clock menu more-arrow no yes

Filed under:

Statistics versus the eye test: Starting pitchers

New, 26 comments

In which the bullpen gets a break.

Minnesota Twins v Oakland Athletics
And this was just the start of July.
Photo by Thearon W. Henderson/Getty Images

There are two kinds of statistics, the kind you look up and the kind you make up.
-Archie Goodwin, Death of a Doxy

While the Twins have remained an above-.500 team in June and July, their division lead has dropped from its peak of 11 games to two at the time of writing. While Cleveland has gone on a tear through the summer, Minnesota has slipped a bit from their dominant spring, with a winning percentage of .531 since June 1 compared to .679 through the end of May.

With team brass looking to upgrade both the rotation and bullpen, it’s worth a dive into the pitching staff to see if statistics back up the struggle suggested by the eyes. The bullpen’s been discussed through and through, so let’s go at the rotation.

The Eye Test

In recent months, watching Twins games has suggested to me that Twins starters are not lasting as long into games, and they’re throwing more balls than in the spring months. To test whether the eye test has been correct, I’ve selected several statistics (all from Fangraphs) to examine from month to month.

  • Starts - for sample size.
  • Innings pitched per start - part of the eye test hypothesis. (In the tables below, this statistic is rounded to the nearest third of an inning.)
  • Strike percentage - part of the eye test hypothesis.
  • K%, BB%, K/BB - to provide additional evidence for or against the second part of the hypothesis, seeing whether changes have resulted in outs or baserunners.
  • ERA - for seeing if changes have resulted in starters allowing more runs.
  • BABIP - for seeing if any changes can be attributed to bad luck.

Let’s dive in.

Rotation stats - March/April

Player Starts IP/Start K% BB% K/BB Strike % ERA BABIP
Player Starts IP/Start K% BB% K/BB Strike % ERA BABIP
José Berríos 6 6.2 26.5% 5.2% 5.13 66.7% 2.97 .245
Jake Odorizzi 6 5.0 25.0% 9.2% 2.73 61.7% 3.34 .263
Michael Pineda 6 5.0 17.5% 4.8% 3.67 67.6% 6.21 .341
Kyle Gibson 5 5.2 20.3% 6.8% 3.00 59.3% 4.88 .272
Martín Pérez 3 6.0 15.1% 4.1% 3.67 67.6% 3.00 .304
Kohl Stewart 1 6.0 3.7% 11.1% 0.33 57.1% 7.50 .286
Totals 27 5.2 20.8% 6.3% 3.31 64.1% 4.21 .284

Rotation stats - May

Player Starts IP/Start K% BB% K/BB Strike % ERA BABIP
Player Starts IP/Start K% BB% K/BB Strike % ERA BABIP
José Berríos 6 6.1 19.0% 3.7% 5.17 69.3% 3.58 .328
Jake Odorizzi 5 5.2 28.4% 8.3% 3.44 64.9% 0.94 .209
Michael Pineda 5 6.0 23.3% 4.2% 5.60 66.5% 4.50 .205
Kyle Gibson 5 6.0 30.0% 5.0% 6.00 64.5% 3.34 .338
Martín Pérez 6 5.2 25.2% 11.2% 2.25 63.0% 3.15 .273
Kohl Stewart 1 6.0 8.0% 8.0% 1.00 63.3% 4.50 .250
Devin Smeltzer 1 6.0 35.0% 0.0% 7.00 76.8% 0.00 .231
Totals 29 6.0 24.4% 6.3% 3.89 65.9% 3.08 .274

Rotation stats - June

Player Starts IP/Start K% BB% K/BB Strike % ERA BABIP
Player Starts IP/Start K% BB% K/BB Strike % ERA BABIP
José Berríos 5 7.0 21.6% 4.3% 5.00 69.7% 2.06 .242
Jake Odorizzi 5 5.1 29.5% 6.3% 4.71 67.7% 3.95 .343
Michael Pineda 5 5.2 22.3% 3.6% 6.25 70.3% 3.58 .325
Kyle Gibson 5 6.0 21.3% 7.1% 3.00 58.0% 4.60 .314
Martín Pérez 5 5.2 18.6% 8.1% 2.30 63.8% 5.08 .273
Devin Smeltzer 1 6.1 7.7% 11.5% 0.67 59.8% 7.11 .059
Lewis Thorpe 1 5.0 36.8% 10.5% 3.50 67.1% 3.60 .444
Totals 27 6.0 22.3% 6.2% 3.59 66.5% 3.91 .289

Rotation stats - July

Player Starts IP/Start K% BB% K/BB Strike % ERA BABIP
Player Starts IP/Start K% BB% K/BB Strike % ERA BABIP
José Berríos 4 5.2 19.4% 9.7% 2.00 66.5% 3.18 .309
Jake Odorizzi 4 4.1 16.9% 10.8% 1.56 60.3% 9.35 .321
Michael Pineda 4 6.0 25.3% 7.1% 3.57 62.3% 2.59 .281
Kyle Gibson 6 4.2 23.6% 7.9% 3.00 61.3% 3.77 .317
Martín Pérez 4 5.2 15.5% 8.3% 1.88 65.4% 5.32 .309
Totals 22 5.1 20.4% 8.6% 2.36 63.1% 4.54 .307

It’s clear looking at these tables that statistics back up this instance of the eye test, especially at the top of the rotation. As a whole, Twins starters have averaged worst in each of these statistics in July, in most cases by a wide margin. Notably, their monthly 8.6% walk rate is 2.3 percentage points worse than their next highest, their 4.54 ERA is highest by 0.33, their 63.1% strike rate is worst by a full percentage point... I could list the rest of these stats, but I think the picture is clear.

With such a significant team-wide drop, these trends should and can be noticed in the players as well, none more noticeably than Jake Odorizzi. In July, Odorizzi has averaged less than five innings pitched in his starts, striking out less than twice as many batters as he’s walked, and posted a nasty 9.35 ERA.

However, all is not lost. The Twins’ .307 BABIP allowed was also significantly their worst month, 18 percentage points above the next month, indicating their July starting struggles have been a combination of bad pitching and bad luck. Twins’ starters have had excellent months (May) and dreadful ones (July), both in controllable and uncontrollable statistics, so it shouldn’t be thought that the trends this month are irreversible. If they clean up the bad pitching, they can minimize the chances of bad-luck hits.

That said, this does clarify why Falvine, Esquire, have been looking to acquire a starting pitcher; adding a player to the rotation who can consistently throw strikes and get batters out (oh, were it that simple) would at minimum kick the rotation’s statistics as a whole onto a better track. At most, it could help spark a turnaround in the rest of the rotation. But as the deadline looms, it looks like that spark will have to come from the existing clubhouse - and with a return to May’s pitching, it can.