SharpEdge

The Ledger

One win, three losses. Below is each play graded this week, the DraftKings number we took and where the market closed.

Four published plays haven't been graded. They are logged as "none" with no game attached (weeks 3, 4, 4 and 5), so there is nothing to grade.

What the scoreboard terms mean:

• Spread: The head start the underdog gets. Seahawks -3 means Seattle has to win by more than 3 for that bet to cash.
• Moneyline: A bet on who wins, no spread. -166 means risk 166 to win 100; +142 means risk 100 to win 142.
• Units: A fixed bet size so results can be compared. One unit is one unit whether that is $10 or $100.
• Hit rate: The percentage of bets that won. At standard -110 odds you need 52.4% just to break even.
• Closing line value: CLV: whether the line you bet was better than where the market closed. The best long-term sign a bettor knows what they are doing, because closing lines are the sharpest numbers there are.

Break-even is 52.4%. We're at 27.8%. That isn't a cold streak sitting just under break-even. It's a long way under it, and -8.45 units is the bill. An average CLV of +0.21 means our numbers have been 0.21 points better than the close, on average. That's the right direction, but the margin is thin.

What the Closing Line Said

Three of the four plays have a closing number. One beat it, two matched it and none lost to it.

Titans +7.5 beat the close by a full point. DraftKings closed at +6.5, and that move went through 7, one of the key numbers: NFL games land on 3 and 7 far more than any other margin, so a half point on either side of those numbers is worth real money. We held the hook, the extra half point, above 7, and the market had taken it away by kickoff. It was our best number of the week, and it lost.

Giants +3.5 and Browns +2.5 closed exactly where we bet them. That's zero CLV. The market didn't say we were wrong, and it didn't say we were early.

The Colts moneyline has no closing price on record. That's a hole in our bookkeeping, and it's ours. From here on, every play gets a closing number.

Why this column matters more than wins and losses: a single football game is full of noise. A tipped pass, a missed kick or a backdoor touchdown can make a good bet lose. The closing line is where every sharp dollar and every piece of news has landed by kickoff. If you beat it consistently, you're seeing things before the market does, and over enough bets the wins follow. Now the honest part. Eighteen graded bets is a small sample. An average CLV of +0.21 points the right way, but it isn't close to enough to call 5-13 bad luck.

The Lean, Revisited

Titans +7.5 (Play), loss: the number was right, the result wasn't. We bet the number here, not the team. Tennessee is last in our power ratings at -6.4. Power rating: A number for how good a team is, in points, against an average team on a neutral field. Subtract two ratings, add home field, and you have a model spread. Houston rates 2.0, and the Titans score 11.4 points a game. Our bet was that 7.5 was too many points to give them. The market agreed and moved to 6.5, and Houston still won by more than 7.5. Verdict: variance on a good number. At the same price I'd make the same bet. What would prove me wrong: if bets like this, a good number on a bad team, keep beating the close and still losing, then our ratings are too kind to the bottom of the league.

Browns +2.5 (Lean), loss: the read, on a weak number. Zero CLV on a Lean means we had an opinion the market didn't share. Start with the number. For an underdog, +2.5 is the wrong side of 3. If a game lands on 3, +2.5 loses and +3 pushes. Our ledger records results, not margins, so I can't tell you whether this game landed there. Our data showed DraftKings and no second book, so line shopping wasn't possible. Line shopping: Checking the same bet at several sportsbooks and taking the best number. Getting +3.5 instead of +3 is free equity, and it costs nothing but a minute. Our own backtest says that's where repeatable edge comes from. Then the read. Cleveland quarterbacks get sacked on 11.2% of dropbacks, and the Jets defense gets a sack on 9.6% of opponent dropbacks. A power rating doesn't price that matchup directly. I can't say it decided the game, but I'd weigh it now. Verdict: the read. A lean the market never moved toward isn't an edge.

Colts moneyline +124 (Lean), loss: probably the read. At +124 the Colts were the underdog. Underdogs lose more often than they win, and the plus price is what pays you for taking that risk. With the result included, our ratings have Pittsburgh at -0.2 and Indianapolis at -0.6, and Pittsburgh was at home. That isn't a case for the Colts winning outright. With no closing price, I can't grade the number. Verdict: the read, plus a process failure.

Giants +3.5 (Play), win. Ratings of -1.5 for the Giants and -2.8 for Washington support the side. The flat close says the market had the same read, so this was a fair bet that cashed, not proof of an edge.

Now the bigger picture. The model's backtest, run on 2023-2025 games it wasn't built on, went 51.7% against closing spreads over 437 plays where it disagreed with the market by 2 or more points. Gaps of 2-3 points hit 54.7%, gaps of 3-4.5 hit 52.1%, and gaps of 4.5 or more hit 47.7%. The market is sharp. The model shows us where to look, and when it disagrees with the market by a lot, the model is usually the one that's wrong. A 27.8% hit rate is far below what that backtest suggests. Some of it is small-sample noise and some of it is plays we shouldn't have made. Eighteen bets can't tell me how much is which.

Looking Ahead

These are next week's games where DraftKings opened on 3 or 7, or a half point off one. The "Model" column is our number for the home team. No team on this slate has an injury designation listed yet.

Seahawks at Broncos, Thursday. The model has Seattle by just 0.6, and DraftKings has already moved from Seattle -3 to -2.5. This is the Cleveland lesson again: an underdog at +2.5 is on the wrong side of 3. Denver +3 was the number to bet. At +2.5, I pass. The projected starters are Drew Lock for Seattle and Bo Nix for Denver.

Bears at Falcons. This is the biggest gap on a key number. The model makes Atlanta a 0.2-point favorite, and DraftKings has Chicago -3, a 3.2-point gap toward Atlanta. Atlanta +3 is the right side of 3, but gaps that size hit 52.1% in the backtest, which is under break-even. I lean Atlanta, and it only becomes a play at +3.5. Tyson Bagent is the projected Chicago starter.

Steelers at Buccaneers. Same setup, bigger gap. The model makes Tampa a 1.2-point favorite, and DraftKings has Pittsburgh -3. That's a 4.2-point gap, and Tampa comes in on 10 days of rest to Pittsburgh's 7. But 4.2 is right next to the 4.5-or-more range, which hit 47.7%. When the model disagrees that loudly, it's usually the model that's wrong. I lean Tampa, but there's no play at +3.

Texans at Jaguars, Wembley Stadium. Jacksonville tops our ratings at 5.0. The model says -4.8 and DraftKings says -3. A 1.8-point gap is under the 2-point minimum the backtest used, so there's no play. If you're on Jacksonville anyway, -3 is the number to own, because a 3-point win pushes. At -3.5, I'm out.

Ravens at Browns. Baltimore opened -3.5, a hook off 3. It's now -4.5, away from the key numbers. The model has Cleveland +2.8, and the market moved a full point the other way. Pass.

The 7s. Indianapolis -7 against a model -7.6 is nothing, and I'm not chasing Tennessee. New England opened -7 and now sits at -7.5. A 1.1-point gap isn't worth paying for the hook. San Francisco -7.5 against a model -8.5 on Monday night gets the same answer.

Ledger Entries

No new plays in the ledger issue.

Plain English

• Spread: The head start the underdog gets. Seahawks -3 means Seattle has to win by more than 3 for that bet to cash.
• Moneyline: A bet on who wins, no spread. -166 means risk 166 to win 100; +142 means risk 100 to win 142.
• Units: A fixed bet size so results can be compared. One unit is one unit whether that is $10 or $100.
• Hit rate: The percentage of bets that won. At standard -110 odds you need 52.4% just to break even.
• Closing line value: CLV: whether the line you bet was better than where the market closed. The best long-term sign a bettor knows what they are doing, because closing lines are the sharpest numbers there are.
• Key numbers: NFL games land on 3 and 7 far more than any other margin, so a half point on either side of those numbers is worth real money.
• Line shopping: Checking the same bet at several sportsbooks and taking the best number. Getting +3.5 instead of +3 is free equity, and it costs nothing but a minute.
• Power rating: A number for how good a team is, in points, against an average team on a neutral field. Subtract two ratings, add home field, and you have a model spread.

Sources

DraftKings lines via ESPN (numbers taken, closing numbers, openers and current lines), nflverse play-by-play (team profiles, power ratings and model numbers), official injury reports, and the SharpEdge graded ledger.

21+ only. If you or someone you know has a gambling problem, call 1-800-GAMBLER.