CY Elite Lab

Confirmed / Projected — whether both teams have ALREADY published their official lineup for the day or not yet. With the lineup confirmed, the Model projects with the nine hitters who will actually play.

The odds next to the team — what the Market is paying now: the consensus price, and if there is none yet, the reference market price.

% win and the odds in parentheses — the probability of victory that the Model gives to each team and, below, the odds that would be fair if the Model were right, with no commission. If the odds above pay more than the ones below, the Model sees edge on that side.

EXP RUNS — the runs that the Model projects for each team. It is a projection, not a line. It is tracked each morning against the runs that actually scored.

o4.5% — the REAL probability that that team exceeds 4.5 runs. You cannot deduce it by looking at the EXP RUNS: a team scores 4.41 on average but 4.0 at the median, and scores zero 7.2% of the time. That's why the balance of an over 4.5 is not at 4.5, it's near 5.2: with the team projected at 5.00 the over still pays at 48%.

Expected Total — the sum of the two projections; below, the Market line with the over odds and the under odds. The expected total is not compared to the line, and it's the easiest error to make: it is a projection, and the line is set where half the games cross it (the median). Over 368 games, at each line they score on average between 0.2 and 0.8 runs MORE than the line, so a well-calibrated model always comes in above it. For the over/under check the chip below, which already compares probability against probability.

Detail ▾ — opens, by team: its profile in five categories, the starter with his last 5 starts, the lineup with each batter's OPS and the real bullpen load in 48 hours.

Profile · CON POW EYE BSR DEF — how that team plays, not just how much: CONcontact (how often it puts the ball in play), POWpower (how much it hits when it connects), EYE (how much it walks), BSR (how much it gains running) and DEFdefense. Two teams with the same OPS can play in opposite ways. It goes in the Detail and not on the card because it is context: the Model projects with the OPS against today's pitcher hand, not with these letters.

How defense is measured — with defensive efficiency: what percentage of balls in play it converts to outs. Not with the fielding percentage, which only counts errors — and a slow outfielder who never gets to the ball never commits one. Among the 30 teams efficiency ranges from .684 to .737; fielding bunches between .983 and .991.

Order and the freeze — the cards are ordered by first pitch time, from the earliest of the day to the latest. When a game starts, its card freezes just as it was: a record of what the Model said BEFORE, and refreshes only touch games that have not yet begun. There is no live score here: the Model shows what it projects, not how the game is going.

ERA · FIP · K-BB — the three for the starter, always in that order and with its note beside it. The ERA is what they have scored against him; the FIP is what truly depends on him; the K-BB%, how many strikeouts minus what he gives away. They are the numbers that everyone recognizes, but the Model does not project with any of them (see «What the Model projects with»).

The two FIP — on the card goes the one from this season, the one you recognize at a glance. In the Detail goes the three-season, which is the pitcher's baseline quality. When the two diverge sharply, that itself is information: that pitcher is having an unusual year relative to his line. The grade (A/B/C) always comes from the three-season.

What the Model projects — from Sep 17, with a composite: two-thirds of the Statcast xERA (the ERA he deserved based on contact quality allowed) and one-third FIP. With it the Model ranks the ~240 starting pitchers in the league in a ranking that is tracked every day, as the reference market does. If a starter has no Statcast sample, the Model uses his three-season FIP. ERA does not enter: splitting the season of 80 starters into even and odd starts, the correlation between the two halves is −0.06 in ERA y +0.32 in FIP. Today's ERA does not predict tomorrow's.

A starter not yet announced (TBD) — is not treated as an average starter: he enters as that team's bullpen, who will actually pitch. Measured across 2,206 team-games: teams without an announced starter allowed 5.13 runs against 4.48 from the rest.

A FIP with no grade — means that pitcher does not have enough sample (fewer than 300 batters over three seasons). It is not an error: it is the way not to hand him a letter he has not earned. Inside the Model does the same: it shrinks him toward the league average.

F (FIP) — measures only what the pitcher controls (strikeouts, walks, home runs); removes luck and defense, which is why it predicts better. Same scale as ERA. FIP lower than his ERA = improvement coming; higher = overvalued.

BP — that team's bullpen: the ERA and FIP of its relievers, equally weighted. Alongside, how they arrive: rested, the innings he has pitched in 48 hours, or how many relievers are burned out (threw on two straight days). Since Sep 17, the Model measures it with a ranking of the 30 bullpens that mixes each reliever's FIP and xERA, weighted by their innings: a bullpen's ERA is one of baseball's noisiest numbers.

OF — that team's offense against the hand pitching today, not its overall index: 100 = the league. Batting .700 against lefties is not the same as against righties, and the Model crosses each offense with the opposing starter's hand.

PF · WX · UMP — the three factors that multiply the projected total, each with the % it moves. They are shown separately so you can see where the number comes from.

PF (park factor) — how many runs this stadium scores relative to the league average: the Statcastrun index, three seasons and by ballpark (not by team, which breaks when a club moves). It is used in full, without trimming: Coors +25%, Globe Life −12%.

WX (weather) — how far the day departs from what is normal in THAT park in THAT season, not the league average: the typical Kansas City heat in August is already inside its park factor, and counting it twice inflated the total. Uses temperature, wind and its direction. Measured over 1,513 games: +0.5% per degree above normal and +0.4% per mph of wind toward center field. Closed = the roof is closed and weather does not apply.

UMP (umpire) — effect of the home plate umpire on runs, measured in 6,934 games over three seasons. Each game is compared to that stadium's average that year (so the park does not contaminate) and the result is shrunk by how many games the umpire has called: at 80 games it retains less than one-third of its number. The true dispersion among umpires is 0.31 runs. If it does not appear, the umpire has not yet been announced.

Fair line — the market consensus with commission already removed: what the market truly believes. It shows how much the Model differs from consensus. The Model's edge, by contrast, is measured against what you get paid (see below), because that difference is what decides whether money is made.

CHC ML +4.2% — the Model's edge on the winner, in percentage points: what the Model gives that team minus what the price you get paid demands, with commission inside. It appears whenever there is an edge, large or small; if neither side has one, the chip says "No ML edge".

Over 9.0 +2.1% — the same for the total, also in percentage points. The Model converts its expected runs into probability of over and under, and each side is compared with what that side pays you; whichever has the most surplus appears. Today is context, not a signal: on Sep 17, over 227 games, the totals model had the side hitting 50.9%, and at −110 it takes 52.4% to break even. Now it is tracked each morning: the side that marks hits 45.8% needing 52.0% (83 games); it will turn green when it hits 3 points above in 150 games. It will turn green from 3 points, like the winner's, the day the side it marks is right at least 3 points above what its prices required, over 150 games; and if it drops later, it goes back to white only.

The chip's color — green = 3 points or more (on the winner; on the total, only when it has earned it, see above), white = less than 3, gray = no edge. Heads up: the color does NOT mean "stronger signal". It is only how far the Model separates from the price. Read the entries below before using it.

The Model's edge is NOT a signal, and this was corrected on Sep 9, 2026 — until that date this said the band of 3 to 6 points was "the only one that made money (50.3%, +4.8% ROI)". That number came from comparing the Model with the history of analysis. Measured the right way, against the ACTUAL results of the games, the opposite came out (with the Model and the measurement method before Sep 17):

3 to 6 points — 40 games: hits 35.0% needing 47.4%. ROI −31.2%, with confidence interval entirely below zero. It was the worst of the three bands, not the best. · 6 to 10 points — 41.2%, ROI −12.4% · More than 10 — 28.6%, ROI −18.9%.

Why it failed — over 507 games with results, the Model had a Brier of 0.2488 versus 0.2386 from the Market, worse than always saying the same thing (0.2480). It compressed everything toward 50%: when the Market gave 68% to the home team, the Model said 58%. And when it disagreed with the Market it was wrong in the direction of its disagreement.

What corrected on Sep 17 — it was recalibrated with 372 games split in two: the slope, which measures how much the Model commits compared to the Market (1.00 = equal), went from 0.47 a 0.95, and the Brier fell from 0.2433 to 0.2378 in the half the Model had not seen. That same day the starter composite and the bullpen ranking entered, so that measure needs to be repeated: it is done with data saved each day since then.

What to do with this — since Sep 9 the Model doesn't reach the brain: the analysis is decided without it, with pitching, bullpen, offense, form and price. The door to return is written in advance: that its Brier equals or beats the Market's over 150 games or more, and each morning's report monitors it alone. Still it would return gradually: first in shadow, with no decision, and then with a small weight. Meanwhile, use it to understand the game —what offense hits, what bullpen holds, how the park and weather weigh—, not to choose the side.