Beta Free full access during beta. All models, tools, and daily picks included.

Simulation NRFI

Research

First-inning scoring simulation using team run environments. Slate: 10/22/2026 50,000 trials per game

Research model only

Not official picks. Not included in the official public W-L record. This is an experimental first-inning simulation model under active research. The official NRFI model uses pitcher-specific features and is separate from this page.

Today's Research Picks

No research picks with positive edge for today's slate.

Selected-Pick Performance

Tracked from frozen sportsbook prices at selection time. Research model only. Not part of the public W-L record.

Total Picks
930
W-L-P
499-415-0
Win Rate
54.6%
ROI
6.3%

By Direction

NRFI 5 picks
4-1-0 (80%) ROI: 65.6%
YRFI 925 picks
495-414-0 (54%) ROI: 5.9%

By Edge Bucket

Edge 2-5%
318-266-0 ROI: 5.4%
Edge 5-10%
170-140-0 ROI: 7.6%
Edge 10%+
11-9-0 ROI: 9.5%
Brier Score
0.2479
Avg Edge
4.7%
Days Tracked
115

Calibration Tracking

Sufficient Sample 914 decided / 16 pending

Brier Score

0.2479

Log Loss

0.6889

ECE

0.0147

Calibration measures how well the model's predicted probabilities match actual outcomes. A perfectly calibrated model would have games predicted at 70% NRFI happening exactly 70% of the time. Lower Brier scores, log loss, and ECE indicate better calibration. This is a research model and is not used for official picks.

Bucket Count Avg Predicted Actual Rate Gap
0.00-0.10 0 -- -- --
0.10-0.20 0 -- -- --
0.20-0.30 0 -- -- --
0.30-0.40 0 -- -- --
0.40-0.50 0 -- -- --
0.50-0.60 860 55.7% 54.3% 1.4%
0.60-0.70 54 61.3% 59.3% 2.0%
0.70-0.80 0 -- -- --
0.80-0.90 0 -- -- --
0.90-1.00 0 -- -- --

By Direction

YRFI

909 graded Brier: 0.2479

NRFI

5 graded

Methodology

The Simulation NRFI model estimates first-inning scoring probability by simulating the first half-inning for each team. It derives per-inning run rates from the Game Market Core run distribution (F5 expected runs / 5 innings), applies a 10% first-inning scoring boost (historically, first innings score about 10% higher than average), and runs 50,000 trials using a negative binomial distribution with dispersion k=1.1.

P(NRFI) = fraction of trials where both teams score zero runs in the simulated first inning. P(YRFI) = 1 - P(NRFI).

This approach is intentionally different from the official NRFI model, which uses pitcher-specific first-inning features: career NRFI rate, first-inning ERA, K rate, BB rate, WHIP, and whiff rate. The simulation model captures the team-level run environment but does not see individual pitcher first-inning tendencies.

Research picks are selected where the simulation probability shows positive edge (2%+) against the sportsbook implied probability. All picks are tracked with frozen sportsbook prices from the time of selection.

Related