Estimate the historical final-score hit probability of a Super Bowl square from its two assigned digits, or compare its estimated fair value with the price paid.
Super Bowl Squares Probability Formula
A standard football squares board has 100 positions formed by digits 0 through 9 on each team axis. After the digits are assigned, a square wins a final-score prize when its two digits match the last digits of the teams’ final scores.
- P(a,b) is the historical probability of assigned digits a and b.
- Count(a,b) is the number of games ending with Team A digit a and Team B digit b.
- N is the number of games in the source sample.
Counting both team orders makes the matrix symmetric. It avoids assuming that the team written across the top is systematically stronger, at home, or otherwise different from the team written down the side.
Dataset Version
The calculator freezes a reproducible matrix labeled NFL final-score digit matrix v2015-2024-r1. It was derived from 2,743 regular-season and postseason games in the 2015 through 2024 NFL seasons. The included game types are regular season, wild card, divisional, conference championship, and Super Bowl.
Each game contributes two ordered observations, producing a denominator of 5,486. Matrix values are stored to four decimal places. Freezing the version prevents the same inputs from changing silently when new games are added to the upstream schedule file.
Before and After Digits Are Assigned
Before randomly drawn digits are placed on a complete board, every purchased position is equivalent: each has a 1% chance of receiving any one particular digit pair. After the digits are revealed, historical score patterns make some assigned pairs more common than others.
This distinction also means that choosing a physical square before the digits are drawn does not let a participant select 0-7 or another historically frequent pair unless the pool rules assign digits first.
Fair Value Formula
The prize input must be the amount allocated to the final score. If a pool divides its money among quarters, halftime, and the final, using the entire pool would overstate fair value.
Example
In the frozen matrix, 0-7 and 7-0 each have a historical final-score estimate of 3.6274%. For a $400 final-score prize, estimated fair value is $400 times 0.036274, or $14.51. If the square costs $10, estimated net value is $4.51.
Why the Calculator Uses Final Scores Only
First-quarter, halftime, third-quarter, and final-score digits have different distributions. A score of 0-0 is much more plausible early than at the end of a modern NFL game. Mixing periods into one matrix would make the result difficult to interpret, so this compact version covers only the final score.
Assumptions and Limitations
- Historical frequency is treated as a planning estimate, not a forecast for one specific Super Bowl.
- The model does not use team strength, betting spread, game total, weather, injuries, overtime likelihood, or rule changes after the 2024 season.
- Team-axis order is deliberately symmetrized.
- The result applies to one final-score prize and does not combine multiple payout periods.
- Pool rules vary. Confirm which team is on each axis, when digits are assigned, whether overtime counts, and how ties or unfilled squares are handled.
- Rounding the stored matrix to four decimal places creates a very small total-rounding difference across all 100 cells.
Browse the complete Football calculator collection or return to the Sports & Fitness Calculators hub.
Sources
nflverse. nfldata games schedule file and nflreadr Schedules Data Dictionary. Seasons 2015-2024 extraction, retrieved August 5, 2026.
National Football League. 2026 NFL Rulebook, Rule 11: Scoring.
SuperBowlSquares.org. Updated Super Bowl Squares Odds Based on the Last 10 Years of NFL Seasons. 2026. Methodology comparison only.