The 2026 season
In numbers, from June to the World Series
242analyses
54%52.5% needed
+2.3%balance
A margin of +1.7 points, above what's needed, but only just.
June30+14.3%
July83+8.8%
August88-9.8%
September36+10.6%
October5-18.8%
By odds range
Worse than −1605 analyses · 4-1
+34.8%
−160 to −110114 analyses · 56-54
-10.9%
−110 to −10059 analyses · 36-23
+19.7%
+100 to +20064 analyses · 33-31
+9.2%
Favorites (−160 to −110) left −10.9% over 114 analyses; even games (−110 to −100), +19.7% over 59. That's why the system asks a favorite for more edge than an even game.
What it said, against what happened
analyseshit
Said under 50%2454%
Said 50 to 55%7545%
Said 55 to 60%3868%
On average it said 52.8% and hit 53.3%: what it says is, more or less, what happens. That's 137 closed analyses where the system gave its probability: the ones that went out as a signal and the ones that didn't. Each band still carries a lot of chance.
What changed in 2026
The Model stops deciding the analysesOver 507 games, its probability missed by more than the market price. It still has its own pageSep
Every signal, in four stepsWhat the market says, what we see, what could go wrong and why it goes in anywaySep
Analyses go liveIn rounds, with the lineups already outSep
Only prices from −162 to +125And the more of a favorite, the more edge it's asked forSep
What would you use first in 2027?
Noted. We already have your email: no need to give it again.
opens with the plans on March 18
Reserve your spot and keep the founder price: −25%, for as long as you stay.
Reserved. Founder price saved. Just one email, on launch day. And the weekly summary, on Mondays: unsubscribe in one click.
The full record