How PinFlag computes its numbers

Strokes gained, against a stated baseline

PinFlag's analytics are built on Strokes Gained (SG), the framework developed by Mark Broadie (Every Shot Counts, 2014) and used across professional golf. Every shot — or every round, depending on how it was logged — is compared to an expected-strokes baseline derived from PGA Tour performance. Your SG is always stated against that Tour baseline: 0 means tour-level, negative means strokes lost to it. We never silently switch baselines.

There are two ways a round's SG can be computed, and we always tell you which one you're looking at:

• Shot-by-shot (measured). When a round is logged shot-by-shot, each shot is scored against the expected-strokes table for its distance and lie. This gives the exact four-way split: off the tee, approach, around the green, putting.
• From round stats (estimated).When only totals exist (score, fairways, greens, putts), a stated formula estimates the split. That formula cannot separate approach from around-the-green — so we show them combined as “ball striking” rather than pretend to a precision the data doesn't have.

The four labels on every stat

Any stat in PinFlag that isn't raw logged data carries one of four provenance labels. They mean exactly this, everywhere:

• MeasuredComputed from shot-by-shot logged rounds — the exact split, no inference.
• EstimatedDerived from round totals (score, fairways, greens, putts) via a stated formula — directionally reliable, not shot-exact.
• ModeledAnchored to published research or modeled values, not this player’s data. Flagged until verified real data replaces it.
• PinFlag estimateCalculated by PinFlag from this player’s logged rounds using a stated method — not an official rating.

One definition of “biggest leak”

When PinFlag names your biggest leak, it means one thing: the category where you lose the most strokes per round vs the Tour baseline, computed from your own rounds. A category at or above the baseline is never called a leak, and deficits inside the noise floor (0.05strokes/round) don't count. Where we show the other lens — how you compare to golfers at your handicap level — it is always labelled as that comparison, and it never silently supplies the headline.

Practice outcomes: observed, gated, never causal

When a coach prescribes practice against a leak, PinFlag records the player's baseline in that category — the average over the last 5 SG-bearing rounds before the assignment, requiring at least 3. No verdict is shown until at least 3SG-bearing rounds have been played since; until then the card says it is still collecting data. Program-level claims state their thresholds: a prescription counts as “improved” only when the category gains at least 0.3 SG within the first 6rounds after assignment. We frame all of it as observed change — “SG improved since assigned” — never “the drill caused this.”

The same discipline applies to projections: no handicap trajectory is shown below 5 rounds, and uncertainty bands widen when the history is thin. A projection is never allowed to look more confident than the data behind it.

Where the benchmarks come from

• Skill-level SG profiles(scratch, 5/10/20 handicap): per-category means follow Mark Broadie's published amateur strokes-gained analysis (Every Shot Counts, 2014), stated in the Tour frame. The round-to-round spreads around those means are modeled — Broadie publishes means, not distributions — which is why percentile views carry a modeled label. Benchmark set version: 2026.2-tour-frame.
• Age and recruiting standardsare modeled guidance anchored to published scoring averages — not official recruiting standards, and labelled that way on every surface. Where we don't have credible data (girls' competitive standards differ materially and we haven't sourced published women's data yet), we suppress the projection instead of estimating it.
• The PinFlag handicap is a WHS-style estimate computed from your recent rated 18-hole rounds. It is not an official GHIN index and is labelled as an estimate everywhere it appears.

What we refuse to do

• We don't fabricate numbers — no invented benchmarks, no synthetic “proof”.
• We don't infer beyond the data — no approach/around-green split from totals, no distance bands without shot-level logging, no projections off a handful of rounds.
• We don't dress estimates up as measurements — the label is part of the number.
• We don't claim causation from correlation — practice outcomes are observed change, stated with their thresholds.
• We don't let any number ship without a defensible answer to “why this number?”

If you ever find a number in PinFlag that doesn't live up to this page, that's a bug — tell us and we'll fix it.