Anyone who has been on Twitter in recent weeks is probably intimately familiar with the grids of Wordle solutions clogging up everyone’s timelines. But those tweets give more information than it would seem. Collecting and analyzing data from millions of these Wordle result tweets can give us some interesting insights into aggregate play patterns and the relative difficulty of daily Wordle puzzles.
The Wordle Stats Twitter account has done a lot of the heavy lifting here. Since January 7, the bot account has used the Twitter API to sort through the public timeline for every tweet formatted as a Wordle result, tracking the total number of players and how many guesses each player needed to complete the puzzle. That account shared its underlying data with Ars to power a deeper analysis of daily play patterns.
#Wordle 242 2022-02-16
289,721 results found on Twitter.
10,740 hard mode players.
2: 🟩 4%
3: 🟩🟩🟩🟩🟩 20%
4: 🟩🟩🟩🟩🟩🟩🟩 31%
5: 🟩🟩🟩🟩🟩🟩 26%
6: 🟩🟩🟩 15%
X: 🟩 3%#Wordle242
— Wordle Stats (@WordleStats) February 17, 2022
This isn’t a perfectly random sample of Wordle players, of course—it’s limited to the group of players who use Twitter and choose to share their results publicly. The vast majority of what The New York Times said were millions of daily players at the end of January are not reflected here.
These results might be skewed if players are more likely to share interesting results than boring ones or if certain demographics of players are less likely to share in the first place. Players can also edit their tweeted results relatively easily, or they could cheat by simply playing the daily puzzle more than once.
All that said, the hundreds of thousands of daily Wordle results shared on Twitter are the best window we have into how the game is played in aggregate. When compared day to day, the results ought to give a decent picture of the relative popularity of the game and the difficulty of puzzles over time.
(For even more analysis going back to December’s games, check out this Medium post by a University College Dublin professor.)
Looking purely at the total shares, it seems like Wordle‘s dominance of the Twitter timeline may already be abating. Shares started at less than 100,000 a day in early January, and total shares on Twitter peaked at 360,155 with puzzle 230 on February 7, shortly after The New York Times purchased the game. Since then, sharing has slowly declined, dipping below 240,000 shares for puzzle 243 on February 17.
That doesn’t necessarily mean Wordle is getting less popular overall—new players might be less likely to be on Twitter, or long-time players may be tiring of sharing their results. But the slow decline in tweets has been interesting to see.
Warning: The remainder of this story has spoilers for Wordle puzzles through February 17. If you’re still working through the archives, finish them up before clicking through to the next page.