How I Designed a Scoring System for Connections

Perfection is in my job description. I lead The New York Times’s Flexible Editing team, a group that’s relentlessly committed to perfecting the language, grammar and quality of writing across our news report. It’s a hard switch to turn off, especially after work when I’m trying to pursue diversions that most people find relaxing.

Like Connections.

When I play, solving the categories in any order should lead to a satisfying victory. But my perfectionist brain can’t handle that. When I first started comparing results with a friend, we often learned that we had both solved the game in four lines. Instead of congratulating each other, we needed to know who had the “better” solve. Who won?

“I have a new game within the game,” that friend, Sean, texted me one morning in November. “Try to get the purple line first.”

“Challenge accepted,” I replied without hesitation. Then I immediately started thinking: What if I created a scoring system that encouraged solving not just the first line, but the entire game, in a specific way? That would make perfection more elusive — and, by extension, more satisfying to achieve. It would also allow me to more objectively compare my results with Sean’s.

We’ve been keeping score ever since.

The idea of turning Connections into a head-to-head competition will surely be polarizing to some players. Part of the game’s appeal for many is how easy it is to pick up. So, you might ask, why muddy the waters by adding a layer of difficulty and introducing a numerical scoring element?

When asked by a reporter a century ago why he would want to climb Mount Everest, the English mountaineer George Mallory is said to have responded, “Because it’s there.” As an editor, I might satisfy my cravings for adventure in less dangerous ways (think: splitting infinitives and pushing the boundaries of The Times’s stylebook), but my answer is no more complicated than Mallory’s: It’s just more fun for me this way.

Think of it as the rough equivalent of playing Wordle on hard mode, though I acknowledge it’s not for everyone. So if a scoring system spoils the game for you, don’t use one.

But if you want to turn Connections into a competition, start by finding an opponent who shares your competitive drive. Then agree on some kind of ground rules. The possibilities are infinite, but I’ll describe the system I’ve been using to compete with Sean for the past few months.

Each color is appointed a value based on its difficulty. Purple (the trickiest category) is worth 4 points, followed by blue for 3, green for 2 and yellow for 1. A strikeout on a line earns nothing.

Each line is also assigned a multiplier, again in descending order:

Line 1 = × 4
Line 2 = × 3
Line 3 = × 2
Line 4 = × 1

The total for each line is determined by multiplying the value of that color by the line’s multiplier. Add the totals from the first four lines together to find your aggregate score.

A perfect score of 30 requires not only that you solve Connections in exactly four lines, but also that you solve it in perfect descending order of difficulty, from purple to yellow. Here’s what a perfect grid looks like:

🟪🟪🟪🟪 (4 × 4 = 16)
🟦🟦🟦🟦 (3 × 3 = 9)
🟩🟩🟩🟩 (2 × 2 = 4)
🟨🟨🟨🟨 (1 × 1 = 1)
30/30 (16 + 9 + 4 + 1)

Pursuing a perfect score in this system adds a twist: You almost always need to solve the entire grid in your head, or on a piece of paper, before entering any guesses in the game. And after you think you’ve solved it in your head, you then need to make a mental judgment about the order of difficulty. This is harder than it sounds because it involves some psychology. You’re determining not which category was actually the most difficult for you, but rather which category the game’s editor determined to be the trickiest. Warning: You’ll frequently disagree.

I land my share of perfect scores, but more often I come painfully close. My most common result since I’ve been keeping track is a 29, evidence of just how subjective this exercise is.

A complete whiff on the first four lines will net you a dreaded zero, but it’s theoretically possible to win with any other score. If your opponent has a tough day and strikes out in four lines, even a pathetic score of 1/30 could still lead to victory. For example:

🟪🟪🟩🟦 (0 × 4 = 0)
🟪🟪🟦🟦 (0 × 3 = 0)
🟦🟦🟦🟩 (0 × 2 = 0)
🟨🟨🟨🟨 (1 × 1 = 1)
🟩🟩🟩🟩 (2 × 0 = 0)
🟦🟦🟦🟦 (3 × 0 = 0)
🟪🟪🟪🟪 (4 × 0 = 0)
1/30

When Sean and I text our results to each other, typically a few minutes after midnight, we leave off those line-by-line numerical scores but manually type in the aggregate below the grid:

Connections
Puzzle#253
🟪🟪🟪🟪
🟦🟦🟦🟦
🟩🟩🟩🟩
🟨🟨🟨🟨
30/30

I’m actually more than a little surprised that I got a perfect score today. It took me a few minutes to see through the misdirection and solve the grid. But after I figured out the categories, three of them seemed to be of roughly equal difficulty. I wasn’t confident at all with my guess for purple. Sometimes you just get lucky.

The more I play, the more I’ve been thinking that Sean and I can’t be the only ones who’ve found a way to make Connections — or any endeavor — more difficult. Have you ever tried making something harder for yourself in order to make you like it more? Or do you associate ease with enjoyment? Tell me your thoughts in the comments. And if I haven’t just ruined the game for you, show me your scores from today.