Which warning signs come before an injury, and who is showing them?
Do warning signs come before injuries?
Relative risk of an IL placement within 30 days, 2024–26. Bars: 95% intervals that resample pitchers, not outings; dashed: fewer than 10 injuries, an exact interval counting each injury once. Verdicts need 10+ distinct injuries.
Same rule as the live check: 8+ earlier outings that season, nothing in the 45 days after a return from the IL, and the primary fastball as of that date. A starter is a pitcher whose typical outing so far had 25+ primary fastballs (80% of flags fall on relievers, flagged at about the same rate). Flags use only earlier outings; the thresholds are round numbers, not fit. The season-end row counts arm placements after the regular season’s last day through May 31 (35 pitchers placed in between are left out).
How we use it
Starters: a velocity drop triggers a check, not a shutdown.
A 1 mph drop over three starts: 17% went on the arm IL within 30 days vs. 5% of the rest (3.6×, interval 1.68–6.13, 14 injuries). The trainer and pitching coach review it before his next start. Warnings come about 0.9 times a season per club for starters, a median 12 days before the placement (14 flagged starters).
Relievers: a watch item, not a warning.
1.4× with an interval of 0.87–2.00: not separable from noise. Measured against his own outing-to-outing spread it does no better (1.4×, 0.94–1.82). 80% of all flags fall on relievers, so an alarm on every one would mostly be false. Note it, watch the next outing, and don’t pull him on it.
Bigger drops are suggestive, not proven.
−1.5 mph: 2.4× on 9 injuries, interval 1.07–4.55 (exact, counting each injury once); −2.0 mph: 4.5× on 3 injuries, interval 0.93–13.40 (exact, counting each injury once). Under 10 injuries a row is suggestive at most. Escalate on the trainer’s read, not on the size of the drop.
A season-end drop is a look, not a forecast.
2 of 37 pitchers who finished with a drop went on the arm IL over the winter, vs. 16% of the rest: too few cases (2 injuries). A check before the throwing program is cheap, so do it; the data can’t say more.
Arm angle and spin are context.
Arm angle ±3°: 1.3×, interval 0.93–1.82; spin: 1.3×. Together with a velocity drop: 2.3× on 8 injuries, so suggestive at most; Chandler shows both. Bat speed doesn’t predict hitter injuries (1.0×).
What the warning catches
Arm injured-list placements, 2024–26: how many the check could see, and how many it flagged in the 30 days before.
MLB, every club
- 971 arm placements, hitters included
- 828 by pitchers (100+ MLB pitches 2024–26)
- 579 had pitched in the majors in the 30 days before
- 257 the check could see (8+ earlier outings, not just back from the IL)
- 39 flagged (15% of those it saw): starters 14 of 73, relievers 25 of 184
Pittsburgh
- 27 arm placements, hitters included
- 22 by pitchers (100+ MLB pitches 2024–26)
- 16 had pitched in the majors in the 30 days before
- 7 the check could see (8+ earlier outings, not just back from the IL)
- 0 flagged (0% of those it saw): starters 0 of 4, relievers 0 of 3
Of the 579 placements of pitchers active in the 30 days before, the check couldn’t see 322: 275 had fewer than 8 earlier outings that season (104 placed in March or April, 171 later in the season), 43 were within 45 days of an IL return, and 4 threw too few fastballs to check. Tested apart from the live rule: a prior-season baseline for those early outings (his previous season’s mean of the same fastball, plus the league’s early-outing gap fit on 2022–23) would see 117 of the 275 (58 of 104 placed in March or April) and flag 31; 84 had no prior season of that fastball, and 74 too few outings or a recent IL return. Its own signal: 1.4×, interval 0.95–1.94, 41 injuries: no clear signal, so the live check keeps its 8-outing rule. Alarm load: about 4.2 new flags a club a season, about 0.9 of them on starters. Lead time from the first flag to the placement: median 12 days (middle half 3–21); starters 12, relievers 12. The 30-day count includes some flags from the injury outing itself; leaving each pitcher’s last outing before the IL out, the all-pitcher signal is 1.6× (0.99–2.37) and starters’ 3.2× (1.41–6.57, 9 injuries).
By season, and 2026 alone
The −1.0 mph rule each season; 2026 scored on its own.
| Season | All pitchers | Starters | Relievers |
|---|---|---|---|
| 2024 | 1.0× 0.46–1.83No clear signal · 11 injuries | 0.9× 0.17–2.94Too few cases (3 injuries) · 3 injuries | 1.1× 0.45–2.29Too few cases (8 injuries) · 8 injuries |
| 2025 | 2.6× 1.29–4.29Strong signal · 14 injuries | 5.7× 1.91–14.37Suggestive (6 injuries) · 6 injuries | 1.8× 0.73–3.77Suggestive (8 injuries) · 8 injuries |
| 2026 | 2.0× 0.95–3.48No clear signal · 14 injuries | 6.0× 1.76–16.70Suggestive (5 injuries) · 5 injuries | 1.4× 0.56–2.55No clear signal · 10 injuries |
Each cell: relative risk, its 95% interval and its verdict; muted where the interval reaches 1 or the cases are too few. The pooled numbers are a time-ordered study, not a forecast made in advance: in 2 of 3 seasons (2024, 2026) the all-pitcher signal doesn’t hold on its own. 2026 alone: starters 6.0×, suggestive (5 injuries); relievers 1.4×, no clear signal. The same rule (−1.0 mph over three outings, 8+ earlier outings: round thresholds, not fit) scored on 2026 outings only. The starter/reliever split was found looking at 2024–26 together, so by role this is a check, not a clean forward test.
Pittsburgh pitchers · season-end check
Primary fastball by appearance, 2026. Dashed line = his season baseline; the dot is his final outing (Last 3 is the three-outing average the check uses).
| Pitcher | Fastball by outing | Baseline | Last 3 | Change | Arm angle Δ | Spin Δ | By season (context) | Signs | Injured list | 2027 |
|---|---|---|---|---|---|---|---|---|---|---|
| Bubba ChandlerStarter · 31 outings · last Sep 25 | 98.4 | 97.1 | −1.3 | −3.2° | −50 | 98.9 → 98.344° → 46° arm angle | Pre-program check Arm angle Both: suggestive | — | Rotation #4 | |
| Brandon EisertReliever · 33 outings · last Sep 25 | 90.0 | 89.1 | −0.9 | +4.7° | −34 | 91.2 → 89.7 → 90.036° → 32° → 35° arm angle | Arm angle | — | Bullpen | |
| Carmen MlodzinskiReliever · 32 outings · last Sep 20 | 95.1 | 94.8 | −0.4 | −3.3° | +44 | 96.2 → 96.1 → 95.030° → 32° → 34° arm angle | Arm angle | — | Bullpen | |
| Mason MontgomeryReliever · 54 outings · last Sep 27 | 98.6 | 98.6 | 0.0 | −6.9° | +95 | 97.3 → 98.7 → 98.554° → 53° → 57° arm angle | Arm angle | — | Bullpen | |
| Mitch KellerStarter · 22 outings · last Aug 2 | 93.4 | 92.6 | −0.8 | +0.2° | −37 | 94.5 → 94.1 → 93.333° → 38° → 35° arm angle | — | On the 60-day IL today (since Aug 3) | Rotation #5 | |
| Luke WeaverReliever · 52 outings · last Sep 22 | 95.4 | 94.7 | −0.7 | +0.6° | −32 | 95.7 → 95.1 → 95.347° → 48° → 48° arm angle | — | — | Bullpen | |
| Paul SkenesStarter · 32 outings · last Sep 24 | 96.9 | 96.4 | −0.5 | +0.9° | +50 | 98.8 → 98.2 → 96.824° → 24° → 27° arm angle | — | — | Rotation #1 | |
| Wilber DotelReliever · 22 outings · last Sep 25 | 98.3 | 97.9 | −0.4 | −0.7° | −35 | 98.3 | — | 15-day IL Jun 13–Jul 20 (Lat / teres) | Bullpen | |
| Yohan RamírezReliever · 34 outings · last Sep 26 | 96.4 | 96.0 | −0.4 | +1.0° | −40 | 94.4 SI → 96.4 FF → 96.3 SI22° → 26° → 22° arm angle | — | — | Bullpen | |
| Jared JonesStarter · 21 outings · last Sep 27 | 98.3 | 97.9 | −0.4 | −0.6° | −24 | 97.3 → 98.243° → 40° arm angle | — | 60-day IL Feb 11–May 29 (UCL / Tommy John) | Rotation #2 | |
| Gregory SotoReliever · 47 outings · last Sep 26 | 96.5 | 96.5 | 0.0 | −1.4° | −24 | 97.4 → 96.8 → 96.434° → 34° → 31° arm angle | — | — | Leaving (free agent) | |
| Lake BacharReliever · 44 outings · last Sep 27 | 94.5 | 94.7 | +0.2 | +2.4° | +25 | 94.2 → 94.8 → 94.435° → 38° → 41° arm angle | — | — | Call-up depth | |
| Isaac MattsonReliever · 35 outings · last Sep 5 | 93.9 | 94.6 | +0.7 | +1.6° | +39 | 94.0 → 93.9 → 94.249° → 49° → 46° arm angle | — | 15-day IL Sep 8–Sep 28 (Groin / hip) | Bullpen | |
| Braxton AshcraftStarter · 28 outings · last Sep 5 | 97.2 | 97.9 | +0.7 | −1.1° | −19 | 97.0 → 97.340° → 37° arm angle | — | 15-day IL Sep 10–Sep 28 (right arm discomfort) | Rotation #3 |
Back from the IL within 45 days, so not checked against the rule (their baseline straddles the injury): Kirby Yates (reliever, back Sep 22, −1.2 mph, a drop the rule would flag; leaving (free agent)); Evan Sisk (reliever, back Aug 5, −0.4 mph; bullpen).
Pre-program check: Chandler. That is the in-season warning for starters (17% vs. 5% on the arm IL within 30 days); read at season end it calls for a check before the throwing program, not a 30-day risk (the season-end test rests on 2 injuries: too few to say). Context only: Eisert (arm angle +4.7°), Mlodzinski (arm angle −3.3°) and Montgomery (arm angle −6.9°).
The Winter Plan’s medical ledger
Every player on the plan’s 26-man and its protected prospects with an injured list, a warning or a medical note (the plan’s health check), then players the winter moves out.
| Player | 2027 | What we see |
|---|---|---|
| Endy Rodríguez | bench, catcher | on the injured list since Aug 10 (moved to the 60-day IL Sep 1) (Left hip inflammation)4 injured-list stints since March 2025, 2 on the 60-day list: the current stint, above (60-day IL, left hip inflammation); Jul 8–Jul 27 (10-day IL, left glute strain); Jun 7, 2025–Nov 6, 2025 (60-day IL, right elbow inflammation); Apr 15, 2025–May 28, 2025 (10-day IL, right hand laceration)Base rate: 9% of hitters’ arm placements in 2024–25 were followed by another arm IL stint within a year (6 of 70; too few hitter elbow cases to split out). A league rate for hitters, not his own risk. |
| Bubba Chandler | rotation #4 | late-season velocity −1.3 mph and arm angle −3.2° against his baseline (last outing Sep 25) |
| Mitch Keller | rotation #5 | on the 60-day IL since Aug 3 (Right arm teres major muscle injury)Base rate: 31% of pitchers’ lat / teres placements in 2024–25 were followed by another arm IL stint within a year (11 of 35). A league rate for pitchers, not his own risk. |
| Termarr Johnson | protected, in the minors | on the minor-league injured list since Aug 9 (moved to the full-season IL Aug 18) |
| Rafael Flores Jr. | starter, C | 1 injured-list stint since March 2025: Jul 24–Aug 1 (7-day IL, concussion) |
| Ryan O'Hearn | starter, 1B | 2 injured-list stints since March 2025: Aug 1–Sep 11 (10-day IL, left quad strain); May 17–May 31 (10-day IL, right quad muscle strain) |
| Nick Gonzales | starter, 2B | 1 injured-list stint since March 2025, 1 on the 60-day list: Mar 28, 2025–Jun 3, 2025 (60-day IL, non-displaced left ankle fracture) |
| Isaac Paredes | starter, 3B | 1 injured-list stint since March 2025, 1 on the 60-day list: Jul 20, 2025–Sep 19, 2025 (60-day IL, right hamstring strain) |
| Konnor Griffin | starter, SS | 2 injured-list stints since March 2025, 1 on the 60-day list: Jul 7–Sep 4 (60-day IL, left ring finger sagittal band injury); May 31–Jun 26 (10-day IL, right elbow strain)Base rate: 9% of hitters’ arm placements in 2024–25 were followed by another arm IL stint within a year (6 of 70; too few hitter elbow cases to split out). A league rate for hitters, not his own risk. |
| Oneil Cruz | starter, CF | 2 injured-list stints since March 2025, 1 on the 60-day list: Jun 10–Aug 17 (60-day IL, left hand fourth and fifth metacarpal non-displaced fractures); Aug 13, 2025–Aug 26, 2025 (7-day IL, concussion) |
| Esmerlyn Valdez | starter, RF | 1 injured-list stint since March 2025: Sep 12–Sep 28 (10-day IL, left hamstring strain) |
| Spencer Horwitz | starter, DH | 2 injured-list stints since March 2025: Jun 25–Aug 2 (10-day IL, left hamstring strain); Mar 27, 2025–May 17, 2025 (10-day IL, right wrist surgery) |
| Jake Mangum | bench | 2 injured-list stints since March 2025: May 8–May 17 (10-day IL, left hamstring strain); Apr 24, 2025–May 30, 2025 (10-day IL, left groin strain) |
| Jared Triolo | bench | 2 injured-list stints since March 2025: Apr 5–May 2 (10-day IL, right knee patellar tendon injury); Apr 7, 2025–Apr 16, 2025 (10-day IL, lumbar spine strain) |
| Ronny Simon | bench | 1 injured-list stint since March 2025, 1 on the 60-day list: Aug 30, 2025–Nov 6, 2025 (60-day IL, dislocated left shoulder)Base rate: 13% of hitters’ shoulder placements in 2024–25 were followed by another arm IL stint within a year (5 of 40). A league rate for hitters, not his own risk. |
| Jared Jones | rotation #2 | 2 injured-list stints since March 2025, 2 on the 60-day list: Feb 11–May 29 (60-day IL); Mar 27, 2025–Nov 6, 2025 (60-day IL, right elbow UCL sprain)Base rate: 29% of pitchers’ UCL / Tommy John placements in 2024–25 were followed by another arm IL stint within a year (16 of 55). A league rate for pitchers, not his own risk. |
| Braxton Ashcraft | rotation #3 | 1 injured-list stint since March 2025: Sep 10–Sep 28 (15-day IL, right arm discomfort)Base rate: 34% of pitchers’ arm placements in 2024–25 were followed by another arm IL stint within a year (192 of 557; too few pitcher arm (general) cases to split out). A league rate for pitchers, not his own risk. |
| Wilber Dotel | bullpen | 1 injured-list stint since March 2025: Jun 13–Jul 20 (15-day IL, right lat muscle strain)Base rate: 31% of pitchers’ lat / teres placements in 2024–25 were followed by another arm IL stint within a year (11 of 35). A league rate for pitchers, not his own risk. |
| Luke Weaver | bullpen | 1 injured-list stint since March 2025: Jun 3, 2025–Jun 20, 2025 (15-day IL, left hamstring strain) |
| Evan Sisk | bullpen | 1 injured-list stint since March 2025: Jul 4–Aug 5 (15-day IL, left elbow inflammation)Base rate: 34% of pitchers’ elbow placements in 2024–25 were followed by another arm IL stint within a year (57 of 169). A league rate for pitchers, not his own risk. |
| Isaac Mattson | bullpen | 1 injured-list stint since March 2025: Sep 8–Sep 28 (15-day IL, right hip impingement) |
| Derek Diamond | protected, in the minors | 2 injured-list stints since March 2025, 1 on the 60-day list: Apr 2–May 23 (7-day IL); May 1, 2025–Sep 2, 2025 (60-day IL) |
| Kirby Yates | Leaving (free agent) | 5 injured-list stints since 2025: Sep 3 (15-day, right shoulder impingement, 19 days); Mar 24 (15-day, left knee inflammation, 40 days); Sep 24, 2025 (15-day, right hamstring strain, 4 days); Aug 1, 2025 (15-day, lower back pain, 22 days); May 18, 2025 (15-day, right hamstring strain, 20 days) |
| Brandon Lowe | Leaving (free agent, if he declines the QO) | 2 injured-list stints since 2025: Jul 22, 2025 (10-day, left ankle/foot tendinitis, 8 days); Jul 11, 2025 (10-day, left oblique tightness, 7 days) |
From the decisions feed: the plan reads the same injured lists and warnings. The plan’s innings already carry league-average attrition through its playing-time model; nothing here is a medical forecast for one player, because the diagnoses aren’t public. Health, all 26 on the 26-man and the protected prospects: 4 warnings (Endy Rodríguez, Bubba Chandler, Mitch Keller, Termarr Johnson); 18 more with injured-list history since March 2025
Not in the ledger, and context rather than a warning: Paul Skenes’s primary fastball (FF) is down 2.0 mph since 2024 (98.8 → 98.2 → 96.8), arm angle 24° → 27°; his decision brief weighs it; Mitch Keller’s primary fastball (FF) is down 1.2 mph since 2024 (94.5 → 94.1 → 93.3), arm angle 33° → 35°; his decision brief weighs it. The season-end check compares a pitcher with his own season, so it can’t see a drop across seasons, and a multi-season drop hasn’t been tested as a warning here.
Workload: the 2027 staff
Innings by season, majors and minors combined; 100+ pitch starts in 2026. Sorted by the jump from 2025.
| Pitcher | 2027 | 2024 IP | 2025 IP | 2026 IP | Change | Pitches | 100+ starts |
|---|---|---|---|---|---|---|---|
| Jared Jones | Rotation #2 | 122 | — | 120 | Back from a lost 2025 | 1,924 | — |
| Noah Murdock | Call-up | — | 38 | 55 | +46% | 944 | — |
| Braxton Ashcraft | Rotation #3 | — | 118 | 165 | +40% | 2,413 | 1 |
| Yohan Ramírez | Bullpen | 45 | 64 | 85 | +32% | 1,411 | — |
| Mason Montgomery | Bullpen | 10 | 55 | 69 | +25% | 1,169 | — |
| Evan Sisk | Bullpen | — | 55 | 67 | +22% | 1,210 | — |
| Bubba Chandler | Rotation #4 | — | 131 | 154 | +17% | 2,686 | — |
| Lake Bachar | Call-up | 9 | 74 | 87 | +17% | 1,448 | — |
| Carmen Mlodzinski | Bullpen | 51 | 115 | 123 | +7% | 2,080 | 1 |
| Antwone Kelly | Call-up | — | 107 | 110 | +2% | 1,831 | — |
| Brandon Eisert | Bullpen | 7 | 72 | 71 | −1% | 1,246 | — |
| Khristian Curtis | Call-up | — | 111 | 107 | −3% | 1,866 | — |
| Luke Weaver | Bullpen | 84 | 65 | 60 | −7% | 903 | — |
| Paul Skenes | Rotation #1 | 133 | 188 | 174 | −7% | 2,890 | 9 |
| Isaac Mattson | Bullpen | 5 | 69 | 59 | −14% | 1,050 | — |
| Mitch Keller | Rotation #5 | 178 | 176 | 116 | −34% | 1,887 | 2 |
| Wilber Dotel | Bullpen | — | 126 | 81 | −35% | 1,373 | — |
Jones (120 innings after 2025 lost to the IL, UCL / Tommy John) came back from a lost season, a bigger change than any percentage. Murdock (+46%), Ashcraft (+40%), Ramírez (+32%) and Montgomery (+25%) threw a quarter more innings than the year before. 8 organization pitchers with 100+ innings jumped 25% or more. Whether a jump like that comes before injuries can’t be tested yet: our minor-league innings start in 2025, so there is one jump and no season after it (2024 is majors only). Relievers’ back-to-back cost is on Bullpen Strategy.
Where injuries happen · MLB 2024–26
Placements by body part; label = median regular-season days missed in the placement’s season (a stint into the next season counts only to this one’s end).
- Shoulder 338 46d
- Elbow 310 65d
- Hand / wrist / finger 215 26d
- Oblique / core 195 29d
- Back / neck 177 22d
- Hamstring 166 24d
- Groin / hip 142 28d
- Forearm / flexor 138 43d
- Knee 134 33d
- Lower leg / ankle / foot 122 29d
- UCL / Tommy John 88 143d
- Not stated 76 43d
- Calf / Achilles 54 29d
- Lat / teres 52 61d
- Head / concussion 51 9d
- Biceps / triceps 34 36d
- Other 32 20d
- Illness 28 15d
- Quad / thigh 28 28d
- Chest / upper back 13 44d
- Arm (general) 11 29d
Shaded = arm injuries (Shoulder, Elbow, Forearm / flexor, UCL / Tommy John, Lat / teres, Biceps / triceps, Arm (general)). 76 placements gave no reason; we label them “Not stated” rather than guess.
Injured-list stints, 2025–26
Every stint Pittsburgh filed, and current Pirates’ stints with other clubs. Injury text is MLB’s; days are regular-season days. 20 of the 2026 stints are Pittsburgh’s own.
| Player | Club | Placed | Injury | IL | Days |
|---|---|---|---|---|---|
| Esmerlyn Valdez | PIT | Sep 12, 2026 | Hamstring Left hamstring strain | 10 | 15 |
| Braxton Ashcraft | PIT | Sep 10, 2026 | Arm (general) Right arm discomfort | 15 | 17 |
| Isaac Mattson | PIT | Sep 8, 2026 | Groin / hip Right hip impingement | 15 | 19 |
| Kirby Yates | PIT | Sep 3, 2026 | Shoulder Right shoulder impingement | 15 | 19 |
| Endy Rodríguez | PIT | Aug 13, 2026 | Groin / hip Left hip inflammation | 60 | 45+ |
| Mitch Keller | PIT | Aug 3, 2026 | Lat / teres Right arm teres major muscle injury | 60 | 55+ |
| Ryan O'Hearn | PIT | Aug 1, 2026 | Quad / thigh Left quad strain | 10 | 41 |
| Rafael Flores Jr. | PIT | Jul 24, 2026 | Head / concussion Concussion | 7 | 8 |
| Endy Rodríguez | PIT | Jul 8, 2026 | Groin / hip Left glute strain | 10 | 19 |
| Konnor Griffin | PIT | Jul 7, 2026 | Hand / wrist / finger Left ring finger sagittal band injury | 60 | 59 |
| Evan Sisk | PIT | Jul 4, 2026 | Elbow Left elbow inflammation | 15 | 32 |
| Spencer Horwitz | PIT | Jun 25, 2026 | Hamstring Left hamstring strain | 10 | 38 |
| Wilber Dotel | PIT | Jun 13, 2026 | Lat / teres Right lat muscle strain | 15 | 37 |
| Oneil Cruz | PIT | Jun 10, 2026 | Hand / wrist / finger Left hand fourth and fifth metacarpal non-displaced fractures | 60 | 68 |
| Konnor Griffin | PIT | May 31, 2026 | Elbow Right elbow strain | 10 | 26 |
| Ryan O'Hearn | PIT | May 17, 2026 | Quad / thigh Right quad muscle strain | 10 | 14 |
| Jake Mangum | PIT | May 8, 2026 | Hamstring Left hamstring strain | 10 | 9 |
| Chris Devenskino longer a Pirate | PIT | May 7, 2026 | Illness Undisclosed illness | 60 | 90 |
| Jared Triolo | PIT | Apr 5, 2026 | Knee Right knee patellar tendon injury | 10 | 27 |
| Kirby Yates | LAA | Mar 24, 2026 | Knee Left knee inflammation | 15 | 40 |
What this can’t see
Medical records, imaging, pain and soreness reports, bullpen and long-toss workloads, sleep and recovery data, and biomechanics are all private. Minor-league innings before 2025 are public (the MLB Stats API’s minor-league season totals), but we haven’t pulled them yet, so workload jumps can’t be tested against injuries yet; we don’t estimate them. See what we’d add.