Method. Each team's talent is its Pythagorean record regressed toward .500, worth about seventy games. We simulate every remaining game, then the entire bracket — best-of-three, best-of-five, two best-of-sevens — and count. Remaining games are treated as league-average opposition; real systems use the actual schedule and ours does not yet. That will overrate teams with a hard finish and underrate teams with an easy one.

Why we track closing line value instead of a win-loss record. A record measures variance for several thousand bets before it measures skill. Whether the market moved toward your price after you took it converges in about fifty. If a site leads with "47-31 last month," it is showing you a coin landing.

The most valuable thing in September and almost nobody publishes it. A team under 2% to reach October will rest nobody and play everybody young. A team over 98% will sit its stars and pull its starter at eighty pitches. Both facts change who you should start tonight, and both fall straight out of the simulation.

These are arguments, not items. The Wire on the front page is short, dated and sourced; this is the long version with a point of view. House rule: the target is always a claim, a cliché, a situation or a team — never a person's competence.