THE BETTING FOOL
By a Man Who Should Know Better
The Chicago Bears scored three points last Sunday. Three. A field goal and then four quarters of standing around in the cold looking at each other. And the machine I built — the one I paid for with a summer of my life and a fair amount of my dignity — has looked at that performance, cleared its throat, and made the Bears the number one play on this week's card. I want you to sit with that a minute before you read another word, because I had to.
First the ledger, because that is the deal here. Week two went four and three on the bets and the bankroll closed at a thousand dollars and fifty-five cents. Fifty-five cents. Two weeks of work, two hundred and seventy dollars pushed across the felt, and I am up the price of a cup of coffee at a place that does not have a name. Across all ten positions it was five and five. The season sits at seven and seven. Nobody is getting rich and nobody is getting wiped out, which if you have ever watched a man bet football is not the worst place to be standing in September.
The wind rule lost one. San Francisco and Miami went forty-eight on a forty-four and a half with the bay blowing eleven miles an hour, and twenty-five dollars walked out the door. That is the first time the only rule I actually believe in has cost me money this year, and I am telling you about it in the second paragraph instead of the twenty-second because that is the difference between a column and a sales pitch. New England and the rain got me even. It is going to lose again. I publish that rule at fifty-four and a half percent, which means it is supposed to lose forty-five and a half times out of a hundred. Anybody selling you a rule that does not lose is selling you something else.
Now. Somebody asked me to stop chasing the market's own number around the block — to quit echoing — and go find something with variance in it, something with rhythm. Fine. I spent two days on it. I am going to tell you exactly what I found, which is nothing, and then I am going to tell you the one thing I tripped over on the way.
The three dead ends. Idea one was pace: two fast teams, lots of plays, take the over. Across twenty-two hundred and fifty-eight games the fastest fifth of the slate went over fifty point nine percent of the time. Fifty point nine. You can get that from a coin and the coin does not charge you a dime of juice. Idea two was variance: a betting line prices the middle of a range, so two teams whose weekly scoring lurches all over the place ought to blow through the tails. The one cell that looked alive — wild offenses, low posted number, take the over — came in at fifty-five point one percent on two hundred and five games, which sounds like something until you remember I looked at nine cells and that is the one that happened to be winning. Idea three was rhythm: fade the team that just exploded, because the market reads last Sunday like scripture. Teams coming off a game more than a standard deviation and a half above their own average went forty-six point nine percent against the spread the following week. Right direction, two hundred and twenty-eight games, and a z of minus nine-tenths. That is not a signal. That is a rumor.
So I sharpened it. If the market overreacts, then the fade should be strongest where the market actually moved — where a team exploded AND is now being laid more points than it usually is. Fifty-one point nine percent. And the mirror, backing the team everybody just gave up on, came in at forty-seven. Backwards. Then the same idea on totals: after both teams blew up, take the under next week. Forty-five point two. Also backwards.
There it is. Rhythm is a thing sportswriters say on television. The market has already read last Sunday, it read it before you did, and it charged you four and a half percent for the privilege of disagreeing with it. Every number in the last two paragraphs is in a file called variance_rhythm.py and you can run it yourself and watch every one of my clever ideas die in alphabetical order.
Now the thing I tripped over. Somewhere around the second dead end I went to check my own inputs and found that the table feeding the model held all of week one and exactly one game of week two. One. The Thursday-nighter. Everything else was still sitting on somebody's server waiting to be published, and my machine, which has no more sense than a parking meter, had cheerfully built a week three card out of a week-old photograph of the league. Chicago's fifty-nine-point opener was in there. Chicago's three-point Sunday was not. I rebuilt the whole 2026 half of the table out of the play-by-play, it now holds thirty-two team-games from week one and thirty from week two — thirty because the Monday-nighter has not landed in my file yet, and I would rather tell you that than round it up — and every number on this card comes from both weeks. That is not a discovery. That is a man finding out his watch stopped. But it is the only thing I found this week that was worth finding, and if I had not gone looking for rhythm I would not have looked at all. It is a script now. It runs every week.
And here is what actually changed. The model reads a team the way a bookkeeper reads a ledger — so many passing yards, so many rushing yards, add it up, here is your team. It never once asks whether the fellow who gained those yards is going to be on the field. This week it does. I took the injury report and the roster file, walked back through every snap of 2026, and worked out what each absent man actually produced and what the fellow behind him is likely to do with the same work. Then I subtracted the difference, in points, before the model was allowed to open its mouth.
It is a small number and I am not going to oversell it. Baltimore loses one point and two-tenths because Zay Flowers pulled a hamstring and a hundred and fifty receiving yards in an afternoon does not get replaced by the next man on the depth chart, it gets replaced by most of the next man. The Jets lose a quarter of a point. Houston lose a tenth. Nobody loses a touchdown. And every deduction is tagged with what it rests on: HARD means a man is on injured reserve, which is a fact and not a forecast, because the rules keep him there four games. FIRM means he was ruled out last week and did not take a snap. SOFT means he was listed, played anyway, and I deducted nothing.
Which brings me to the honest hole in all of it, and I would rather you hear it from me. There is no week three injury report yet. It comes out Wednesday. Everything above is built on what last week's report said and what last week's roster showed, and a man can be cleared on a Thursday. The biggest single name on that soft list is Joe Burrow, who was questionable with a bad back, practiced in full, played, and has thrown for two hundred and thirty a game. If he sits Sunday, Cincinnati comes off my board by nearly nine points and the Pittsburgh game is a different animal entirely. I am not on that game. That is why.
Chicago. All right. The three points. Here is the case, and you can throw it out the window if you like. The Bears have run for 212.5 yards a game, the most of anybody on this slate and twenty-six and a half more than Kansas City behind them, split between D'Andre Swift and Kyle Monangai, and Caleb Williams has run for another hundred and seven himself. That is a ground game with three sets of legs in it. They put five hundred and sixty yards on Carolina in week one and scored fifty-nine. Then Minnesota came to Soldier Field and the whole thing seized up, nine to three, the kind of game where both punters get standing ovations. Two games, one a fireworks show and one a wake. The model averages them and gets a good team. The market looked at the same two games and made Philadelphia a three-point favorite on the road.
The one thing I can add that the arithmetic cannot is this: when the injury report came out last week, Chicago did not have a single man carrying a game-status designation. Not one questionable, not one doubtful, nobody out. Ten Bears were listed with something bothering them and every last one of them was cleared to play. Two clubs in football managed that, Chicago and Jacksonville. Philadelphia, meanwhile, lost Jonathan Greenard off the edge — torn-up chest, ruled out. So I have a healthy team with the best ground game on the board, getting three points at home on a Monday night, against a club that has won two games by a combined six points. Fifty-two point nine percent to cover, and it needs fifty point five. That is a real edge and it is also a small one, and I have one Sunday of evidence that says I could be a complete fool about the Bears. Fifteen dollars. Same as everything else.
Something new on the board. Player props, eight of them, and I want to be very clear about what they are and are not. I have no feed of posted prop numbers. None. Every side and total on this card gets checked against a real price at a real window, and anything that does not clear gets labeled a lean and five dollars. I cannot do that with a prop because there is nothing to check against. So what you get is the Fool's number — what I think the man does — and beside it the flip, which is the line where the play stops being a play. Take the over if your book hangs it at the flip or under. Do not take it if they hang it higher. Three dollars apiece, less than a lean, because a bet nobody priced is not really a bet. They will be graded in public at the flip like everything else, and if they are garbage you will watch them be garbage.
Twenty positions, a hundred and seventy-four dollars out of a thousand. Three weather unders at twenty-five, three sides at fifteen, two sides and four totals at five that do not clear their price and say so, and eight props at three. The wind is up at Hard Rock, at EverBank and in Landover, and it is up hardest of all at MetLife — twelve point six miles an hour for Tennessee and the Giants, the windiest game on the slate, and nobody has posted a line on it, so there is nothing for me to bet. It is not raining anywhere. And my best idea is a football team that scored three points.
Five sides, every week, whether or not any of them is a bet worth making — because a pool makes you pick five. Each one is tagged with what is actually behind it.
Chicago plus three at home, Monday night, and I have already told you what is wrong with this play: they scored three points on Sunday. Here is what is right with it. The Bears run for 212.5 yards a game, the most of anyone playing this week, split between D'Andre Swift's thirty-four carries and Kyle Monangai's twenty, and Caleb Williams has run for another hundred and seven himself in two games. That is a ground game with three sets of legs in it. Philadelphia has won twice by a combined six points and lost Jonathan Greenard, their edge rusher, to a pectoral — ruled out, practiced limited. And when the injury report came out last week Chicago had no man carrying a game status — ten Bears listed with something sore, all ten cleared. Two clubs in football managed that. The regression has it Bears 27, Eagles 21, a six-point Chicago win against a line that says Philadelphia by three — nine points of disagreement, the widest on the card. At even-ish money it needs 50.5 percent and the logistic gives it 52.9. It is the best number I have and it is attached to the team that scored three points. Both things are true. Fifteen dollars.
Indianapolis plus two and a half at home and this one is about who Houston is missing. Nico Collins is out with a hamstring — ruled out last week, did not practice, did not take a snap — and he was carrying eleven and a half percent of the Texans' targets before he went down. What is left is Dalton Schultz, a tight end, at twenty-five percent, and a collection of men named Hutchinson, Wayne and Boutte. Houston also lost Jadeveon Clowney off the edge and Ed Ingram off the guard spot: four men ruled out, which ties the Jets for the most in football. C.J. Stroud threw the ball fifty-nine times against Cincinnati and Houston scored six points, which is what happens when a passing game has no outside receiver and everybody in the building knows it. The Texans are throwing for 314 a game and scoring eighteen and a half. Indianapolis has Jonathan Taylor and forty-three carries in two weeks. The model says Texans 24, Colts 23 — eight-tenths of a point — against a line that says two and a half. Needs 51.2, gets 52.2. Fifteen dollars.
Tampa Bay plus a point and a half at home, and this is the personnel play of the week. Minnesota put Jordan Mason on injured reserve — that is not a designation, that is the rules, four games minimum — and Mason was taking twenty-four percent of the Vikings' carries. Every one of those goes to Aaron Jones, who is thirty-one years old and already carrying it seventeen and a half times a game. Minnesota also has Carson Wentz under center because Kyler Murray has a concussion, and Jauan Jennings is out on a personal matter, and the whole operation is throwing for a hundred and forty-seven yards a game, the least of anybody on this card but Atlanta. The Bucs run for a hundred and thirteen of their own behind Bucky Irving and have lost two games by a combined ten points. The regression makes it Buccaneers 22, Vikings 19 — Tampa by three where the market has them getting a point and a half. It needs 52.4 and it gets 52.4 and it clears by seven hundredths of a percentage point, which is to say by nothing at all. Fifteen dollars because a pool needs five sides. Do not mistake it for conviction.
The Jets plus six and a half in Detroit and I am labeling this a lean before you ask, because it does not clear its price — 52.1 against a 52.4 break-even, three-tenths short at minus one-ten. It is on the card because a pick-five pool makes you pick five and because the number is close enough to be interesting. What I do not like: New York has four men ruled out, tied with Houston for the most in football, including Minkah Fitzpatrick in the secondary with a groin and Omar Cooper off the receiver depth. What I do like: Detroit is giving six and a half with two offensive linemen out — Christian Mahogany's hip and Blake Miller's knee, both did not practice — and Jahmyr Gibbs runs behind those men. The model has Lions 25, Jets 19, which is exactly six, which is exactly the kind of number that makes a six-and-a-half-point line a coin flip with juice on it. Five dollars. It is a pool pick, not a bet, and it says so.
Dallas plus three against Baltimore, and before anything else: this game is not in Dallas. It is in Rio de Janeiro, at the Maracanã, a neutral field forty-eight hundred miles from Dallas and forty-seven hundred from Baltimore, in a city an hour ahead of one of them and two ahead of the other. Nobody is the home team, nobody sleeps in his own bed, and the crowd belongs to whoever sold more jerseys in Brazil. I mention it first because every instinct you have about a home dog is wrong here and so was my own first draft of this paragraph. Now the football. Baltimore comes in with seven men on the injury report, tied with Arizona for the most in the league. Zay Flowers is doubtful with a hamstring and he had a hundred and fifty receiving yards in the opener. Nnamdi Madubuike is out. Ronnie Stanley's toe and John Simpson's groin are both questionable, and those two are the line Derrick Henry runs behind. Against that, Dallas has two defensive starters out — Malik Hooker and DeMarvion Overshown — which is most of why this total is 52.5, the highest number on the slate. My model has it Cowboys 25, Ravens 23 after the Flowers deduction: Dallas by two against a line of Baltimore by three, a five-point disagreement, the widest of any side on the card but Chicago. And it still does not clear, 52.5 against a 52.8 break-even at minus one-twelve. Three-tenths short. Five dollars, labeled a lean, and if you are in a pool this is the one I would think about twice.
Under forty-five and a half in Miami, and this is a weather play, which means it is the only kind of play on this card with real evidence behind it. Eleven and a half miles an hour of wind forecast across the game window at Hard Rock, over the ten-mile threshold, and outdoor games in that wind have gone under 57.7 percent across 603 games since 2016. Discounted to 54.5 for forecast error. Needs 51.2 at minus one-oh-five, which means it clears by 3.3 points of probability, the biggest margin over a price of anything on this card. The game script agrees for once: Kansas City is laying eleven and a half, and a team that far ahead in a wind hands the ball to Kenneth Walker and goes home. The model happens to agree too — 44.9 against a 45.5 line — but I want to say plainly that its agreement is decoration. The totals arm of this thing grades at 49.5 percent over nine seasons and I would be playing this under if it screamed over. Twenty-five dollars.
Under forty and a half in Landover, ten point three miles an hour of wind, and here is the part where I overrule my own machine in public for the third week running. My model projects 46.6 against a 40.5 line. That is six points of disagreement — the widest gap between my number and the market's on any total this week — pointing straight at the over, and I am taking the under anyway, because a 49.5-percent totals model does not get to veto a 57.7-percent weather rule. It never will. Now the football: Seattle is starting Drew Lock because Sam Darnold is out with a glute injury — ruled out, did not practice — and Lock has thrown fifty passes for 422 yards across a start and a relief appearance. Washington is without Chig Okonkwo at tight end and Frankie Luvu on defense, and is running for a hundred and fifty-seven yards a game, which is a club that has decided what it is. Forty and a half is already the lowest number on the board, which tells you the market has done this arithmetic too. Minus one-eighteen is an ugly price and it needs 54.1; I get 54.5. It clears by four-tenths of a point and not an inch more. Twenty-five dollars.
Under forty-five and a half in Jacksonville, ten point one miles an hour, and it clears the wind threshold by a tenth of a mile an hour, which is exactly the kind of margin that should make you check my arithmetic — so the card now prints the forecast to a tenth instead of rounding it to a friendly whole number. The model is at 45.9, over the line again, and again it does not get a vote. New England has A.J. Brown on injured reserve — the hard tier, four games minimum — along with Carlton Davis and a tackle. In fairness Brown had only four targets before he went down, so this is not a team losing a star so much as a team that did not have one: the Patriots are scoring fifteen points a game — not the worst in the league, five clubs are below them, but well under what it takes to get to forty-five and a half by yourself — with Drake Maye throwing to Hunter Henry and a committee. Jacksonville is the other club that put no man on a game status last week. A healthy home team in a stiff breeze against a road offense scoring fifteen is the shape of an under whatever the projection says. Needs 51.9 at minus one-oh-eight, gets 54.5. Twenty-five dollars.
Under forty-two and a half in the Vikings-Bucs game, and it is a lean because it does not clear — 48.9 against a 50.5 break-even, a point and a half short. Printed anyway because it is the same story as the side. Carson Wentz at quarterback, Jordan Mason on injured reserve, Aaron Jones about to carry it twenty-four times, and Tampa Bay grinding it out with Bucky Irving. Two clubs that want to run, and Minnesota throwing for a hundred and forty-seven a game whether it wants to or not. The model has it at 40.5, two points under. The totals logistic turns two points of disagreement into 48.9 percent, which is below a coin flip, because that curve starts below a coin flip — the totals arm has graded 49.5 percent for nine years and the math does not flatter it. Five dollars, a lean, and an honest note that if the model were any good at totals this would be a real bet.
Under forty-three and a half in the Superdome, indoors, no weather, and a four-and-seven-tenths-point disagreement — the model says 38.8. That is the widest totals gap on this card among the games where the weather is not doing the talking, and it still comes up a point and seven-tenths short of its price at minus one-oh-eight. Which is the whole lesson of this operation in one line: a big number is not a good bet, because the curve that turns disagreement into probability is nearly flat and the vig is not. Las Vegas has Ashton Jeanty carrying it twenty-two times a game for seventy-five yards, which is a lot of work for not much. New Orleans throws for three hundred and thirty-one a game, the most on this slate, and has Chris Olave, who was questionable with a hamstring last week, played anyway, and is at a hundred and thirty-four yards a game. If Olave sits on Wednesday's report this number gets more interesting. Five dollars for now.
Under forty-three and a half at Lambeau on Thursday night, four miles an hour of wind, nothing in the sky. The model has it at 42.0, a point and a half under, and turns that into 48.7 percent, which does not clear 50.5. A lean and a small one. The reason it is on the board at all is Atlanta, who are scoring eight points a game and throwing for a hundred and forty-four, both the worst figures on this card by a distance that is not close. Michael Penix is out with a knee and Tua Tagovailoa was doubtful with an oblique, so Cooper Rush has started both games — and last week he threw eighteen passes before Jackson Strand came in and threw seventeen more. A club that needs two quarterbacks to get through a Sunday and scores eight points doing it is not going to solve Green Bay in late September. Five dollars.
Over forty-four and a half in the Monday nighter, and it is the only over on the card, and it is a lean. The model has 48.3, nearly four points over, and turns that into 49.8 percent against a 51.9 break-even — two points short. The case is that Chicago has scored 62 points in two games and gained eight hundred and eighty-six yards doing it, Philadelphia has Saquon Barkley and Jalen Hurts, and both clubs are basically healthy. The case against is that this is the same Chicago that managed three points nine days ago, and that an over on a team capable of scoring three is an over with a trapdoor in it. Five dollars, printed because I am also on the Bears side and you deserve to see the whole picture rather than the flattering corner of it.
This is the cleanest personnel play on the card and it has nothing to do with my regression. Jordan Mason is on injured reserve. Not questionable, not doubtful — on the list, four games minimum, he is not walking through that door. Mason was taking twenty-four percent of Minnesota's carries, fifteen of them in two weeks. Aaron Jones already has thirty-five. Add Mason's share to what Jones is doing and you get 24 carries. Multiply those by three point eight nine a carry — which is halfway between Jones's own four point one four and Minnesota's team rate of three point six five, because two games is not a rate and I regress everybody the same way — and you get 93.9 yards. He has been averaging 72.5. And Minnesota is playing behind Carson Wentz and throwing for a hundred and forty-seven a game, which means they are going to hand it to him until somebody stops them. The flip is 85.9. Take the over at 85.9 or lower and leave it alone above that.
Nico Collins is out with a hamstring, and when a team loses its number-one receiver the targets do not evaporate, they go to whoever is still standing. In Houston that is a tight end. Schultz already owns twenty-five percent of the targets — twenty-two of them in two games, sixteen catches, 175 yards — and Collins' eleven and a half percent share redistributes largely to him. That puts him at 12.4 targets and 94.2 yards. Understand what is going on here: C.J. Stroud threw it fifty-nine times last week and Houston scored six points, which means the volume is enormous and the efficiency is dreadful. A tight end in that situation catches a great many balls for eight yards apiece, and eight yards apiece twelve times is a number. Flip is 86.2.
Baltimore's most productive receiver is doubtful with a hamstring and did not take a snap last week. Zay Flowers is eleven percent of the targets and, more to the point, he is the man who gets thrown to when Lamar Jackson needs something. Andrews is already the highest-targeted Raven at 24.5 percent — thirteen targets, ten catches — and the Flowers share pushes him to 7.8 targets and six catches. Two things cut against it and I will name them: Baltimore has both a tackle and a guard carrying questionable tags, which means more help blocking and fewer routes, and this is a neutral-site game in Rio against a defense missing two starters, which could turn into a track meet where Derrick Henry never leaves the field. Flip is 5.2 catches.
The other side of the Flowers absence. Bateman is Baltimore's second-most-targeted receiver at 19 percent — ten targets, seven catches, 88 yards — and he inherits the deep end of the route tree when Flowers is out, because he is the only man on that roster who runs it. Redistribute and he is at six targets for 58.3 yards. This is the highest total on the slate at 52.5, played on neutral ground in Rio de Janeiro, against a secondary without Malik Hooker at safety. Baltimore throws for two hundred and eighty a game. If they have to keep up, Bateman is where the yards come from. Flip is 50.3.
Twenty carries, 105.5 yards, and the honest note first: this is the prop on the card I am least sure of, because Baltimore's offensive line is held together with tape. Ronnie Stanley's toe and John Simpson's groin are both questionable and both practiced limited, and those are the men Henry runs behind. What is on the other side of the ledger is everything else. Henry owns 68 percent of Baltimore's carries — forty in two games, 212 yards — with Flowers out, which thins the passing game and pushes the offense toward him, in the game with the highest total on the board against a Dallas defense missing DeMarvion Overshown at linebacker. A thirty-two-year-old back with two-thirds of the carries in a game nobody expects to be low-scoring, on a neutral field where neither club gets the benefit of a crowd. Flip is 97.5.
The biggest number I am printing and the simplest argument behind it. Kansas City is laying eleven and a half points in eleven and a half miles an hour of wind. A team favored by that much in weather like that does one thing in the second half, and it does it over and over until the clock says zero. Walker already has forty-seven carries and 290 yards in two weeks — twenty-three and a half carries a game, the heaviest workload in football — and Patrick Mahomes has thrown it seventy-eight times, which will come down in this wind. I have him at 23.5 carries and 137.7 yards. Miami is without Chop Robinson and Ronnie Harrison on that defense and is scoring thirteen points a game, which is the other half of why Kansas City will be ahead. Flip is 129.7 and I would rather have it at 125.
Seventeen carries, 93.9 yards, and this is the prop version of the Bears side. Chicago runs for 212.5 yards a game, the most on the slate, and Swift has thirty-four of the carries to Kyle Monangai's twenty. He is not a bell cow and I am not pretending he is — that split is the risk here, and if Monangai gets hot this goes nowhere. What makes it worth three dollars is that Chicago came out of last week with nobody carrying a game status, and is playing at home on a Monday night against a Philadelphia defense that just lost Jonathan Greenard off the edge to a pectoral. A healthy line against a short-handed front is where rushing yards come from. Flip is 85.9.
Twenty-one and a half carries for 93.2 yards. Taylor has forty-three carries and 190 yards in two games and Indianapolis has shown you exactly what they intend to be: Daniel Jones has thrown it sixty-four times for 376 yards, which is five point nine a throw, which is nothing. When the passing game produces five point nine a throw the running back gets the football. Houston's front is missing Jadeveon Clowney with a knee and Jacob Hummel at linebacker, both ruled out, both did not practice. The Colts are getting two and a half at home in a game my model calls an eight-tenths-of-a-point Houston win. Taylor is how they keep it close. Flip is 85.2.
“Bet less than you brag, learn more than you lose, and never trust a tout who smiles too much.”
— The Betting Fool
Disclaimer: The Betting Fool is for entertainment purposes only. Nothing on this site is financial advice, and past performance does not guarantee future results. Must be 21+ to wager where legal. If you or someone you know has a gambling problem, call 1-800-GAMBLER.