Day Seventeen: The World Cup Group Stage Ends, Plus a Fight Night Surprise
Today the group stage takes its final breath. Groups J, K and L close the book, the last twelve teams find out whether they go on or go home, and the round of 32 stops being a forecast and starts being a fixture list. It is the biggest day we have run, and we have a little surprise saved for the end that has nothing to do with the World Cup at all.
Yesterday
Groups G, H and I wrapped up, and they were not shy about it. France saw off Norway 4-1 with a heavily rotated side, the Haaland against Mbappe billing turning into a comfortable France win and a top seed protected. Senegal needed a statement and delivered a 5-0 demolition of Iraq, the kind of result that gives them a real shot at sneaking through as one of the best third-placed teams. Spain edged Uruguay 1-0 in a tight one to top their group and send a strong Uruguay side home. Belgium, who had drawn their way through the first two games, finally exploded for a 5-1 over New Zealand. And the two that got away from us both finished level: Cape Verde and Saudi Arabia played out a 0-0, and Egypt and Iran shared a 1-1 in which the underlying numbers said Iran were the better side by a distance.
For our model it was a split day. Four winners from six, with both draws beating us, which is the honest hazard of calling tight games. The bruise was on the published bets. We put out three totals plays and all three lost: the Under in Belgium’s game, blown apart by a 5-1, the Under in Senegal’s, buried under five, and the Over in Egypt against Iran that a cagey 1-1 never threatened. Here is the part worth being straight about. Our actual read on those totals, the side the model itself leaned, went a perfect six from six. It was the contrarian edge filter sitting on top, the one that sometimes bets against our own lean to chase a price, that went nought from three. We have flagged that pattern before. Yesterday it cost us a clean sweep, and it is getting fixed.
Today closes the group stage
Three groups, six games, and the meaning model splits them cleanly into the desperate and the done.
Group K runs through Colombia and Portugal. Colombia sit top and need only a draw to win the group; Portugal, with Ronaldo still finding the net at his sixth World Cup, must win to take first off them. In the other K game DR Congo have to beat an already eliminated Uzbekistan and then pray the maths of the best third-placed teams falls their way. A win, or they are out.
Group J is a tale of two temperatures. Algeria against Austria is a straight fight for second place behind Argentina, with a twist of history given the 1982 Gijon ghosts that still hang over the fixture. A draw sends Austria through on goal difference and into the knockouts for the first time since 1954, which is exactly why our numbers, the market and Opta all expect a cagey, low-scoring night. Algeria have to win or sweat on the third-place places. The other J game, Jordan against Argentina, is the purest dead rubber of the day. Argentina have already won the group and are expected to rest Messi and rotate hard. Jordan are out. We can show it on the meaning meter, which reads a flat zero for both.
Group L hinges on Croatia against Ghana. Ghana need only a point; Croatia, Modric and all, must win or hope a draw is enough. And Panama against England is England playing for top spot, a win putting them above Ghana and onto the softer side of the bracket, against a Panama side already eliminated.
Meet the panel, and how they did
Yesterday the four AI forecasters we have been building had their first full slate graded, so for the first time we can show how they actually did, not just what they said.
A quick reminder of who they are. The Sharp is a cold, price-driven professional who only bets when he thinks the number is wrong. The Analyst lives in the underlying numbers and trusts the model over the eye test. The Diehard is pure fan instinct, no spreadsheets, all class and bottle and romance. The Scout works the team news, the lineups, the injuries, who is rested and who needs it.
The early scoreboard tells a clean story, and it is one we half expected. The two disciplined heads, the Sharp and the Analyst, hugged the consensus and matched the best of the field, four winners from six, with a strong day on totals for the Sharp in particular. The two bold ones paid for their boldness. The Scout had the worst day of anyone, two winners from six, because his angle, fading the big sides that were rotating, ran straight into France, Belgium and Spain winning anyway. The Diehard landed in the middle, as erratic as a gut read tends to be. It is six games, far too few to crown anyone, and the whole point of the bold voices is the occasional divergence that lands big. But the first lesson is the oldest one in this sport: the further you stray from the market, the more often it stings.
All four call every game, and we grade every one of them.
I bet you did not know we also do MMA
Here is the surprise. We do not only model soccer. We have MMA models running too, and tonight hands us the perfect way to show them off: the UFC main event, Rafael Fiziev against Manuel Torres at lightweight. It is a fitting test case, because our model calls it a coin-flip.
We give Torres the thinnest of edges, 52 to 48, which lines up almost exactly with the no-vig market at 51, an edge of barely a point on our pick. When the model is this close, the panel earns its keep, and here the two voices genuinely disagree. The Analyst takes Torres at 65 percent, pointing at a blistering 7.29 significant strikes a minute and a finishing record that has not missed in his UFC wins, against a Fiziev who has been stopped in three of his last five. The Diehard takes Fiziev at 60 percent and reads the room instead of the sheet: word of a brutal travel and a rough weight cut for Torres, and a Fiziev fighting back at home with everything to prove.
The deeper data is where it gets fun, and it leans Torres on almost every row. The tale of the tape hands him the reach, the youth, the output, the accuracy and the takedown threat; Fiziev’s one clear edge is that he has never been an easy man to put away, until lately. League history piles on in the same direction: at lightweight the younger fighter, the longer-reach fighter and the rating favorite all win more often than not, and all three arrows point at Torres. His own record is the eye-opener, a perfect finishing run whenever he has been the younger man, the underdog or the longer-reach fighter, which tonight he is on all three counts.
And how does it end? The model says a finish is the likeliest outcome at 62 percent, with a knockout the single most probable result at 48, most of that danger loaded into a first round where Torres has ended people before. Expect a little over three rounds if it goes the distance. Method and round props like these are a market almost nobody prices well, and they are exactly the kind of edge we are building toward.
The biggest day of our World Cup, a panel that is starting to earn its name, and a fight night to round it off.
The knockouts start tomorrow. The bracket is almost real.