THE CONTROL
New scoring is not new behavior.
We also rescore the exact Stroke Play shots under Double Birdies. This isolates the scoreboard effect. Only the adapted simulation can produce additional long makes.
DISC GOLF · SCORING EXPERIMENT / 01
More runs.
Better drama?
A pro-stat-informed simulation of Stroke Play vs Double Birdies. Same players, same course, different decisions.
DOUBLE BIRDIES Albatross −6 · Eagle −4 · Birdie −2 · Par 0 · Bogey +1 · Double +2Double every stroke under par, on each hole. Par or worse stays unchanged. Lowest total wins.
A scenario experiment, not a validated prediction of professional play. Long-shot accuracy, miss costs and player adaptation are assumptions.
01 / THE RESULT
More highlights. A mixed finish.
How many more long makes?
● Stroke Play● Double Birdies
Distances are represented by discrete modeled lies, not measured shot-level data. Zero at 100+ feet in the central scenario means the model chooses to lay up there; real players sometimes run and make these shots.
What does “more entertaining” mean?
Keep highlights, competitive suspense and viewer response separate. All results below are simulated; viewer enjoyment is not measured.
| Metric | Stroke Play | Double Birdies | Interpretation |
|---|
A close finish is a final margin within one birdie–par swing: ≤1 stroke or ≤2 Double Birdies units, including ties. Contenders are within two such swings with six holes left. Comebacks measure the winners’ share that trailed after round two; tied winners share credit. Lead changes count changes in the sole leader after synchronized hole completion, ignoring intervening ties.
THE UNCERTAINTY
Test the assumptions.
This range is a sensitivity test, not a confidence interval or a bound on reality. Other courses, player behavior and miss patterns could produce results outside it.
02 / WHY PLAYERS CHANGE
The birdie run gets a bigger upside.
Consider a birdie chance where laying up guarantees par. Let p be the chance of making birdie and b the chance the run ends in bogey, with par otherwise.
Stroke Play: run if p > b
Double Birdies: run if 2p > b
Double Birdies: run if 2p > b
A 10% birdie chance with a 15% bogey risk is a bad Stroke Play bet (+0.05) but a good Double Birdies bet (−0.05), relative to a safe par.
Selective aggression, not “run everything.”
- More marginal birdie runs. Good comeback putting and forgiving greens make aggression cheaper.
- Par saves receive no direct bonus. Par versus bogey still differs by one. Eagle versus birdie differs by two, but both are doubled; the biggest asymmetry is crossing par.
- Danger still matters. OB, rollaways and weak comeback putting can overwhelm the bonus.
- Tee strategy may change too. More aggressive routes could create birdie chances, but tee-line adaptation is outside this model.
03 / THE WATCHABILITY VERDICT
A promising format to pilot.
The model supports more long-putt attempts and makes. It does not establish that tournaments become more enjoyable—or consistently closer.
Measure a real pilot
Track run rate by distance and score opportunity; makes per broadcast hour; birdie-to-bogey tradeoffs; late lead changes; players within a realistic comeback; and final margins measured in birdie swings.
Then test the actual audience: watch time, retention through the final holes, and a blinded preference survey using matched highlight packages. Rotate formats across comparable events and record green danger, wind and player intent.
What could improve over time?
More attempts could encourage practice of long putts and comeback putts. This simulation holds skill fixed: it estimates the immediate decision effect, not learning.
Learning would face diminishing returns, accuracy–speed tradeoffs, limited practice time, physical constraints and opportunity cost to other skills. Repeated shot-level data would be needed to estimate any long-term gain.
04 / OPEN METHOD
Real statistical anchors.
Explicit modeling choices.
Sixteen selected MPO professionals, a synthetic 18-hole course, three rounds. The central study runs 10,000 paired tournaments (8.64 million player-holes per format). The same seeded random inputs are used across formats to reduce simulation noise.
Data sources and player attributes
StatMando: 2025 DGPT MPO season statistics supplies parked, circle-in-regulation, putting and OB attributes. Source accessed October 28, 2026; page updated October 15, 2026. Completed 2025 season, minimum 40% of rounds tracked. These are selected recognizable pros, not a representative full-tour field.
C1X excludes tap-ins; C2 putting covers roughly 33–66 feet. These broad percentages do not identify accuracy at each exact distance or distinguish a deliberate run from a layup. UDisc definitions.
| Player | C1X | C2 putt | C1 regulation | C2 regulation | OB / 18 |
|---|
How the simulation works
- Place a player at a synthetic lie with two throws remaining for par, using nested parked/C1/C2 regulation rates adjusted by hole difficulty. Add a danger-weighted OB event using the published OB rate as an input; allow a small modeled early arrival on qualifying par 4s/5s.
- Use each player's C1X and C2 percentages to anchor assumed distance curves. C2 make rates at 40/50/60 feet are 1.5×/1×/0.5× the aggregate rate. Beyond 60 feet: 0.5 × C2 × exp[−(distance−60)/35], scaled by the scenario.
- For every lie, compare running and laying up using dynamic programming. A run can make, leave a comeback putt, roll away or take an OB penalty. A layup can also fail. Choose the action with the lowest expected final hole score in each format.
- Run both strategies with paired randomness and independently score the original stroke-policy shots under Double Birdies as a control. End tied tournaments as ties; allocate win credit equally. No playoff model.
Central adoption is 100%. The partial scenario adopts the Double Birdies policy in a random 50% of player-rounds. Strategies minimize expected hole score, not the probability of winning given the leaderboard.
Outside C2, central lie shares are 35% at 80 ft, 30% at 100, 20% at 150, 10% at 200 and 5% at 250. Additional near/far distributions are included in the downloadable study. Full probabilities, hole attributes and the seeded generator are in the source.
Calibration limits and what the model cannot establish
This is a policy experiment, not a fitted digital replica of the pro tour. Source stats anchor abilities but the complete model has not been validated against shot-level tournament data. It omits flight physics, scrambling detail, tee-line choices, weather, correlated player form, gallery pressure and adaptive winning strategy.
Birdie/bogey source averages and modeled birdie-or-better/bogey-or-worse are not perfectly matched definitions. The simplified course also understates scoring trouble. Consequently, treat leaderboard and player winner predictions with particular caution. A precise Monte Carlo interval does not cover these structural uncertainties.