Caddie

The caddie is graded on the decisions he adds. Not the bag he carries.

The caddie is a named actor whose product is on-course decisions. The engine grades him on the contribution he adds over a baseline, what the golfer would decide alone, and whether that edge travels across golfers, not on the reputation of the bag. It separates his decision-and-management value from the golfer's own game, holds it at the wide confidence a hard-to-isolate signal deserves, and never lets the read touch a golfer's KR.

Case 01 · the on-course contribution

Small, inferred, and stated plainly.

The caddie's product is on-course decisions: the reads and the club-and-target and risk-selection support, the same in-round decision set the round-ops engine surfaces, plus the green reads, the wind and lie judgment, and the course knowledge. The engine reads that contribution against a baseline, what the golfer would decide alone, and it carries its data tier and both confidence dials.

The caddie
The product is on-course decisions, not shots. The contribution is read against a baseline, what the golfer would decide alone, and it is small, largely inferred, and held at wide confidence, because a caddie's effect is one of the hardest signals in the game to isolate. The engine states that plainly rather than overselling it.
+0.42strokes / round over baselineread conf 43%variance Wide
Green reads+0.12
INFERREDLine and speed on the greens, reasoned from outcomes, hard to separate from the golfer's own read.
Club and target+0.1
INFERREDThe club-and-target choice for the shot in front, the same decision set the round-ops engine surfaces.
Risk selection+0.09
PROXYWhen to attack and when to bail, partly visible in the shot pattern and the situations.
Wind and lie judgment+0.06
INFERREDReading the conditions and the lie, a judgment call the data rarely captures cleanly.
Course knowledge+0.05
INFERREDLocal knowledge of the misses and the pins, accumulated and hard to quantify.

The numbers are small on purpose, and mostly inferred. A caddie's effect is largely hidden inside the golfer's own decisions and hard to isolate, so the engine reads it conservatively, tags most of it INFERRED, and holds it at wide confidence rather than dressing up a precise figure the data cannot support. A caddie's contribution is small, largely inferred, and stated plainly, never oversold.

Illustrative engine read on the real caddie-contribution structure (the on-course decision domains read against a baseline, each tiered, both dials carried, mostly INFERRED). Composite caddie, demonstration figures.

Case 02 · separating the caddie from the golfer's game

Credit the decisions. Never the shots he struck.

The caddie's contribution is not the golfer's ball-striking, and the read separates them. A great ball-striker with poor course management can gain real strokes from a caddie who fixes the decisions, while an elite decision-maker gains little on that axis. The engine credits the caddie for the decision-and-management gain, and never for the shots the golfer struck. Pick a golfer and watch the gain move.

Ball-striking the golfer's, not the caddie's92
Course management the golfer's baseline58
Decision-and-management gain the caddie can add
+0.58
He strikes it beautifully but leaks strokes to bad decisions, so a caddie who fixes the management adds real value. None of it is credited for the shots he struck.

The gain lives entirely in the decision-and-management headroom, the gap between what the golfer would decide alone and the optimal call, and it shrinks to almost nothing for a golfer who already decides well. The read is confidence-gated and conservative, because on-course decision value is one of the hardest signals to measure, and the engine will not manufacture a precise caddie number where the data cannot support one. Credit the decision-and-management gain, never the shots the golfer struck.

Illustrative engine read on the real separation structure (the caddie gain as decision-and-management headroom, held apart from the golfer's ball-striking, conservative and confidence-gated). Composite golfers, demonstration figures.

Case 03 · the portability test

A real edge travels. A partnership does not.

The engine runs the caddie through the portability test: does the contribution follow him across golfers, or did one strong partnership flatter him. A real caddie edge shows up across bags; a single good run does not. The output is a caddie read that informs the golfer's-team construction in Career and Management, and never modifies a golfer's KR.

Does the contribution follow him across bags
Current bag
+0.42
2 seasons · conf 45%
Prior bag, a different golfer
+0.35
3 seasons · conf 40%
The contribution follows him across bags, modest but consistent, so it reads as a real caddie edge rather than one strong partnership flattering him. A borrowed edge, a single good run with one golfer, would collapse toward zero on the next bag.
Actor readInforms the golfer's-team construction in Career and Management, at wide confidence, and never modifies a golfer's KR.

A modest, consistent contribution across different bags is a real edge; a big number with one golfer and nothing with the next is a partnership, not a caddie. The read is an actor read, held at the honest, wide confidence its hard-to-isolate data supports, it informs how a golfer builds his team, and it never touches his KR, because a caddie is an actor in the golfer's team, not a term in his rating. A real edge travels across bags, and the read informs the team, never the rating.

Illustrative engine read on the real portability structure (the contribution tested across bags, informing Career and Management, never touching a KR, at wide confidence). Composite bags, demonstration figures.

The law underneath
Graded on the decisions he adds, not the bag he carries.

The engine credits the on-course decision-and-management gain, separates it from the shots the golfer struck, holds it at the wide confidence a hard-to-isolate signal deserves, and never lets the read touch a golfer's KR, because a caddie is an actor in the golfer's team, not a term in his rating. A real edge travels across bags, and the honest read prices the actor, not the reputation.

The caddie is a member of the team. See where the read goes.

The caddie read is an actor read that informs how a golfer builds the team around him. It sits beside the coach and the program as the people on the golfer's side, and it feeds the team-construction read, never the rating.

Get access