The organization is the third bet. And it belongs to the org, not the player.
The same player carries different odds in different organizations, because player-development competence varies more in baseball than in any sport, and it is a property of the org. The engine grades the program as a developer and a destination: it measures development as a confidence-gated residual against the curves players were projected for on entry, separates developing talent from acquiring it, and keeps the axes apart, because the same prospect is a different wager in a factory than in a meat grinder. The institutional read adjusts a projection and its confidence, and it never touches a current KR.
Realized against projected, read as the org's signal.
Organizational development is measured as a residual: the engine compares the org's players' realized development against the trajectories they were projected for on entry, across a sample, and reads the persistent over- or under-performance as the org's development signal. An org that consistently beats its projected curves is a real, quantified strength; one whose players stall is a real weakness.
A +3.6 KR pitching-development residual at 63% confidence is a real factory signal that lifts every arm the org acquires; the +1.2 hitting residual is milder and less certain, and reported separately. Both adjust the projection and its band, and neither moves a player's current KR. The residual re-prices the projection, model-derived and versioned, and never the current rating.
Illustrative engine read on the real organizational development residual (realized development against projected entry trajectories, the pitching-factory and hitting-factory read, confidence-gated and model-derived). Composite org, demonstration figures.
Two different skills, plotted apart.
Developing players and acquiring them are different skills, and an org can be strong at one and weak at the other. The engine splits the two: a program that drafts and signs well but stalls talent is a different institution from one that develops what it lands, and the two reads are reported separately, never blended into one.
drafts well, stalls talent
lands less, develops what it has
Org A is an 82 acquirer and a 48 developer; Org B is a 55 acquirer and an 84 developer. Blend them and both look average; keep them apart and you can see that one wastes talent and the other reclaims it, which is the difference a scout is paid to know. Drafting well and developing well are different skills, and the engine refuses to average them into one.
Illustrative engine read on the real development-versus-acquisition split (two independent skills plotted separately, the two reads never blended). Composite orgs, demonstration figures.
A safe destination and a real factory are different bets.
The program is rated across axes that stay separate: developmental value, especially the pitching-development and hitting-development competence, institutional stability, and resource and payroll posture relative to the tax, plus, on the amateur side, scouting and international infrastructure. They never collapse into a single number.
A rich, stable org that cannot develop and a poor, unstable one that can would grade about the same on a single scale, and be completely different bets. The axes stay apart so the read tells you which org you are actually joining or drafting into. A scalar would hide the difference between a safe destination and a real factory, so the engine keeps the axes apart.
Illustrative engine read on the real organizational axes (developmental value, institutional stability, resource and payroll posture, amateur scouting and international infrastructure), kept separate and never collapsed. Composite orgs, demonstration figures.
The institution is the third bet, and it is the org's property, not the player's. The engine measures development as a residual against the curves players were projected for, separates developing from acquiring, and keeps the axes apart, because the same prospect is a different wager in a factory than in a meat grinder, and the honest read says which one you are betting on. The residual is model-derived, versioned, and confidence-gated, and it adjusts the projection and its band without ever touching a current KR.
Grade the developer. Grade the destination.
Program measures organizational development as a confidence-gated residual, separates the skill of acquiring talent from the skill of developing it, and keeps the axes apart, so the same prospect is priced as the different bet he actually is in each building.