Where does every number come from?
Sources
80 raw datasets, listed with seasons, rows and fetch times in the pipeline’s manifest (Oct 21, 2026).
| Source | What we pull | Used for | As of |
|---|---|---|---|
| MLB Stats API | Rosters DSL→MLB, every club’s opening-day 40-man 2021–26, bios, season, postseason and minor-league stats (season totals), standings and team totals 2021–26, splits, transactions league-wide, draft | Org depth, player pages, the playing-time model and its team-wins backtest, minor-league lines, the injured-list model, the Season Review | Oct 20, 2026 – Oct 21, 2026 |
| Baseball Savant (Statcast) | Pitch-level tracking 2021–26 (MLB), Triple-A and FSL tracking, bat tracking, ABS challenges with each pitch’s original call, Outs Above Average by position and direction, Fielding Run Value, outfield jump and catch probability, arm strength, catcher blocking and throwing, park factors by batter side, sprint speed, baserunning, basestealing and pitchers’ running game | Keystone Stuff, swing decisions, expected outs and the defense map, catcher framing, fielding by position, the depth chart’s gloves, PNC Park Fit, the Running Game, ABS policy, projection tests | Oct 20, 2026 – Oct 21, 2026 |
| FanGraphs | WAR and season lines (2005–26), minor-league leaders, The Board (FV/ETA only), RosterResource contracts, service time and team payroll (CBT and benefits estimates), and the published prospect-value table | Projections and their 13-season backtest, aging curve, surplus value, payroll vs. the proposed floor, prospect grades and trade values, Rule 5 queue | Oct 20, 2026 |
| Chadwick Bureau / SFBB | Player ID crosswalks | Joining sources on MLBAM ID | Oct 20, 2026 |
| Public reporting | CBA proposals and calendar, arbitration projections, front-office context: each figure is in the reference table below, each claim about the organization cited where it appears | Product Thinking evidence, the decision calendar, the owners’ lens | See below |
Coverage by level · Pittsburgh Pirates system
Public data only. Cells marked — are where internal tracking would plug in.
| Level | Affiliate | Season stats (2025–26) | Pitch tracking | Bat tracking | ABS | Fielding |
|---|---|---|---|---|---|---|
| MLB | Pittsburgh | Yes | Yes | Yes | Yes | Yes |
| AAA | Indianapolis Indians | Yes | Yes>99% of 150 games’ pitches | — | Yes | Data only |
| AA | Altoona Curve | Yes | — | — | — | — |
| A+ | Greensboro Grasshoppers | Yes | — | — | — | — |
| A | Bradenton Marauders | Yes | Partial95% of 127 games’ pitches | — | — | Data only |
| FCL | FCL Pirates | Yes | — | — | — | — |
| DSL | DSL Pirates Black / | Yes | — | — | — | — |
Minor-league lines are season totals, not game logs. Every Triple-A park has Hawk-Eye. Below Triple-A, pitch tracking is public only in Florida State League (Single-A) parks with Hawk-Eye, and not every FSL park has it, so an FSL affiliate’s road games in an untracked park are dark (“Partial”: the share is of its 2026 pitches, home and road, that Baseball Savant marks tracked). 10 of the 30 clubs have an FSL affiliate; the rest have no tracking below Triple-A. High-A and Double-A, the complex leagues and the DSL are dark for every club: 11 of Pittsburgh’s 24 ranked top-30 prospects still in the system are at High-A or Double-A today. Complex Hawk-Eye, academy TrackMan or KinaTrax would fill them. Fielding: the majors have Savant’s Outs Above Average and Keystone’s zone model; Triple-A and tracked FSL parks publish batted balls with fielder IDs (“Data only”), which Keystone doesn’t model yet, so prospects’ gloves fall back on a positional prior.
Publishing rules
- Derived aggregates only. No raw pitch-level dumps are republished (MLBAM terms: individual, non-commercial use).
- Keystone builds its own models. FanGraphs Stuff+ is used only to check our model, never displayed. FanGraphs scouting text is never republished.
- Every claim about the organization is a hypothesis drawn from public evidence, cited where it appears.
- Small samples are labeled and shrunk toward the average where decisions depend on them: ABS accuracy, catcher framing, hitters’ true talent and platoon splits, matchup edges, rival pitch-type splits, and fielding (three seasons, regressed). Each model card names its method.
How to read color
On an opponent’s page (Rival Intel) the rule holds, so warm marks a rival’s weakness. On player pages warm-cool matches Baseball Savant’s percentile colors. Pitch-type colors were chosen by a color-vision-deficiency validator.
Outside figures the site uses
Numbers we don’t compute, with where each came from. If one changes, this is the list to check.
| Figure | Value | Source | When | Used by |
|---|---|---|---|---|
| Owners’ proposed payroll floor and cap (a proposal, not a rule) | $171.2M floor, $245.3M cap, benefits included | NBC Sports (AP) | Reported in 2026 CBA talks; in our notes Oct 20, 2026 | Contract Studio, Winter Plan, Free-Agent Board (the owners’ lens) |
| Qualifying offer, 2026–27 | $23.15M | Just Baseball (also Yardbarker), reporting MLB’s figure | In our notes Oct 20, 2026 | Qualifying-offer brief, Contract Studio |
| Rule 5 protection and non-tender deadlines | Rule 5 protection Nov 17, 2026; non-tender Nov 20. ⚠︎ Some sources give a day later for each; we use the earlier dates | Just Baseball, Yardbarker and other offseason calendars | In our notes Oct 20, 2026 | Decision calendar, the Rule 5 and non-tender briefs, the 40-man at each deadline |
| Pittsburgh’s 2027 arbitration projections (10 players) | Used only as a check on our own estimates | Pittsburgh Baseball Now | Oct 2026 | Track Record (agreement test), Contract Studio check |
| Prospect value by Future Value grade | As published for prospects without major-league time; blended with Keystone’s surplus engine for players with a major-league record (weighted by how much of the projection is his own); 2026 draftees carry a provisional draft-curve value | FanGraphs (Ben Clemens, “Introducing an Updated Method for Prospect Valuation”) | Jul 2, 2026 | Trade Builder, Prospect Pipeline, every trade package |
| Major-league minimum salary today | $780K (2026). 2027 under today’s rules is our assumption: no agreement sets it yet | The 2022–26 Basic Agreement (MLB and the MLBPA) | Set in the 2022 agreement | Contract Studio, Winter Plan payroll, every pre-arbitration salary |
| The owners’ other proposed terms | $1M minimum salary; contract-length caps (the terms aren’t in our sources); an international draft in place of bonus pools | CBS Sports; CNBC | Reported Oct 2026; in our notes Oct 20, 2026 | The owners’ lens notes in Contract Studio |
| World Series and Winter Meetings | World Series opens Oct 23 (Brewers vs. Rays); Winter Meetings Dec 7–10, Orlando, where the Rule 5 draft closes | Just Baseball and other offseason calendars | In our notes Oct 20, 2026 | Decision calendar, the Rule 5 brief |
| CBA expiry and expected lockout | Dec 1, 2026; lockout expected Dec 2 | CBS Sports | In our notes Oct 20, 2026 | Decision calendar, every brief’s deadline |
What we’d add with more data
Analyses we want but can’t do honestly with public data, plus any marked “public data, not modeled yet”, a to-do of ours rather than a data gap. We list them instead of estimating them, the ones that could change this winter’s calls first.
- MedicalHealth & Workload
Missing Medical records, imaging, pain and soreness reports, rehab progress. We see injured-list moves, not diagnoses, so when an injured player returns is an assumption in the Winter Plan, not a medical read. The Winter Plan assumes that Endy Rodríguez and Mitch Keller, who ended 2026 on an injured list, are ready for spring and that Termarr Johnson’s medicals clear by midseason (101 PA as a call-up).
Would unlock True injury-risk models; return-to-play timelines; medical inputs to the qualifying-offer, extension, Keller and catching calls
Could change Extension: Paul Skenes (≈ $112M over 5 years at comparable prices); Mitch Keller: keep, trade, or add a starter ($18.4M 2027 salary); Qualifying offer: Brandon Lowe ($23.15M one-year offer); Tender or non-tender, and the third catcher (Henry Davis non-tendered); Rule 5 protection (3 protected of 54 eligible; 9 at notable pick risk left exposed)
- BiomechanicsPitch Lab · Development
Missing Motion capture, force plates, high-speed video of the delivery and swing
Would unlock Whether a pitch design is physically safe for a given arm; mechanical causes of velocity loss; swing changes tied to movement patterns
Could change Extension: Paul Skenes (≈ $112M over 5 years at comparable prices); Mitch Keller: keep, trade, or add a starter ($18.4M 2027 salary)
- Makeup and fitAll modules
Missing Scouting interviews, coach evaluations, clubhouse context
Would unlock Every player decision; today they’re outside every number on this site
Could change Fix third base ($13.35M salary and $16.4M of prospects); Extension: Paul Skenes (≈ $112M over 5 years at comparable prices)
- Other clubs’ draft-pick statusFree-Agent Board · Draft Value
Missing Each club’s revenue-sharing and luxury-tax status (its qualifying-offer forfeiture class); the 2027 draft order and Competitive Balance picks, set by MLB later; trades of picks (our data lists who made a pick, not who traded one)
Would unlock What a qualifying-offer signing costs another club; the exact pick Pittsburgh would forfeit or receive (its 2026 Competitive Balance pick is inferred)
Could change Qualifying offer: Brandon Lowe ($23.15M one-year offer); Spend to the floor ($34.0M under the proposed floor)
- Outside free-agent price forecastsFree-Agent Board
Missing A published forecast of this winter’s free-agent contracts to check the board’s prices against. None is in our data, so the price model is checked only against last winter’s deals, held out one at a time.
Would unlock An outside benchmark for every price on the Free-Agent Board before the market sets them, starting with the prices the biggest calls rest on (Brandon Lowe’s, behind the qualifying-offer decision, and the floor list’s picks)
Could change Qualifying offer: Brandon Lowe ($23.15M one-year offer); Spend to the floor ($34.0M under the proposed floor)
- Prior prospect grades for graduated playersFree-Agent Board
Missing FanGraphs grades from before a player graduated (old Future Value): our board carries current prospects only. A first-round draft record stands in for a former top prospect, so international signings, who have no draft record, aren’t recognized. Years of club control, which the contract and service-time data do give, could stand in instead; Keystone doesn’t use them for this yet.
Would unlock Shopping lists that never release a graduated top prospect of any origin to make 40-man room (they protect first-round picks still under club control)
Could change Spend to the floor ($34.0M under the proposed floor)
- Club finances2027 Outlook
Missing Revenue per win, playoff revenue, attendance and media effects
Would unlock The win curve in dollars: exactly what the 85th win is worth to Pittsburgh
Could change Spend to the floor ($34.0M under the proposed floor)
- High-A and Double-A trackingProspect Pipeline · Call-Up Readiness
Missing Pitch and bat tracking at High-A (Greensboro) and Double-A (Altoona), and in the complex league and the DSL. Triple-A and Florida State League parks are tracked, Bradenton included.
Would unlock Stuff, Location and swing metrics for prospects in the two-level gap between Bradenton and Indianapolis, where push-or-hold calls get made
Could change Rule 5 protection (3 protected of 54 eligible; 9 at notable pick risk left exposed); Fix third base ($13.35M salary and $16.4M of prospects)
- Minor-league fielding Public data, not modeled yetFielding Lab
Missing Fielder positioning, catch probability and Outs Above Average below the majors. Triple-A and FSL batted balls with fielder IDs are public, but Keystone doesn’t model them yet.
Would unlock Glove ratings for prospects and minor-league depth (a utility infielder at third or short); today they fall back on a positional prior
Could change Fix third base ($13.35M salary and $16.4M of prospects); Rule 5 protection (3 protected of 54 eligible; 9 at notable pick risk left exposed)
- Updated prospect gradesTrade Builder
Missing Post-season Future Value grades, and FanGraphs grades for the 2026 draft class
Would unlock Grades that reflect the 2026 season (today’s are from before it), and graded values for this year’s draftees: until then each 2026 draftee carries a provisional value, Keystone’s draft curve at his pick scaled to FanGraphs’ table on the 2025 class, with a bootstrap range
Could change Fix third base ($13.35M salary and $16.4M of prospects)
- Trade marketTrade Builder
Missing What other clubs are asking for and offering
Would unlock Which Trade Builder deals are actually available, and what a club would really ask for a target
Could change Fix third base ($13.35M salary and $16.4M of prospects); Mitch Keller: keep, trade, or add a starter ($18.4M 2027 salary)
- Options and service timeRoster Builder
Missing Official minor-league option counts, service time and waiver status. Option years are estimated from transactions, and shown as unknown when a player’s 40-man time began before our records.
Would unlock Which demotions would expose a player to waivers; exact costs of 40-man cuts in the Roster Builder
Could change Rule 5 protection (3 protected of 54 eligible; 9 at notable pick risk left exposed); Tender or non-tender, and the third catcher (Henry Davis non-tendered)
- Minor-league contractsCall-Up Readiness
Missing Minor-league contract terms: length, opt-outs, minor-league free-agent dates
Would unlock Who becomes a minor-league free agent in November (estimated today), and who must be re-signed to keep depth
Could change Tender or non-tender, and the third catcher (Henry Davis non-tendered); Rule 5 protection (3 protected of 54 eligible; 9 at notable pick risk left exposed)
- Rule 5 selectionsProspect Pipeline
Missing Who clubs took in past Rule 5 drafts and what those players looked like when taken
Would unlock A pick-risk model fit to real selections; today each prospect’s pick risk is a rule, not fit to past selections
Could change Rule 5 protection (3 protected of 54 eligible; 9 at notable pick risk left exposed)
- Challenge confidenceABS Challenge Coach
Missing How sure a player was on each close call, including the ones he let go (only challenged pitches are in the public log)
Would unlock Whether challenging more would pay: the value of the overturnable calls clubs never challenged
Could change 2027 ABS challenge policy
- Workload off the fieldHealth & Workload
Missing Bullpen sessions, long toss, weighted-ball and side work; wearables (sleep, heart-rate variability)
Would unlock Acute-to-chronic workload and recovery models; the full picture behind velocity drops
- Workload historyHealth & Workload
Missing Minor-league innings before 2025. These are public: the MLB Stats API’s minor-league season totals, which Keystone pulls for 2025–26 only. Not built yet rather than out of reach.
Would unlock A test of whether an innings jump comes before arm injuries (today there is one measured jump and no season after it), and workload lines for pitchers who came up through the minors before 2025
- MLB ids for 13 board prospectsTrade Builder · Opportunity Finder · Prospect Pipeline
Missing An id crosswalk for the FanGraphs board rows that carry no MLB id: Randy Arias (HOU), Micah Bucknam (TOR), Carlos Castillo (LAA), Connor Cooke (TOR), Joe Elbis (AZ), Sean Episcope (MIL), Joshua Flores (MIL), Ruben Gallego (AZ), River Hamilton (DET), Griffin Hugus (SEA), Peter Kussow (NYM), Dahian Santos (TOR), Neomar Urbina (PIT)
Would unlock Their values in the trade pools and packages (Pittsburgh’s Neomar Urbina, a 35+ reliever in the DSL, is $1.5M on FanGraphs’ table) and their grades on /prospects and their player pages; until then they aren’t in any pool
- Amateur dataDraft Value
Missing College and high-school tracking (TrackMan, Hawk-Eye at showcases), amateur performance databases, scouting reports
Would unlock Valuing individual draft prospects, not just draft slots
- International amateursDraft Value
Missing International signing bonuses by player; the 2027 international bonus pools (not yet published); amateur tracking
Would unlock International signing strategy and pool allocation for the DSL pipeline
- Intended locationPitch Lab
Missing The catcher’s target on each pitch
Would unlock Separating a bad plan from a missed spot in the command model