NFL Immaculate Grid: Rules, Rarity Score and Real Strategy

NFL Immaculate Grid is a daily 3x3 puzzle where every square sits at the intersection of a row and a column, and you name one NFL player who satisfies both. Nine squares, nine guesses, no player used twice. It’s run by Sports Reference, the company behind Pro-Football-Reference, and it’s free. The part most people get wrong is the scoring: you’re not trying to score high, you’re trying to score low.

Who actually runs it, and where it lives now

The grid was built by Brian Minter, an Atlanta software engineer, and the first one went live on April 4, 2023 as a baseball puzzle. Sports Reference bought it on July 11, 2023, and the football version launched nine days later, on July 20. Basketball, hockey, women’s basketball and a soccer edition called Immaculate Footy followed within weeks, according to Wikipedia’s record of the game.

Two things changed since, and both are worth knowing because stale guides still get them wrong.

First, the address moved. In September 2025 Sports Reference folded the game into its own domain. The old immaculategrid.com URLs now 301-redirect to sports-reference.com/immaculate-grid/football, which we verified directly. Old links still work. They just don’t land where they used to.

Second, the ownership rumour. In October 2025 Religion of Sports, the production company co-founded by Tom Brady and Michael Strahan, picked up rights to develop a TV game show from the format. That is a licensing deal for a show, not a sale. Sports Reference still runs the game, and it’s still free and still daily.

What time does the grid actually reset? Sources conflict, and here’s the truth

This is the single most-argued detail about the game, so it’s worth being precise.

Grids originally premiered at midnight. On July 20, 2023, Sports Reference announced that from July 26 new grids would launch at 9:00 a.m. ET instead, with a refreshingly human reason: so staff wouldn’t have to pull all-nighters to fix problems. That announcement is the last primary statement from the operator that we can find, and no later announcement of another change exists.

So where does “6 a.m. ET” come from? It’s everywhere, and it doesn’t trace back to Sports Reference. The guides repeating it are either dated before the July 2023 change or are describing a different game entirely. Unofficial grid clones each set their own reset times, and their reset times get copied into articles about the official one. Sports Reference’s own pages block automated checking, so we can’t read today’s live time off the page, and we’d rather say that than pretend.

The practical version: if you’re in the UK or Australia, don’t set an alarm off a number you read in a blog. Load the grid, note the grid number, and see when it increments. That’s the only reset time that’s true for you.

The rules, stated plainly

• Nine guesses total. One per square. A wrong guess burns a guess and the square stays empty.
• No player twice. Each name can only be used once per grid.
• Eligibility is one game. For a team header, the player needs to have appeared in at least one regular-season or playoff game for that franchise. Not a start, not a snap count, one game.
• Career stats are career-wide. If a square asks for a team plus a career statistical threshold, the player needs the stat at some point in his career and one game for that team. The two don’t have to have happened together.
• Champion headers are season-specific. For a Super Bowl champion header paired with a team, the player has to have appeared for that team during the winning season. If you’re fuzzy on which seeds and rounds produce a champion, our guide to the NFL playoff format covers the bracket itself.
• Looking things up isn’t cheating. There’s no rule against it. The game is powered by Pro-Football-Reference, and half the fun is that the answer key is public.

Headers come in two flavours: the 32 franchises, and category headers such as career yardage thresholds, Pro Bowl and All-Pro selections, draft position, college, and Super Bowl champions. A team-versus-team square is a career-path question. A team-versus-category square is a roster-history question. They reward different memories.

How the rarity score really works

Sports Reference added the rarity score in late June 2023 with a one-line summary: the lower, the better.

Here’s the mechanic. When you fill a square correctly, the game tells you what percentage of everyone else who played that grid put the same player there. Your rarity score is the sum of those nine percentages. Every square you leave empty scores a flat 100.

That flat 100 is the whole game, and it’s the thing casual players miss. Consider the maths:

Outcome	Rarity contribution
Correct, and 42% of players picked the same guy	42
Correct, and 3% picked the same guy	3
Empty square (wrong guess or no guess)	100
Worst possible grid (all nine empty)	900
Best realistic grid (nine deep cuts)	under 50

A perfect nine-for-nine grid of obvious answers can score worse than 400. A nine-for-nine grid of genuinely obscure answers can beat 100. Same “immaculate” result, wildly different score. Anything under 200 with all nine filled is a good day.

The design lesson is that this is golf, not trivia. Sports Reference borrowed the scoring shape from golf deliberately, and the reason it works is that it converts a puzzle with a ceiling into one without a floor. Once you’ve filled nine squares you’ve stopped competing against the grid and started competing against everyone else’s obviousness.

Why the football grid is harder than the baseball one

Baseball came first and it’s still the biggest version, roughly half the game’s traffic. It’s also easier, for a structural reason: baseball players move constantly, careers run twenty years, and a journeyman reliever can rack up eight franchises. Team-versus-team squares have deep answer pools.

NFL careers are short. The average is a handful of seasons, rosters turn over hard, and a player who appeared for both the Chargers and the Jaguars is a much thinner category than the baseball equivalent. That squeezes the answer pool on team-versus-team squares, which is why a two-team NFL intersection often has a handful of viable names and everyone lands on the same one.

Which has a consequence: on NFL grids, the rarity gap between the popular answer and the deep cut is bigger than it looks. If only six players ever suited up for both franchises and 60% of the field picks the famous one, the fifth-most-famous name is worth roughly 55 points to you.

Strategy that actually lowers your score

Fill first, optimise second. A 100-point empty square costs more than almost any bad-but-correct answer. If you have one certain name and one clever name, take the certain one. Chasing rarity into a wrong guess is the most expensive mistake in the game.

Solve the hardest square first, while your guesses are cheap. Most people work left to right and arrive at the brutal intersection with one guess left and no room to be wrong. Scan all nine, identify the one square where you have exactly one candidate, and place it before your remaining flexibility is gone.

Practice squads and one-game cameos are free rarity. The one-game eligibility rule is the most under-used thing in the ruleset. Everyone thinks about who starred for a team. Almost nobody thinks about the veteran who was there for six weeks at the end of a career. Those are the 2% answers.

Backups over Hall of Famers on stat headers. A “team plus 10,000 career passing yards” square makes everyone think of the franchise legend. The backup quarterback who passed through for one season and earned his yardage elsewhere satisfies exactly the same criteria for a fraction of the percentage.

Don’t spend rarity on a square that has no depth. Some intersections have three valid answers in history. Take the one you’re sure of and bank the guess. Save your ambition for the squares with a hundred candidates, where the rarity spread is wide enough to be worth hunting.

Know your headers’ era. Category headers such as Pro Bowl selections and draft rounds span the entire history of the league. Team headers do too. If you only think about the last decade, you’re playing with 10% of the answer pool and picking from the same names as everyone else.

If you like this, the adjacent format worth trying

The grid tests breadth: who was where, across the whole history of the league. The other big daily NFL format tests deduction instead, and it’s a genuinely different exercise.

We run Guess the NFL Player, which is not a grid and doesn’t try to be. You get eight guesses at one hidden player, and each guess returns attribute clues, so you narrow the field by inference rather than by recall. There’s a daily puzzle and a free unlimited mode with no account, which is the one thing the grid deliberately doesn’t offer. It won’t scratch the same itch as a 3x3, but if what you actually enjoy is the daily-NFL-puzzle habit, it’s the same habit in a different shape. The rest of our free sports games live in the same place.

The honest verdict

NFL Immaculate Grid is the best-built game in this category, and the reason is boring: Sports Reference owns the database. The answer validation is right because Pro-Football-Reference is right. No clone can match that, and most of them don’t try.

Its real weakness is the one it chose. One grid a day, no unlimited mode, and a reset time that even its own fans argue about. That’s a deliberate scarcity play, and it’s why an entire ecosystem of unofficial grids grew up around it. If you want more than one puzzle a day, the official game has never been the place to get it.