Stats

Since 1901 the 30 modern franchises have logged 2,736 team-seasons; Houston owns 65 of them. With 275 franchise-mapped no-hitters in that span, a team in Houston's shoes expects 275 × 65 ⁄ 2736 ≈ 6.5. They threw 18. The Poisson tail P(X ≥ 18 | λ = 6.5) ≈ 1.6×10⁻⁴, under the Bonferroni bar of 0.05 ⁄ 60 ≈ 8.3×10⁻⁴ for sixty franchise-level tests.

285 no-hitters across 126 modern seasons is a mean of 2.26 per year. A Poisson process would put the season-to-season variance near that mean; the observed variance is 3.51, a dispersion index of 1.55. The test statistic (index × 125 degrees of freedom) comes to 193.8, p ≈ 8×10⁻⁵ against pure randomness.

At the long-run rate of λ = 2.26 per season, P(X ≥ 9) ≈ 5.7×10⁻⁴: about one season in 1,750. Across all 126 modern seasons, the expected count of nine-no-hitter years is 0.07. We got one.

Before 1990: 4 combined among 220 no-hitters (1.8%). Since: 18 of 110 (16.4%). A two-proportion z-test against the pooled rate (22 ⁄ 330) gives z = 5.0, p ≈ 3×10⁻⁷.

From 53,336 pitching starts since 2015, the chance a hitless bid survives each inning, given that it got there: 42%, 45%, 47%, 43%, 47%, 45%, 43%, 47%, 50% for innings 1 through 9. The full product is 0.0008, about 1 start in 1,250. These are the same rates drawn in the bid survival curve above.

46 of the 330 won 1–0 (13.9%, about one in seven). Across all MLB games, a 1–0 final turns up roughly once in 40.

330 no-hitters over roughly 460,000 team-games (230,000 games, two pitching sides each) is one per ~1,394 starts. A 773-start career at that rate expects λ = 0.55. P(X ≥ 7 | λ = 0.55) ≈ 2×10⁻⁶, about 1 in 500,000. Koufax's four: P(X ≥ 4) ≈ 1 in 400.

Same exposure model as the Astros entry: 276 franchise-mapped victims over 2,736 team-seasons puts the Phillies' 126 seasons at 12.7 expected. They have suffered 19. P(X ≥ 19 | λ = 12.7) ≈ 0.059, a result that chance alone produces about 6% of the time.

Run 4,000 random scatters of 330 no-hitters across a season's calendar (weighted by each month's real share): more than 90% of the simulated histories produce at least one date holding six or more. Real history's busiest dates hold exactly six. The only two double-no-hitter days ever (April 22, 1898 and June 29, 1990) also sit on expectation; the per-season pairing math predicts about two.

Observed vs. expected by month, using each month's approximate share of games played: April 0.95×, May 1.06×, June 1.02×, July 0.86×, August 0.82×, September 1.27×. The whole pattern tests at p ≈ 0.27 (chi-square, 5 degrees of freedom): consistent with noise. The exposure shares are estimates, but the conclusion survives any reasonable version of them.

Model: first-hit depth ∼ park + pitching-team-season + batting-team-season + home⁄away, ordinary least squares over every pitching side since 2015. Each team’s road games pin down its own quality; whatever remains at its home park is the venue. The park block is highly significant (F = 4.01 on 31 degrees of freedom, p ≈ 4×10⁻¹³): venue adds real information after both teams are controlled. It is a thin slice of a noisy outcome, though; no single block explains even 1% of game-to-game variation (park ≈ 0.3%, each team-season block ≈ 0.9%). Venue-only effects span ±0.22 hitless innings and correlate 0.85 with the raw park table. They are persistent: even- and odd-season estimates correlate 0.69. The signal fades past the sixth inning, where 455 events since 2015 are too few for park-specific claims.

Across 54,274 pitching sides since 2015, home staffs average 0.815 completed hitless innings to start a game, visitors 0.733: an edge of +0.081, p ≈ 3×10⁻¹⁵ clustered by game. In first-hit terms: inning 1.81 versus 1.73. Fitting a separate home edge per park on top of team-season controls improves nothing (F = 0.69, p = 0.90). That is an absence of evidence for park-specific familiarity, not proof that every park’s edge is identical or even positive; a handful of raw park splits currently lean the visitors’ way.

First-pitch temperature from MLB’s schedule feed, 45,396 outdoor pitching sides since 2015. Mean hitless depth by temperature band, a broad downward gradient: below 55°F 0.81, 55–64 0.82, 65–74 0.78, 75–84 0.74, 85+ 0.70. In first-hit terms: inning 1.81 in the coldest band, 1.69 in the hottest. Adjusted for park, team-seasons, and day⁄night, the association is −0.030 innings per +10°F (F = 31.4, p ≈ 2×10⁻⁸), and adding month-of-season controls strengthens it to −0.044 (p ≈ 5×10⁻¹⁰) rather than explaining it away. Observational, not causal. Domes and closed roofs are excluded; their “weather” is a thermostat.

Same fixed-effects framework as the ballpark entry, with MLB’s first-pitch weather joined to all 54,274 sides. Day⁄night: +0.003 innings, p = 0.79 (the fussiest model version reaches p = 0.07 with a coefficient still too small to matter). Wrigley’s own split: 0.800 hitless innings by day, 0.797 by night. Wind, outdoor games only: in⁄out speed slopes per +10 mph are −0.005 and −0.011 (joint p = 0.73), direction classes alone p = 0.99, direction plus direction-specific speed p = 0.13; nothing tested clears significance. Roof closed, tested within the seven parks that actually open and close one, with month-of-season controls: +0.041 innings, p = 0.18. The roof decision follows the weather, so the archive cannot separate the two.

The full fixed-effects model reaches R² = 0.024 over 54,202 deduplicated pitching sides: about 2.4% of the variance in how long a game’s opening no-hit bid survives, from more than 700 parameters. Every effect in this section is real and repeatable in aggregate, and nearly useless for calling tonight’s game. That is why we watch.

104 line outs across the 40 games, 15% of their balls in play. At a league line-drive batting average near .650, P(all 104 caught) = 0.35¹⁰⁴ ≈ 10⁻⁴⁷. The .650 is a league-wide figure; the 104 line outs are counted directly from the pitch logs of the 40 no-hitters. Selection does the heavy lifting: a game only makes this list if every liner finds a glove.

372 strikeouts in 1,162 plate appearances (32.0%) and 89 walks (7.7%). League rates run near 22% and 8.5%, respectively. Called strikes plus swinging strikes came to 32.5% of all 4,707 pitches, against a league norm near 29%.

Median xBA .615; 63% graded .500 or better; 7.4% under .200. By exit velocity, 19.9% left the bat under 85 mph. Batted-ball type is recorded for 393 hits, one more than carry an xBA grade: 47% line drives, 33% ground balls, 20% fly balls, and zero popups.

Of the game's 147 pitches, 146 carry no type code in MLB's feed; Estadio de Beisbol Monterrey had no tracking system installed. The chart's Other bucket holds 398 pitches: 114 sweepers, 78 knuckle curves, 59 splitters, and 147 untyped, 146 of those from Monterrey.

Luplow's outs came against Rodón (April 14, 2021) and Miley (May 7, 2021). Realmuto's came in the five-pitcher Mets no-hitter (April 29, 2022) and the Astros' Game 4 World Series no-hitter (November 2, 2022). Suárez's came against Arrieta (April 21, 2016) and Fiers (May 7, 2019). Guerrero Jr.'s came against Manning and two relievers (July 8, 2023) and Blanco (April 1, 2024). It's less cosmic than it looks: a team that gets no-hit twice sends mostly the same hitters to the plate both times, and somebody has to bat last.

Six plate appearances since 2015 show a bunt attempt in the pitch log against a live no-hit bid, in 2015 (twice), 2020, 2021, and 2024 (twice). The outcomes were three strikeouts, a groundout, a walk, and one completed sacrifice bunt, laid down by the opposing pitcher, Alex Wood, against Arrieta in 2015. The count is a floor; a batter who squares and pulls back leaves no trace in the pitch record.

Among the 392 Statcast-graded spoiler hits, the maximum exit velocity is 115.7 mph (April 12, 2018) and the minimum is 23.0 mph (September 25, 2024, sixth inning); the median is 98.5 mph. The minus-63-degree hit came off Tyler Anderson on June 15, 2022, at just 48.7 mph: Jared Walsh chopped it nearly straight down and beat it out for a single that never left the infield.

Of 2,243 dates with games since 2015, 407 saw at least one bid finish the 6th hitless (18.1%), 183 the 7th (8.2%), and 90 the 8th (4.0%). These are the same per-day maxima that color the season calendar.

Combined no-hitters excluded; this counts the 308 thrown start to finish by one pitcher. For scale: no-hitters happen about once every 1,400 starts, so even a pitcher who goes on to make 150 more starts — several full seasons — has roughly a 1-in-10 shot at throwing one. Among pitchers who already had one, one in seven did it again.

At the league-wide rate of 330 no-hitters over 2,736 team-seasons (0.121 per team-season), one particular team going 45 seasons dry is P = e^−5.4 ≈ 0.004. But thirty franchises roll those dice at once: the longest drought among 30 should run about ln(30) ⁄ 0.121 ≈ 28 seasons, and the chance that at least one reaches 45 is about 12%. Runners-up: Toronto (Dave Stieb, 1990), Kansas City (Bret Saberhagen, 1991), Atlanta (Kent Mercker, 1994).

Mean 118, median 114 across the 40 pitch logs in the Anatomy of a no-hitter data. 26 of 40 exceeded 110 pitches; 12 exceeded 120. Maximum: 159 (the five-pitcher Mets game, April 29, 2022). Minimum: 93 (Valdez, August 1, 2023).

The four: Edinson Volquez (June 3, 2017), John Means (May 5, 2021), Domingo Germán (June 28, 2023 — the perfect one), and Framber Valdez (August 1, 2023). Perfection isn't about facing 27 batters; it's about nobody reaching base at all. Means lost his to the dropped third strike, Volquez and Valdez to walks — and every one of those runners was wiped out on the bases, which is the only way 27 batters can produce 27 outs.

The 2000 through 2024 seasons averaged λ = 3.12 no-hitters per year. P(zero) = e^−3.12 ≈ 4.4%, about 1 in 23.