | 45 | 16.0% | 11.8% | +27.2% | ±20.9% | 27.5% | 15.6% | |
| of which BEST ANYTIME TD — the sheet's top five on a day of two or more games, the one best on a one-game day | 70 | 14 | 20.0% | 18.4% | -28.6% | ±17.4% | 36.5% | 22.7% | |
| of which the TOUCHDOWN READ — tight ends since it went live | 28 | 3 | 10.7% | 12.0% | -14.3% | ±49.3% | 18.1% | 17.0% | insufficient sample |
Is the edge the unders, or the model? The no-model rows back one side of EVERY priced two-sided projection, lean or not, at the best price the build captured. If the model's under leans return no more than every under does blind, the edge is in the side, not the projection. Anytime touchdowns have no under quote, so they sit out of both the over rows and the no-model rows and get a line of their own: a handful of long-priced hits swings their return far more than any other market's.
QB UNDERS AS A GROUP· every priced line, no model
| SEGMENT | N | HITS | HIT % | NEEDS | RETURN | ± LUCK | MODEL | FAIR | NOTES |
|---|
| QB carries under, every line this season | 144 | 91 | 63.2% | 52.9% | +19.2% | ±7.7% | 58.3% | 51.1% | |
| of which the last four weeks | 45 | 34 | 75.6% | 52.8% | +44.2% | ±12.6% | 55.4% | 51.1% | |
| QB rushing yards under, every line this season | 283 | 162 | 57.2% | 52.8% | +8.3% | ±5.6% | 50.5% | 50.7% | |
| of which the last four weeks | 112 | 67 | 59.8% | 52.6% | +13.3% | ±8.8% | 48.7% | 50.3% | |
A trend, not a model read. These rows back the under on every priced quarterback line, lean or not, at the best price the build captured, graded the way the books settle, with kneel-downs as carries. Through week 8 the projection added nothing beyond the line on these markets, so this shows how the lines have run, and the last four weeks show whether that is holding.
BY MARKET AND SIDE· every lean, each over its no-model twin · sorted by return less one swing of luck · green = still in profit two swings down
| SEGMENT | N | HITS | HIT % | NEEDS | RETURN | ± LUCK | MODEL | FAIR | NOTES |
|---|
| anytime TD · yes | 281 | 45 | 16.0% | 11.8% | +27.2% | ±20.9% | 27.5% | 15.6% | |
| rushing yards · under | 275 | 161 | 58.6% | 52.4% | +11.7% | ±5.7% | 69.2% | 50.8% | |
| no model | 839 | 464 | 55.3% | 52.7% | +4.9% | ±3.3% | 48.8% | 50.4% | |
| receptions · under | 462 | 254 | 55.0% | 51.8% | +4.9% | ±4.5% | 70.6% | 54.4% | |
| no model | 1671 | 856 | 51.2% | 52.0% | -3.7% | ±2.3% | 51.0% | 51.3% | |
| carries · under | 258 | 144 | 55.8% | 52.4% | +6.2% | ±5.9% | 73.6% | 50.7% | |
| no model | 535 | 293 | 54.8% | 52.4% | +4.4% | ±4.1% | 54.2% | 50.6% | |
| receptions · over | 532 | 266 | 50.0% | 48.6% | +1.3% | ±4.5% | 64.6% | 50.2% | |
| no model | 1671 | 815 | 48.8% | 49.6% | -3.5% | ±2.5% | 49.0% | 48.7% | |
| receiving yards · over | 546 | 279 | 51.1% | 52.7% | -3.1% | ±4.1% | 67.2% | 52.6% | |
| no model | 1679 | 837 | 49.9% | 52.9% | -5.7% | ±2.3% | 50.8% | 49.7% | |
| receiving yards · under | 439 | 225 | 51.2% | 52.7% | -2.7% | ±4.5% | 68.7% | 53.9% | |
| no model | 1679 | 842 | 50.1% | 52.7% | -4.8% | ±2.3% | 49.2% | 50.3% | |
| completions · under | 111 | 57 | 51.3% | 52.0% | -2.8% | ±9.0% | 67.3% | 53.3% | |
| no model | 291 | 146 | 50.2% | 52.0% | -4.4% | ±5.6% | 54.9% | 51.3% | |
| passing yards · over | 70 | 35 | 50.0% | 52.9% | -5.4% | ±11.4% | 64.8% | 49.6% | |
| no model | 291 | 156 | 53.6% | 53.1% | +1.1% | ±5.5% | 48.9% | 49.7% | |
| pass attempts · under | 114 | 55 | 48.2% | 52.2% | -8.3% | ±8.9% | 68.8% | 50.3% | |
| no model | 291 | 154 | 52.9% | 52.3% | +1.0% | ±5.6% | 54.9% | 50.3% | |
| rushing yards · over | 286 | 127 | 44.4% | 52.2% | -15.1% | ±5.6% | 72.8% | 49.9% | |
| no model | 839 | 375 | 44.7% | 52.6% | -15.1% | ±3.3% | 51.2% | 49.5% | |
| completions · over | 43 | 19 | 44.2% | 51.8% | -15.0% | ±14.8% | 64.4% | 51.4% | |
| no model | 291 | 145 | 49.8% | 52.0% | -4.7% | ±5.6% | 45.1% | 48.7% | |
| carries · over | 156 | 63 | 40.4% | 52.5% | -23.3% | ±7.5% | 75.2% | 49.9% | |
| no model | 535 | 242 | 45.2% | 52.0% | -12.9% | ±4.2% | 45.8% | 49.4% | |
| passing yards · under | 76 | 30 | 39.5% | 52.8% | -25.3% | ±10.7% | 67.9% | 50.3% | |
| no model | 291 | 135 | 46.4% | 52.9% | -12.3% | ±5.5% | 51.1% | 50.3% | |
| pass attempts · over | 52 | 18 | 34.6% | 52.5% | -33.9% | ±12.7% | 65.5% | 49.9% | |
| no model | 291 | 137 | 47.1% | 52.5% | -10.7% | ±5.6% | 45.1% | 49.7% | |
PROPS BY MARKET· every lean, split by what was projected · sorted by return less one swing of luck · green = still in profit two swings down
| SEGMENT | N | HITS | HIT % | NEEDS | RETURN | ± LUCK | MODEL | FAIR | NOTES |
|---|
| anytime TD | 281 | 45 | 16.0% | 11.8% | +27.2% | ±20.9% | 27.5% | 15.6% | |
| receptions | 994 | 520 | 52.3% | 50.0% | +3.0% | ±3.2% | 67.4% | 52.2% | |
| receiving yards | 985 | 504 | 51.2% | 52.7% | -2.9% | ±3.0% | 67.9% | 53.2% | |
| rushing yards | 561 | 288 | 51.3% | 52.3% | -2.0% | ±4.0% | 71.0% | 50.4% | |
| carries | 414 | 207 | 50.0% | 52.5% | -4.9% | ±4.7% | 74.2% | 50.4% | |
| completions | 154 | 76 | 49.4% | 52.0% | -6.2% | ±7.7% | 66.5% | 52.8% | |
| passing yards | 146 | 65 | 44.5% | 52.9% | -15.8% | ±7.8% | 66.4% | 49.9% | |
| pass attempts | 166 | 73 | 44.0% | 52.3% | -16.3% | ±7.4% | 67.8% | 50.1% | |
STRONGEST BY MARKET· sorted by return less one swing of luck · green = still in profit two swings down
| SEGMENT | N | HITS | HIT % | NEEDS | RETURN | ± LUCK | MODEL | FAIR | NOTES |
|---|
| carries | 124 | 72 | 58.1% | 52.0% | +11.7% | ±8.6% | 79.6% | 50.1% | |
| receptions | 145 | 78 | 53.8% | 50.0% | +6.8% | ±8.4% | 73.9% | 54.7% | |
| rushing yards | 66 | 31 | 47.0% | 52.2% | -10.4% | ±11.8% | 75.1% | 50.0% | |
| receiving yards | 54 | 23 | 42.6% | 52.6% | -19.1% | ±12.9% | 76.0% | 54.7% | |
| anytime TD | 1 | 1 | 100.0% | 33.9% | +195.0% | — | 44.8% | 33.7% | insufficient sample |
| completions | 15 | 9 | 60.0% | 50.6% | +18.1% | ±25.8% | 70.3% | 52.6% | insufficient sample |
| passing yards | 10 | 6 | 60.0% | 52.8% | +14.5% | ±31.2% | 71.9% | 49.8% | insufficient sample |
| pass attempts | 16 | 8 | 50.0% | 51.7% | -3.9% | ±24.8% | 81.6% | 49.6% | insufficient sample |
Calibration, every priced projection. The model said over 42% of the time on average; the over hit 42% (n 7,638). A gap of more than 8 points in either direction is the overconfident verdict for this lane.