Pick No. 3: Detroit released this pitcher in April 2024. Tuesday he starts against the team standing between them and October
The call: Tigers beat Cleveland on Tuesday night at Comerica. Confidence: low. Detroit is 0-6 against this team, and the man they are handing the ball to was released by their own Triple-A club twenty-seven months ago.
Start with that, because it is the most Detroit sentence available. On April 27, 2024, the Toledo Mud Hens released a 30-year-old right-hander named Drew Anderson. Tuesday night, in a game the Tigers cannot really afford to lose, he takes the mound at Comerica Park against the best starter Cleveland has.
The four years in between
Anderson signed with Philadelphia in 2012, a kid out of Reno. He gave the Phillies seven years, most of them injured, and 21 major league innings with a 7.71 ERA. They released him in September 2019. The White Sox signed him, used him for one and a third innings in 2020, and released him too. That winter he pitched for the Melbourne Aces in the Australian Baseball League. Texas gave him 22 innings in 2021 and released him in October.
Then he left. Two years with the Hiroshima Toyo Carp in Japan. A spring training invite from Detroit in 2024 that ended with Toledo cutting him in April. And then the part that matters: two seasons with the SSG Landers in Korea, where in 2025 he went 12-7 with a 2.27 ERA and 245 strikeouts in 171 and two thirds innings. That is not a guy hanging on. That is one of the best pitchers in a good league.
Detroit signed him back in December, one year, seven million dollars, with a ten million dollar club option for 2027. He is 32.
What he has actually done here
The Tigers used him as a long reliever for four months and he was good at it: 41 appearances, 67 and a third innings, a 4.01 ERA, 78 strikeouts against 25 walks. That is a 27 percent strikeout rate, which is genuinely excellent, and an 8.7 percent walk rate, which is not.
Then the deadline happened. Detroit traded Tarik Skubal and Casey Mize, which is a sentence that still looks strange in print, and on August 4 A.J. Hinch announced that Anderson and Jackson Jobe were being stretched out as starters. Anderson made his first one the next day.
So the rotation that is supposed to carry a wild card push includes a converted long man making his fourth start and a kid coming back from Tommy John surgery.
The number that decides Tuesday
Here is the honest problem, and it is not about how good Anderson is.
He has never faced more than 18 batters in a major league game. Not this year, not ever. Eighteen was his ceiling on May 20, and he needed 63 pitches to get there.
| Anderson's three starts | Line | Batters faced | Pitches |
|---|---|---|---|
| May 20 vs Cleveland | 4.2 IP, 0 ER, 2 H, 7 K | 18 | 63 |
| Jun 15 at Houston | 2.2 IP, 1 ER, 2 H, 5 K | 12 | 59 |
| Aug 5 at Seattle | 3.2 IP, 0 ER, 3 H, 0 K | 13 | 42 |
One earned run in eleven innings across the three. The stuff is not the question. The question is the third column.
A Detroit start this season has faced a median of 22 batters. Nineteen of the team's 118 starts have ended at 18 batters or fewer, and five of those ended at five batters or fewer, which is a bullpen game wearing a starter's name. Anderson's best day is still below the middle of that distribution.
| Batters faced by the Detroit starter | Starts |
|---|---|
| 2 to 3 | 1 |
| 4 to 5 | 4 |
| 8 to 9 | 1 |
| 10 to 11 | 2 |
| 12 to 13 | 3 |
| 14 to 15 | 3 |
| 16 to 17 | 3 |
| 18 to 19 | 6 |
| 20 to 21 | 18 |
| 22 to 23 | 31 |
| 24 to 25 | 28 |
| 26 to 27 | 12 |
| 28 to 29 | 5 |
| 30 to 31 | 1 |
Which means the bullpen covers four innings, maybe five, against a division rival. That same bullpen has 23 saves and 26 blown saves, a 47 percent conversion rate, second worst in baseball, behind a 3.49 team ERA that says the arms are fine and the sequencing is not. On Sunday in San Francisco it blew another one in the seventh and Detroit won in the tenth anyway. That is not a plan, that is a coin landing correctly.
So why am I taking the Tigers
Because of one number that everybody covering this series is going to skip.
Cleveland's offense is 28th in baseball. Not bad, not slumping: 28th of 30, 3.97 runs a game. Detroit scores 4.57 and sits 11th. Detroit's run differential is plus 87. Cleveland's is minus 27. That is a 114-run gap between two teams a single game apart in the standings, and the standings are the thing that is lying, not the run differential.
Cleveland has lost seven of ten. Detroit has won seven of ten and is a game out of the last wild card spot. Detroit is 30-27 at home. And the six-game sweep that everyone will lead with, I already went and tested: hold every team's game-by-game run totals fixed, shuffle which opponent they came against, two thousand times, and 74 percent of those shuffled seasons produce a worst matchup at least as extreme as Detroit's against Cleveland. It is the worst offensive split in baseball and it is still the kind of thing chance hands out most years to somebody.
That test ran yesterday, on this same series, and it said the sweep is the kind of thing chance produces. Picking Cleveland tonight would mean deciding that a finding published one day ago stops applying the moment it is inconvenient.
What scares me
The best argument against everything above, and I did not see it until I went looking for it. The 28th-ranked offense line is doing a lot of work in this entry, so here is what that offense has actually done to Detroit: 24 runs in six games, 4.00 a game, against a season rate of 3.97. Cleveland has not been a bad offense against the Tigers. They have been exactly themselves, and they have won every time, because Detroit scored 1.83. The gap in this series has been entirely on Detroit's side of the ball. So the pillar I am leaning on, that a weak lineup will not hurt Anderson, is the one pillar that has produced nothing at all in six tries.
Tanner Bibee, next. He has made two starts against Detroit this year and thrown 15 innings with three earned runs, a 1.80 ERA, and Detroit put four hits on him in eight innings on May 20. The rest of the league has done better: 3.85 for the season, and 23 home runs allowed in 138 innings, which is 1.50 per nine and the one seam in him. Detroit has to find it, at home, off a guy who has not given them anything all year.
Third, the shape of the May 20 game is the exact nightmare. Anderson threw four and two thirds shutout innings with seven strikeouts against this same lineup, Bibee threw eight, and Cleveland won 3-2. Detroit's starter being excellent has already failed to be enough once in this precise matchup, with these two pitchers, in this ballpark, eighty-three days ago.
The call
Better team, at home, with the better lineup, against the third worst offense in baseball, riding a converted reliever who has been terrific in a role he has never held for more than 63 pitches. If Anderson gives them four and the bullpen does not do the thing it does half the time, Detroit wins.
That is a lot of ifs, which is exactly what Low means.
Tigers win. Low confidence.
Sources: MLB Stats API, pulled 2026-08-10, for all records, run totals, standings, run differentials, pitching lines, game logs, batters faced and save and blown save counts. Every figure in this entry and the chart come from one execution of scripts/anderson_start.py in the repository. Anderson's years in Japan and Korea and his KBO lines are from his Wikipedia) entry; his contract terms, the rotation move and Hinch's comments are per MLB Trade Rumors, August 4, 2026. His release dates and minor league assignments are from the league's own transaction log. The shuffle test on the Cleveland split is from scripts/opponent_splits.py, published in full here. Chart generated from live data by scripts/anderson_chart.py.
Not betting advice. Just calls, made in public and kept in public.
Got a Detroit number you want looked at? Email projectunmuted@proton.me and it gets looked at properly. A stat somebody quoted that smells wrong, a thing you have always assumed about one of these 4 teams and have never seen checked. Every question that arrives is listed with its answer, including the ones where the answer is no.