What's a Good Wordle Score? A Solver Benchmark

An average near four is a solid Wordle score. This site's solver averages 3.52 guesses, but that is a software benchmark rather than a target every human player ought to match. Row four is ordinary, and even this solver needs a fifth guess for 76 answers.

The useful comparison is direction, not competition: a balanced opener and a second guess that separates the remaining answers can bring your own average down.

How we tested this

The test applies the same setup to 2,358 answers. It opens with ROATE, considers a limited Wordle guess list, and then chooses the word expected to leave the fewest answers. That makes it a fair comparison across answers, but not a proven minimum or a like-for-like comparison with a person playing from memory.

What the solver run looks like

Of the 2,358 answers tested, 60 finish in two guesses, 1,095 in three, 1,127 in four and 76 in five. The longest route in this run is 5 guesses.

In other words, 49.0% finish in three turns or fewer, and 96.8% finish within four. Many of the longest routes contain groups of words with nearly identical patterns, which are costly for the solver to separate.

The four-turn count is useful perspective: even a systematic solver with the full candidate list spends four guesses on 1,127 answers. It does not follow that every human four is equally efficient, or that every difference is luck. It does show why row four is an ordinary outcome rather than a failure.

So is your average any good?

Do not treat 3.52 as an average every player ought to match. The program sees the entire answer list and evaluates thousands of guesses without fatigue, memory lapses or time pressure. An average near four remains close in raw turns, but the conditions are different.

The most transferable improvements come early. A balanced opener usually leaves a more manageable shortlist, and a good second guess tests the letters that distinguish the surviving words. When the shortlist forms a trap family, a deliberate splitter can be worth more than an immediate answer attempt.

The Opener Lab compares first words using the same first-turn data. Use the 3.52 result as a consistent software baseline, not a universal definition of a good score.