"Aim for the center" is folklore until it has a number
Each oval is one standard deviation of that tier's shot pattern at the 155-yard shot, drawn to scale in model yards and centered on the same reference point.
Golf instruction calls it playing to the fat part of the green. Course-management columns call it taking your medicine. Nobody attaches a number to either phrase, and the advice sounds the same whether the reader carries a 2-handicap or a 22.
This piece runs the aim decision through a model instead: one verdict per handicap tier, per pin, stated as a line and a club. Three pins guard the 12th at Augusta National. Two of them turn out fine to attack at every tier. One doesn't, and the chapters ahead show why in each case.
Every golfer's shot pattern around a single target forms a dispersion oval: how far the ball lands long or short, how far it drifts left or right. The oval sets the odds before the swing happens. Widen it, and the same aim point sends more balls into the bunker or the creek.
The five ovals in the figure belong to the five handicap tiers this piece tracks, all aimed at the same reference point and drawn to the same scale. Scratch's oval runs about 18.3 by 7.3 yards (down-range by left-right); a 20-handicap's runs 32.9 by 13.1, almost double the spread on both axes. The down-range axis mixes two components: a solid-strike core, plus a short mishit tail that comes up well short of the target and almost never comes up long by the same margin. The tail is modeled: a mishit rate that runs from 6.3% of swings for scratch to 31.3% for a 20-handicap (sensitivity range a 0.5x to 1.5x multiplier per tier), each mishit losing 25% of the shot's distance (range 20% to 30%), sized so the full mixture reproduces each tier's published green-hit rate at its own anchor distance. The 15 and 20-handicap ovals carry a further modeled tag: the proximity data behind them was measured at a shorter distance, so their size at this 155-yard shot comes from an extrapolation, not a direct read. The ratio between the down-range and left-right axes (anisotropy, in the model's vocabulary) is itself modeled at 2.5-to-1, with a sensitivity range of 2:1 to 3.5:1 tested throughout this piece.
The left pin: welcoming, and priced that way
Scratch through 20-handicap walk this pin below; the 5-handicap tier sits between them on the same scale but isn't called out in prose.
The left pin sits mid-depth on the green's roomier lobe, the fat target a course-management column would send a reader to first. The model agrees, at every tier.
Attack the flag or bail 4 yards toward the center. Either works.
The model's aim sits 4.0 yards right of the flag (toward the green's wider middle) and about half a club long, worth 0.045 strokes over playing it dead at the pin, inside the 0.05 tossup line. Water risk at the flag is 11.5%, plus another 11.2% short of the creek; aiming there instead drops water to 10.3%. Green odds climb from 30.9% to 41.0%.
Same read: aim 3.8 yards right and about half a club long. Either works.
Worth 0.027 strokes over the flag. Water risk at the flag is 12.4%, plus 21.1% short of the creek; the aim trims water to 11.5%. Green odds rise from 23.2% to 28.8%.
Aim 5.5 yards right and about half a club long. Either works.
The smallest edge of any tier at this pin: 0.018 strokes. Water risk at the flag is 12.0%, plus 26.2% short of the creek; the aim ticks water up to 12.9%, still well inside the tossup line. Green odds climb from 20.5% to 27.2%.
Aim 3.9 yards right, one club long. Either works.
Worth 0.024 strokes, ticking back up from the 15-handicap's edge. Water risk at the flag is 12.7%, plus 32.2% short of the creek, the widest short-miss share of any tier at this pin; the aim trims water to 11.8%. Green odds rise from 16.6% to 21.4%.
The center pin: not the sucker pin I expected
Scratch through 20-handicap walk this pin below; the 5-handicap tier sits between them on the same scale but isn't called out in prose.
I expected the center pin to demonstrate the glossary's sucker-pin idea: fair for scratch, marginal for a 15. The model found something flatter. At every tier, in calm air, attacking the center pin costs next to nothing.
Attack it. Aim 1.0 yard left, about half a club long. Either works.
Worth 0.024 strokes. Water risk at the flag is 14.1%, plus another 14.2% short of the creek; the aim drops water to 10.8%. Green odds hold flat, 45.5% to 46.2%.
Same call: aim 1.0 yard left, about half a club long. Either works.
Worth 0.013 strokes. Water risk at the flag is 13.7%, plus 23.8% short of the creek; the aim trims water to 12.4%. Green odds hold flat, 32.2% to 32.7%.
The closest thing to a pure tossup on the whole hole: aim 2.0 yards left, about half a club long.
Worth 0.012 strokes, the smallest edge this piece finds anywhere. Water risk at the flag is 12.6%, plus 29.2% short of the creek; the aim trims water to 12.0%. Green odds hold flat, 27.6% to 28.1%.
Aim 0.9 yards left, one club long. Either works.
Worth 0.018 strokes. Water risk at the flag is 13.1%, plus 35.1% short of the creek, the widest short-miss share of any tier at this pin; the aim trims water to 11.6%. Green odds rise from 20.8% to 23.3%.
This site's glossary defines a sucker pin as a flag whose direct line sends a meaningful share of the oval into a hazard or a short-sided lie. The center pin's front-third depth runs about 10.5 yards on average (a modeled range of 9 to 12), enough room that even a 20-handicap's much wider oval spreads its misses across green, rough, and the near edge of the creek without concentrating them anywhere lethal. That isn't what a sucker pin looks like. The one the model finds is next.
The Sunday pin: the real sucker pin
Scratch through 20-handicap walk this pin below; the 5-handicap tier sits between them on the same scale but isn't called out in prose.
The Sunday pin sits in the green's back-right corner, about 15 yards deeper than a front-left flag and closest to Rae's Creek. Every tier's verdict says the same thing: bail.
Bail. Aim 7.1 yards left of the flag and about half a club less.
Worth 0.095 strokes, well clear of the tossup line. Water risk at the flag is 19.0%, plus another 15.7% short of the creek; bailing left drops water to 14.7%. Green odds rise from 34.2% to 45.8%.
Bail again: aim 7.0 yards left and about one club less.
Worth 0.078 strokes. Water risk at the flag is 16.0%, plus 25.7% short of the creek; bailing trims water to 15.0%, a small drop from the flag's number. Green odds rise from 26.8% to 32.0%.
Still bail: aim 7.8 yards left and about one club less.
Worth 0.070 strokes. Water risk at the flag is 15.9%, plus 30.5% short of the creek; bailing drops water to 14.7%. Green odds rise from 21.4% to 27.8%.
Bail hardest here: aim 8.0 yards left and about one club less.
Worth 0.053 strokes, the thinnest edge of any tier at this pin. Water risk at the flag is 13.7%, the lowest of any tier here, plus 37.4% short of the creek, the widest short-miss share on the hole. Bailing trims water to 13.0%. Green odds rise from 17.6% to 21.5%.
I expected the water rate to climb with handicap here, the same assumption this piece opens with. At the flag, calm air, it runs the other way: scratch's 19.0% water rate is the highest of any tier at Sunday, and a 20-handicap's 13.7% is the lowest. The creek band sits 10 to 24 yards short of this pin, close enough that a scratch player's tighter miss pattern lands there often; a 20-handicap's much wider pattern spreads past that band in both directions, landing more often in the fairway well short of it or clearing the green on the long side. Add the two short-side outcomes together and the ranking flips back: a 20-handicap's total short-side miss rate, water plus short of the creek, comes to 51.1% (13.7% water, 37.4% short); scratch's total is 34.7% (19.0% water, 15.7% short), lower in both pieces.
The delta for attacking Sunday sits in a tight band, 0.053 to 0.095 strokes in calm air, across every tier this piece tests: the pin costs every tier close to the same amount, and the direction runs backward from what handicap alone would predict. Scratch pays the most; a 20-handicap pays the least. One thing resolves outright: the strict, unclamped search now agrees with the published, full-swing aim at every tier through the 15-handicap, and edges 0.002 strokes deeper for a 20.
Wind adds eight yards of trouble, and turns bail into a coin flip
15-handicap, Sunday pin, published aim in both conditions.
The source log carries wind at this hole as a narrative fact, not a number (Anchor 6): the tee sits sheltered while the green sits exposed, and the 2019 final round played through gusts of about 20 mph. The model prices wind as two modeled multipliers: an 8-yard carry penalty (range 4 to 12) and a 1.3x dispersion inflation (range 1.15 to 1.5), applied on top of whichever tier and pin the golfer is playing.
A 15-handicap's oval at Sunday measures 27.5 by 11.0 yards, and the bail is worth 0.070 strokes.
The oval swells to 35.7 by 14.3 yards, and the bail signal crosses the tossup line: 0.049 strokes, a hair under the 0.05 threshold, reading as "either works" instead.
This flip sits close enough to the line that pushing the model's sensitivity ranges to their edges calls three windy Sunday cells a tossup too. Calm air at every tier, and every other wind state tested here, still call Sunday "bail." Left and center change little: scratch is the single tier where wind flips a label, from "either works" to "bail" at both pins (0.059 strokes at left, 0.053 at center, both past the tossup line); every tier from the 5-handicap up stays "either works" at both pins in the wind, the same as calm air.
April 14, 2019: four balls in Rae's Creek within the hour
In the final round of the 2019 Masters, four contenders found Rae's Creek off the tee at the 12th within about an hour of each other. Brooks Koepka's 9-iron drifted right, caught the bank, and rolled back into the water. Ian Poulter's 8-iron came up short and never carried the creek. Francesco Molinari's tee shot caught the same bank and rolled back into the creek. Tony Finau's tee shot found the water too. All four made double bogey.
Wind gusted to about 20 mph that afternoon, the kind of push the wind chapter above prices as an 8-yard carry penalty and a wider oval, not a fixed stroke count.
The week's numbers show how rare that hour was against how the hole plays most of the time: 2019's tournament-wide scoring average on 12 sat at 3.053, with 52 birdies, 200 pars, 38 bogeys, and 14 double-bogeys-or-worse across 304 plays. One bad hour produced four of those doubles.
Does the model know this hole?
Per-year scoring averages, hole 12, 2019-2025 (2020 not played). Dashed line: the model's Tour season mean.
The amateur ovals come from published amateur proximity data. The Tour oval runs through the same construction, and it is the one the hole's scoring record can check, so before trusting the amateur verdicts the Tour oval has to reproduce that record.
At the 155-yard shot, the model's Tour oval measures 8.4 by 5.4 yards, tight enough that Rae's Creek almost never enters a Tour player's equation. Run through a full season of pin positions and wind, the model's Tour season average comes out to 3.1142 strokes.
That number lands inside the modern-era band this analysis checks against: the mean of six published per-year averages (2019, 2021 through 2025, no usable 2020 figure) runs 3.053 to 3.233 and averages 3.132, with a sample standard deviation of about 0.073. The hole's all-time scoring average, 3.27 to 3.28, describes a different era: it spans tournaments back to 1934, and every modern year on record scores below it. Comparing a model built on 2016-2023 proximity data against a 90-year average would grade it against a different hole, so this piece checks the Tour oval against the six modern seasons instead.
One gap stays open and disclosed. The model's Tour oval runs more bogey-heavy than the published record in every year it can be checked against. For 2019 and 2023 the published averages (3.053 and 3.058) sit below the model's calm-air floor, so no wind setting reaches them; at calm air the 2019 bogey share runs 5.7 points high. For 2024 and 2025 the fitted wind frequency matches the year's average, and the bogey share still runs 6.6 and 3.3 points high against a 3-point tolerance. The season average checks out; the shot-by-shot shape runs too rough, and this piece publishes with that gap stated rather than tuned away.
The move
Three pins, four tiers each, one verdict per line. Every number below comes from outputs/003_moves.csv, the full-shot aim clamped to a normal swing (carry within 10 yards of the flag), never the deeper layup a strict, unclamped search sometimes prefers.
Left pin
Scratch: aim 4.0 yards right of the flag and about half a club long. Either works. 0.045 strokes; water 11.5% at the flag (11.2% short of the creek) vs 10.3% at the aim.
10-handicap: aim 3.8 yards right and about half a club long. Either works. 0.027 strokes; water 12.4% at the flag (21.1% short of the creek) vs 11.5% at the aim.
15-handicap: aim 5.5 yards right and about half a club long. Either works. 0.018 strokes; water 12.0% at the flag (26.2% short of the creek) vs 12.9% at the aim.
20-handicap: aim 3.9 yards right, one club long. Either works. 0.024 strokes; water 12.7% at the flag (32.2% short of the creek) vs 11.8% at the aim.
Center pin
Scratch: aim 1.0 yard left of the flag, about half a club long. Either works. 0.024 strokes; water 14.1% at the flag (14.2% short of the creek) vs 10.8% at the aim.
10-handicap: aim 1.0 yard left, about half a club long. Either works. 0.013 strokes; water 13.7% at the flag (23.8% short of the creek) vs 12.4% at the aim.
15-handicap: aim 2.0 yards left, about half a club long. Either works. 0.012 strokes; water 12.6% at the flag (29.2% short of the creek) vs 12.0% at the aim.
20-handicap: aim 0.9 yards left, one club long. Either works. 0.018 strokes; water 13.1% at the flag (35.1% short of the creek) vs 11.6% at the aim.
Sunday pin
Scratch: aim 7.1 yards left of the flag and about half a club less. Bail. 0.095 strokes; water 19.0% at the flag (15.7% short of the creek) vs 14.7% at the aim.
10-handicap: aim 7.0 yards left and about one club less. Bail. 0.078 strokes; water 16.0% at the flag (25.7% short of the creek) vs 15.0% at the aim.
15-handicap: aim 7.8 yards left and about one club less. Bail. 0.070 strokes; water 15.9% at the flag (30.5% short of the creek) vs 14.7% at the aim.
20-handicap: aim 8.0 yards left and about one club less. Bail. 0.053 strokes; water 13.7% at the flag (37.4% short of the creek) vs 13.0% at the aim.
Your handicap, your pin, your wind.
The sandbox behind this piece runs the same model in the browser: pick a tier, a pin, and a wind state, then drag the aim point and watch expected score, water risk, and the strokes lost against the model's optimum update live.
Open the aim sandboxMethod and sources
Every oval in this piece comes from the same three-step chain. First, published proximity and greens-in-regulation figures at the 155-yard band get inverted into a radial miss scale per tier, direct for scratch through 10-handicap and interpolated for 15 and 20. Second, that radial scale splits into a down-range axis and a left-right axis using an anisotropy ratio. Third, the resulting oval gets placed against the hole's geometry, bunker positions, the green's diagonal shoe-sole shape, and Rae's Creek's finite band, and integrated over every outcome region to price expected strokes.
| Input | Value | Status | Sensitivity range |
|---|---|---|---|
| Radial miss scale, scratch/5/10-handicap | varies by tier | Anchored | - |
| Radial miss scale, 15/20-handicap | varies by tier | modeled | same formula, longer extrapolation |
| Anisotropy ratio | 2.5 | modeled | 2.0 to 3.5 |
| Mishit-tail rate by tier | 6.3% to 31.3% | modeled, anchorless | 0.5x to 1.5x |
| Mishit-tail short-distance fraction | 25% | modeled, anchorless | 20% to 30% |
| Green width | 25.5 yd (midpoint) | Published range | 16 to 35 yd |
| Green overall depth | 26.5 yd (midpoint) | Published range | 20 to 33 yd |
| Center pin front-third depth | 10.5 yd (midpoint) | Published range | 9 to 12 yd |
| Creek width + bank rollback | 14.0 yd | modeled | 9 to 20 yd |
| Wind carry penalty | 8 yd | modeled | 4 to 12 yd |
| Wind dispersion inflation | 1.3x | modeled | 1.15x to 1.5x |
| Up-and-down rate by tier | 30% to 54% | modeled, weak sourcing | 0.8x to 1.2x |
| Pitch-over-water dunk rate by tier | 2% to 12% | modeled, anchorless | 0.5x to 1.5x |
A flip set sweeps every verdict across the anisotropy range and the front-third-depth range together, at each pin, tier, and wind state, and records which qualitative labels change. At the sweep's baseline settings, calm air calls Sunday "bail" at every tier, and every wind state does too. Push both ranges to their sensitizing extremes at the same time, and a handful of cells move:
- Scratch's left pin reads "bail" at the sweep's baseline in calm air, and flips to "either works" at one corner: the narrowest anisotropy (2.0:1) paired with the widest front-third depth (12 yd).
- Scratch's center pin reads "bail" at baseline under wind, and flips to "either works" at one corner: the widest anisotropy (3.5:1) paired with the widest front-third depth.
- The 5-handicap's left pin flips the other direction, from "either works" at baseline to "bail" at the widest anisotropy: in calm air at the widest front-third depth, and in wind at both front-third-depth corners tested.
- The Sunday pin, windy, is no longer untouchable: the 10, 15, and 20-handicap tiers each flip from "bail" to "either works" at one or more corners, the 15-handicap at all three of the high-anisotropy, wide-depth corners tested, the widest reach of any cell in the set.
None of these flips touch the piece's central finding at the sweep's baseline settings: Sunday reads "bail" at every tier and wind state tested there. Push both sensitivity ranges to their extremes together, though, and three windy Sunday cells call it a coin toss instead, a real dent in how bulletproof that finding is, not a reversal of it. A separate sweep of the up-and-down rate table across its 0.8x to 1.2x range changes no pin's label at any tier: left and center stay "either works," Sunday stays "bail."
- Hole geometry and 2019 final round
- PGA.com and golfguidebook.com (155-yard tee yardage, bunker layout, diagonal green shape); NBC Sports/Golf Channel and The Fried Egg (bunker count and the azalea-lined back ledge); SI.com and NBC Sports/Golf Channel (all-time scoring average, corroborated); Racing Post and a second independent source (2019's 52/200/38/14 outcome breakdown); PGATOUR.com's retrospective (the 2019 water-ball sequence and the day's wind).
- Amateur proximity, greens-in-regulation, and putting
- Arccos Golf's "magic number" GIR-distance table and its 125-to-149-yard proximity figure; Shot Scope's putting make-percentage table across six handicap bands and its three-putt-rate figures. modeled where no distance-specific breakdown exists in the published data.
- Anisotropy
- Broadie, "Assessing Golfer Performance Using Golfmetrics" (Science and Golf V, 2008): a 3-to-1 amateur distance-to-line dispersion ratio, measured on short-game and sand shots, carried across distance bands as this analysis's weakest anchor. The default here sits at 2.5-to-1, inside Broadie's range, chosen along with the mishit tail's two parameters to best fit the water-rate and 200-yard-carry sanity checks in Chapter 1, not Broadie's literal figure. modeled, sensitivity range 2:1 to 3.5:1.
- Mishit tail
- No published source splits amateur distance error into a solid-strike core and a short-miss tail; this analysis builds both from scratch, solved so the full mixture reproduces each tier's published green-hit rate at its own anchor distance (Anchor 1's GIR-by-distance table). modeled, anchorless; mishit rate sensitivity range a 0.5x to 1.5x multiplier per tier, short-distance fraction range 20% to 30%.
- Tour proximity and recovery rates
- Mike Bury Golf, Golf Insider UK, and The Left Rough (150-to-175-yard Tour proximity and greens-in-regulation, three independent sources); SwingU Clubhouse (Tour scrambling rate); a corroborated sand-save figure and a derived missed-up-and-down cost built on Broadie's published bunker strokes-to-hole-out figure.
- Modern-era scoring average
- PGA Tour's course-stats pages and Today's Golfer, six per-year averages (2019, 2021 through 2025), the modern seasons this analysis checks the model against instead of the hole's 1934-2025 all-time average.
- Wind
- SI.com and PGATOUR.com's retrospective: a narrative anchor only (Anchor 6), the 20-mph gusts of the 2019 final round. The carry penalty and dispersion inflation applied in this piece are modeled, with stated sensitivity ranges.
Monte Carlo fidelity: a 300,000-run harness matches this piece's analytic amateur surface within 0.012 strokes and its Tour surface within 0.011 strokes, both inside the 0.03-stroke tolerance the test suite asserts. Full source rows, retrieval dates, and every anchor's verdict live in docs/sources/003_Source_Log.md; this piece's per-number ledger lives in docs/sources/003_Chapter_Numbers.md.
None of these numbers come from release 002: every one traces to this hole's own model, CSVs, or the source log above, so no cross-check against 002_results.csv applies here.