The Train Keeps Rolling: World Cup Day Five!
The model is performing really well. Today we'll take a deeper look at it.
I want to do something a little different today. The picks are below as always, but the real story on Day Five is the model itself, because this is the point in the tournament where it stops being a static set of predictions and starts being a living thing that moves a little every single day. So today is a model talk day. Stick with me, because watching this thing breathe is half the fun.
The model is live, and you can watch it move
Here is the part I am most excited about. The model is not a one-and-done prediction we published before the tournament and walked away from. Every day, as results come in, it recomputes. Ratings shift, win probabilities re-settle, advancement odds swing. It is a living forecast, and you can watch it update in real time on the site.
You can find the live model here: World Cup 2026 Live Model
Bookmark that. Over the next few weeks it is going to be the most honest scoreboard we have, because it shows you not just what we think now, but how much our thinking has changed since kickoff. Which brings me to three moves from the last few days that are worth your attention.
Three teams the model just repriced
Sweden surged. After their 5-1 win, the model jumped Sweden by more than fifty rating points in a single result. They are now a top-fifteen side in the entire field, and their odds to get out of the group went from about 82 percent to 98.5. That is what a statement win does to a living model. It does not politely nudge the number, it grabs it.
Brazil slid. This is the one that will raise eyebrows. After the 1-1 draw with Morocco, our model now has Brazil down as roughly its twentieth-ranked side, and here is the wild part: it now rates Brazil below Morocco, the very team they drew. Their odds to win the whole thing have dropped to around 6 percent, off a couple of points since we started. The market still treats Brazil like Brazil. Our model is quietly telling a different story, and we are going to let it ride and see who is right.
Ivory Coast locked in. Before their game with Ecuador, the market made Ecuador the favorite. Our model disagreed and made Ivory Coast the side to back, and we said so out loud. They won 1-0. The model has now moved Ivory Coast all the way to a flat 100 percent chance of escaping the group. We called that one against the grain, and it is the cleanest example yet of the model seeing something the market did not.
Yesterday, quickly
Day Four was a good one. Ivory Coast swept all three of our reads, Sweden won, and the only miss on the published card was the Netherlands under in a 2-2 game that got away from us. Our key picks, the plays where our edge clears ten points, are now sitting at 12-2 with a return of about 82 percent. Small sample, variance is patient, same caveats as always, but the discipline is holding.
A quick word on the rules, because today tests them
Two things shape what we publish. First, we only post plays where our model and the market disagree by ten points or more. Second, and this is new, we will not fade a heavy favorite in the winner market. If the market makes one side 65 percent or more to win, we do not bet the underdog or the draw against them, even when our model technically likes the price. You will see exactly that on Spain today. Totals and both-teams-to-score are untouched by that rule, because a low-scoring read can be right even when the model is wrong about who wins.
Today’s four games, in kickoff order
Spain vs Cape Verde, 12 ET
This is the perfect game for a model talk day. The market makes Spain a monster here, around 91 percent to win on Kalshi against a World Cup debutant. Our model does not see a monster. It has Spain as a solid, mid-tier side, only about 2 percent to win the whole tournament, a clear notch below the Argentinas and Frances of the field, and just 63 percent to win this match. That is a massive gap, and it is going to be fascinating to watch over the coming weeks whether the model is onto something or whether it is simply underrating a heavyweight.
But here is the discipline in action. Even though our model technically likes the draw and even Cape Verde at these prices, we are benching both. We do not fade a 91 percent favorite, full stop. That is a rule we just made precisely so we do not talk ourselves into a longshot against a confident, sharp market. The one play we publish is the Under 2.5 goals, our biggest edge on the board, because the goals market stands on its own and our model sees a far tighter game than the blowout everyone expects.
Belgium vs Egypt, 3 ET
Belgium are favorites and our model likes them even more than the market does, so this one is simple. The play is Belgium to win. No fading here, we are with the chalk and getting a better number than the market is charging.
Saudi Arabia vs Uruguay, 6 ET
A classic low-event read. The model sees Uruguay managing a tidy, low-scoring win, and it lands on two plays: Under 2.5 goals and both teams to score, No. Uruguay should have enough, and Saudi Arabia are not expected to find the net.
Iran vs New Zealand, 9 ET
Another grind. Iran are the rightful favorites and our model is more confident than the market, so Iran to win is the play, alongside both teams to score, No. Expect a tight, tense, low-scoring night, exactly the kind of game these two produce.
The card, in one line
One published under on Spain with two benched longshots we are proudly skipping, a Belgium moneyline we like more than the market, and a pair of low-event reads on Saudi-Uruguay and Iran. Every price is logged, the model is live for you to watch, and by tomorrow we will know how it went with the receipts out. Go watch the model move. Enjoy the football.