Where Your Arccos Data Is Lying

16 findings from 148 rounds that the standard dashboard won't tell you

Contents 1. Your "best round" is a 2-hole fragment scored as -27 2. GPS is lying about your putting distances 3. The 22% three-putt rate on greens you already hit 4. One bad hole = 32% of all your over-par strokes 5. You lose strokes from 100-150y but gain them from 200y+ 6. Penalty strokes add 3.1 strokes per round when they show up 7. Your miss is 64% right — but rough doesn't hurt your approach 8. 9-hole normalization flatters your average by 1.5 strokes 9. Par 3 course mix inflated your "improvement" trend 10. You play 6 strokes better in the evening than the morning 11. The fall regression is real — two years running 12. Bad holes do cascade, but only slightly 13. You don't fade on the back 9 14. Back-to-back rounds don't hurt you 15. City Park improvement is real; Common Ground is flat 16. The priority practice map is wrong

1. Your "best round" is a 2-hole fragment scored as -27

The Data Lie

Your dataset shows a round of -27 over par per 18 at Common Ground on Aug 20, 2025. This is a 2-hole fragment (an ace on a par 3 + a bogey on a par 4) that got normalized: -3 over par × (18/2) = -27. It's dragging your mean, min, and percentiles in the wrong direction. Same issue with 3 other fragment rounds.

When you filter to data_quality = 'full' only (127 rounds), the picture barely changes (+8.4 vs +8.5 overall) — but your best/worst extremes and p1/p5 percentiles are unreliable without this filter.

2. GPS is lying about your putting distances

The Data Lie

Arccos reports 1,859 putts "inside 5 feet" with an average distance of 1.6 ft. But Arccos uses phone GPS, which has ±3-10 ft accuracy on a golf course. A putt reported at 1.6 ft could easily be 8-12 ft in reality.

The proof: your make rate from "inside 5 ft" is 70%. PGA Tour pros make 95%+ from true 5 ft and in. You're not missing 30% of 4-footers — the bucket is contaminated with much longer putts.

This means the previous finding that "putting inside 5 ft is your #1 leak" is likely overstated by the distance bucketing. Your real putting problem is probably concentrated in the 6-15 ft range, but the GPS can't resolve it.

3. The 22% three-putt rate on greens you already hit

The Hidden Number

When you hit a GIR, you average 2.12 putts. Scratch = ~1.75. That's +0.37 putts per GIR × 758 GIR holes = ~280 extra putts on greens you already reached in regulation.

Even more telling: 22.1% three-putt rate on GIR vs only 10.5% one-putt rate. You three-putt more than twice as often as you one-putt.

This is where the real strokes are hiding. Your approach game is getting you on the green. Your putting is giving it all back. The standard SG dashboard spreads this across "putting" generically — but it's specifically lag putting and speed control on GIR that's the issue.

4. One bad hole = 32% of all your over-par strokes

The Variance Tax

Across 96 rounds with 9+ holes: your average total over par is +6.0. If you remove just the single worst hole from each round, it drops to +4.1.

That means one hole per round accounts for 1.9 strokes — 32% of all your over-par scoring. This isn't about "getting better at golf." It's about damage control on one hole per round. The boring play — punch out, chip to the fat part of the green, two-putt for bogey — would be worth almost 2 strokes/round.

5. You lose strokes from 100-150y but gain them from 200y+

The Inverted Distance Curve

Most golfers lose more strokes the farther they are from the pin. You're the opposite:

DistanceShotsSG/ShotVerdict
100-150y401-0.122Your only losing approach range
150-200y321+0.024Near scratch
200y+325+0.087Better than scratch

Every club you use from 100-150y loses strokes: PW (-0.109), GW (-0.114), SW (-0.147), 9i (-0.136). The SG baseline expects more precision from this range, and you're not delivering it. This is a distance control and target selection issue in the scoring zone, not a swing issue.

6. Penalty strokes add 3.1 strokes per round when they show up

The Hidden Split

Clean rounds (no penalties): +7.0 avg (79 rounds)
Penalty rounds: +10.1 avg (69 rounds)

That's a 3.1 stroke penalty tax, and it hits 47% of your rounds. The 151 total penalty strokes across the dataset average just over 1 per round — but the downstream damage (recovery shots, mental reset) roughly triples the direct cost. SG analysis buries this because penalties get absorbed into driving or approach categories.

7. Your miss is 64% right — but rough doesn't hurt your approach

The Counterintuitive Finding

Miss direction: 36% left, 64% right. Clear right miss pattern.

But here's the surprise: approach shots from rough (SG -0.009) are better than approach shots from fairway (SG -0.092).

This is likely Simpson's Paradox at work. From the rough, you're probably playing more conservatively — aiming for the center of the green instead of attacking pins. The "safe" play from the rough is actually producing better outcomes than aggressive pin-hunting from the fairway. Your fairway approach strategy may be costing you more than missing fairways.

8. 9-hole normalization flatters your average by 1.5 strokes

The Data Lie

18-hole rounds: +9.2 avg (76 rounds)
9-hole rounds: +7.7 avg (72 rounds)

Your 9-hole rounds are 1.5 strokes better per 18. This could mean you play better in shorter formats, or it could mean the normalization (doubling a 9-hole score) smooths out variance. Either way, your "true" 18-hole scoring average is closer to +9.2 than the blended +8.5.

9. Par 3 course mix inflated your "improvement" trend

The Data Lie

Before Jun 2025: 3% of rounds on the Par 3 course
Jun 2025+: 14% of rounds on the Par 3 course

Par 3 rounds average +3.3 vs regulation courses at +8.0+. The increased Par 3 mix in the second half artificially improves your overall trend. This is textbook Simpson's Paradox — the trend looks better in aggregate but part of the improvement is just playing easier courses more often.

10. You play 6 strokes better in the evening than the morning

The Time-of-Day Effect

Morning (+13.7 avg (5 rounds)
Afternoon (12-4pm): +9.4 avg (51 rounds)
Evening (4pm+): +7.6 avg (92 rounds)

Caveat: only 5 morning rounds, so that's noisy. But the afternoon→evening difference (1.8 strokes, 143 combined rounds) is robust. Could be warmup effect, course conditions, cognitive state, or just that evening = more relaxed golf.

11. The fall regression is real — two years running

The Pattern

Summer 2024: +10.6 → Fall 2024: +12.6 (+2.0)
Summer 2025: +5.0 → Fall 2025: +9.7 (+4.7)

The fall regression is getting worse, not better. +4.7 stroke regression in fall 2025 vs +2.0 in 2024. Possible drivers: end-of-season fatigue, course conditions, shorter days, or Adderall/medication timing shifts with daylight savings. Worth tracking against your sleep and medication data.

12. Bad holes do cascade, but only slightly

The Good News

After a double+: +0.52 over par on next hole
After par/birdie: +0.40 over par on next hole

The cascade penalty is only +0.12 strokes. You recover reasonably well mentally. This is not where your strokes are hiding — the damage is in the double itself, not the aftermath.

13. You don't fade on the back 9

The Good News

Front 9: +0.46/hole · Back 9: +0.42/hole

You actually play marginally better on the back 9. Fatigue and stamina are not issues.

14. Back-to-back rounds don't hurt you

The Good News

First round: +8.8 · Second round: +8.6 across 61 pairs.

No fatigue penalty from playing consecutive days. Volume of play is not a concern.

15. City Park improvement is real; Common Ground is flat

The Simpson's Paradox Test

City Park: first 28 rounds +9.0 → last 28 +6.4 = -2.7 real improvement
Common Ground: first 24 +8.7 → last 24 +8.4 = -0.3 (flat)

Your improvement is disproportionately at one course. At Common Ground you've barely moved. This suggests course-specific learning (you know City Park's greens, miss spots, etc.) rather than broad swing/skill improvement. The transferable improvement is much smaller than the aggregate trend suggests.

16. The corrected practice priority map

What the standard analysis said vs. what the data actually says

Standard FindingRealityRevised Priority
Putting inside 5ft is #1 leak GPS accuracy makes this bucket unreliable. 70% make rate proves distance contamination. Real issue is lag putting/speed. #2: Lag putting & speed control (reduce three-putts from 22% to 15%)
Short game inside 20y is worst per-shot True per-shot, but only 235 shots. The variance tax from blow-up holes is bigger. #3: Short game inside 20y (still real, lower volume)
Approach is near scratch Only from 150y+. From 100-150y you lose -0.12/shot on 401 shots — scoring zone problem. #4: Scoring zone (100-150y) distance control
Not mentioned One bad hole per round = 1.9 strokes = 32% of all over-par scoring #1: Course management / damage control. Avoiding the blow-up hole is your single highest-ROI change.
Not mentioned Penalty strokes hit 47% of rounds, costing 3.1 strokes each time #1b: Penalty avoidance (layup decisions, miss-side awareness)
Not mentioned You approach BETTER from rough than fairway — conservative play works Mindset shift: Apply your rough strategy to fairway shots. Center-of-green from 100-150y.

The bottom line: Your #1 priority is not a swing change. It's a decision change. Play the boring shot when you're in trouble. Aim center-green from 100-150y. Lag putts to 3 feet, not at the hole. The data says you're a +6 golfer trapped in a +9 golfer's decision-making.