AETHER ENGINE ONLINE◆CADDIE ONLINE◆MODELS UPDATED 2026-04-04T06:00Z◆PROJECTIONS 12,847
// PREDICTION_ENGINE


|

> GOLF DFS INTELLIGENCE // COURSE-ADJUSTED
> OWNERSHIP MODELING // FIELD-AWARE
> ENSEMBLE PREDICTIONS // MONTE CARLO
[ENTER CADDIE →][HOW IT WORKS]
// LIVE_PROJECTIONSCADDIE ONLINE
#PLAYERPROJOWN%
1Tommy Fleetwood65.419.7%
2Ludvig Aberg64.921.5%
3
RM
Robert MacIntyre64.124.6%
4Russell Henley63.422.6%
5
SK
Si Woo Kim63.318.8%
6
HM
Hideki Matsuyama61.815.2%
7
MM
Maverick McNealy61.223.0%
8
RF
Rickie Fowler60.512.4%
PAUSED
SCOTTIE SCHEFFLER • PROJ 52.1 • OWN 28.4% • EDGE +6.2RORY MCILROY • PROJ 49.8 • OWN 22.1% • EDGE +4.7COLLIN MORIKAWA • PROJ 48.3 • OWN 14.6% • EDGE +5.1XANDER SCHAUFFELE • PROJ 47.9 • OWN 18.3% • EDGE +3.8WYNDHAM CLARK • PROJ 46.1 • OWN 9.2% • EDGE +7.4LUDVIG ABERG • PROJ 45.7 • OWN 11.8% • EDGE +5.9SAM BURNS • PROJ 44.2 • OWN 7.1% • EDGE +8.1TOMMY FLEETWOOD • PROJ 43.8 • OWN 6.4% • EDGE +6.6HIDEKI MATSUYAMA • PROJ 43.1 • OWN 8.9% • EDGE +4.3TONY FINAU • PROJ 42.6 • OWN 5.7% • EDGE +7.8SCOTTIE SCHEFFLER • PROJ 52.1 • OWN 28.4% • EDGE +6.2RORY MCILROY • PROJ 49.8 • OWN 22.1% • EDGE +4.7COLLIN MORIKAWA • PROJ 48.3 • OWN 14.6% • EDGE +5.1XANDER SCHAUFFELE • PROJ 47.9 • OWN 18.3% • EDGE +3.8WYNDHAM CLARK • PROJ 46.1 • OWN 9.2% • EDGE +7.4LUDVIG ABERG • PROJ 45.7 • OWN 11.8% • EDGE +5.9SAM BURNS • PROJ 44.2 • OWN 7.1% • EDGE +8.1TOMMY FLEETWOOD • PROJ 43.8 • OWN 6.4% • EDGE +6.6HIDEKI MATSUYAMA • PROJ 43.1 • OWN 8.9% • EDGE +4.3TONY FINAU • PROJ 42.6 • OWN 5.7% • EDGE +7.8
// THE_ENGINE

AETHER

The quantitative prediction engine behind every projection.

Aether is not a wrapper around a data API. It is a purpose-built quantitative prediction system — a stacked ensemble with a meta-learner that learns when to trust each signal.

Every course has a fingerprint. Every golfer has a profile. Aether matches them with 32 course attributes mapped to historical performance across strokes gained decomposition.

Ownership is modeled as a first-class prediction target — not an afterthought. Aether forecasts what the field will do before they do it.

10,000 Monte Carlo simulations per slate. James-Stein shrinkage for SG normalization. Bayesian updating as rounds complete. Fractional Kelly for lineup sizing.

This is institutional-grade quantitative analysis. Built for any sport. Deployed first in golf.

DATA_INGESTION
Real-time field data • Form signals • Weather modeling • Salary inefficiency mapping
COURSE_FINGERPRINT
32 attributes • Par composition • Wind exposure • Rough penalty
ENSEMBLE_MODEL
XGBoost • LightGBM • Random Forest • Meta-learner
OWNERSHIP_PREDICTION
Salary • Narrative • Historical patterns • Price efficiency
MONTE_CARLO_SIM
10,000 simulations • Wave correlation • Bayesian updating
YOUR_EDGE
FPPG projection • CI bounds • Ownership forecast • Kelly sizing
AETHER IS LIVE. CADDIE IS POWERED BY IT.[VIEW CADDIE PROJECTIONS →]
// PRODUCT_SUITE
CADDIEONLINE
GOLF_DFS_INTELLIGENCE
> Course-adjusted projections // 130+ players
> Ownership modeling // field-aware
> Monte Carlo simulation // 10K sims
> Lineup optimization // Kelly-sized
[ENTER CADDIE →]
MORE SPORTS COMING.
Arcline Analytics is expanding. Golf is first.
NBA, MLB, and beyond are in development.
[JOIN THE WAITLIST]
// ACCURACY_RECORD

WE SHOW OUR WORK

0.74
OWNERSHIP_MODEL_r
n=516, CI: [0.70, 0.78]
0.401
RANK_CORRELATION
2025 holdout set
42%
WINNER_IN_TOP_10
114 of 274 tournaments
18
PREDICTION_SIGNALS
5,592 holdout obs
// EDGE_IN_NUMBERS
0
PREDICTION SIGNALS
Validated on 5,592 holdout observations
0.00
OWNERSHIP MODEL r
n=516 Augusta observations, 95% CI: [0.70, 0.78]
0%
WINNER IN TOP 10
114 of 274 tournaments in historical dataset
0.000
RANK CORRELATION
2025 holdout set — data the model never trained on

Performance metrics based on historical backtesting on holdout data. Past performance does not guarantee future results. Golf DFS involves substantial variance.