Pascal Initiative @PascalInitiativ

Most pitching changes are too close to call. Across 84,501 keep-or-pull moments in 2026 MLB starts, the typical gap between keeping the starter and going to the bullpen was 0.16 points of win probability, inside about 1.25 points of talent uncertainty. pascalinitiative.com/insights/when-…

Our rating-based version (opponent-adjusted EPA/play) has Jacksonville No. 1 at 115.2, the unbeaten 49ers half a point back at 114.7, and... the Rams still 3rd even with a 2-3 record. We also score last week's calls: favorites won 10 of 15 vs ~8.5 expected. pascalinitiative.com/power-rankings…

@EvanHAbrams A -175 price implies about a 64% win chance. So a -175 favorite still loses roughly 1 time in 3. That isn’t a broken favorite, it’s what 64% means. The real test is whether 64% calls win ~64% over many games. pascalinitiative.com/ratings/track-…

@tangotiger @mike_petriello @JoeyVotto Method: Statcast's bat-tracking intercept on every swing, compared with what each hitter's own regular-season swings predict for the same pitch location, speed and type. Data through Oct. 14.

@tangotiger @mike_petriello @JoeyVotto How do you show a hitter is late? Measure where bat meets ball. Adjusted for the pitches he saw, Ohtani met the ball 5 inches later than his 2025 norm in NLCS Games 1–3.

@tangotiger @mike_petriello @JoeyVotto The NLDS looked late too, but 73% of the pitches he swung at were fastballs (47% in August). Adjusted for that, it matched his regular season. Some of it is just October: last postseason, while hitting well, he ran 1.3 inches behind his norm.

@tangotiger @mike_petriello @JoeyVotto It isn't the parks or the tracking: the other 21 hitters in those games averaged 0.0 inches off their own 2026 norms. But with about 17 swings each, 3 inches either way is routine. Three games is mostly noise.

@tangotiger @mike_petriello @JoeyVotto Here's where contact actually happened, seen from above. NLCS: 13 inches behind the front of the plate, vs. 1.4 inches in 2025. That's raw. Part of the gap is where he stands (2 inches deeper in the box) and the pitches he's seen.

We publish win probabilities before kickoff, then score them. Week 6: favorites won 10 of 18 games, against 12.6 expected. That was an upset-heavy week. Our average miss was 12.3 points, better than the model's typical error of 15.9. Biggest upset: North Carolina over Pittsburgh, as a 26% underdog. pascalinitiative.com/power-rankings…

NFL favorites won 10 of 15 in Week 5. We expected about 8.5. Average miss: 10.6 points vs. a 13.0-point typical error. Biggest upset: Buffalo 37–34 at the Rams, a 37% underdog. Biggest miss: Green Bay 31–7 over Chicago, 26.6 points off. Scorecard: pascalinitiative.com/power-rankings…

@MikeRegaladoLA Our ratings have them nearly identical: UCLA No. 22 (109.5), USC No. 23 (109.4). Oregon (113.3) stays ahead of UCLA despite the 17–15 loss. A 2-point result moves the ratings only a little, and Oregon’s full season still grades out better. pascalinitiative.com/power-rankings…

@Coach_Yac By the numbers it’s basically a coin flip at the top: Jaguars 115.2, 49ers 114.7. pascalinitiative.com/power-rankings…

@jaredsmithbets @KillerSportsCom Interesting wrinkle: our Pascal Rating makes Indiana a 3.7-pt home favorite (59%) vs Ohio State, the other side of the market. pascalinitiative.com/power-rankings…

Jacksonville holds No. 1 in the Pascal NFL Power Rankings. Unbeaten San Francisco is half a rating point back after winning at Seattle. Week 6 Top 10: 1 Jaguars 2 49ers 3 Rams 4 Chiefs 5 Seahawks 6 Bills 7 Patriots 8 Texans 9 Ravens 10 Broncos All 32: pascalinitiative.com/power-rankings…

@CJNitkowski Agree the hesitation shows the value. We tried to price it at Pascal: in the early-game spots we've modeled, a failed challenge gives up ~0.3 runs of future challenge value. More challenges = lower cost = more calls worth contesting. pascalinitiative.com/insights/what-…

@yarritsblake @UmpScorecards The two Turang calls weren’t equal chances to challenge. Pascal’s Decision Opportunity Score: 0-2, 1st, 1 out, empty = 27. Full count, 3rd, runner on 2nd = 67 (fix worth ~0.62 runs vs ~0.30 lost on a miss). It scores the the moment, not future results: pascalinitiative.com/insights/prici…

Same at-bat, two challenges, two different bars. 1-0 pitch: worth 0.16 runs, needed a 66% chance. Overturned. 2-1 pitch: worth 0.21 runs, needed only 59%. Upheld. A lost challenge cost the same both times. The bigger pitch was the easier call, and it’s the one that failed.

@FansOfCFB Resume and strength answer different questions. Our Pascal ratings are a forecast of strength, not a resume: Notre Dame No. 1 and Georgia No. 2 (both 6–0), while Kentucky sits No. 31 even after beating LSU. Our rankings + last week's scored expectations: pascalinitiative.com/power-rankings…

Pascal Power Rankings after Week 6, heading into Week 7. Full Top 25: South Carolina, UCLA, USC and Louisville move in. Texas A&M drops 12 spots to No. 25. pascalinitiative.com/power-rankings…

Notre Dame is back at No. 1 in the Pascal Power Rankings after a 58–7 win over Stanford. Georgia moves to No. 2 after winning 20–17 at Alabama. The top six are within 2.5 rating points of each other. Biggest climb: Louisville, up 12 to No. 24.