Dordle Solver

Dordle's arithmetic is simple: seven guesses, two answers. Treat it as two separate Wordle games and you need twelve guesses and you lose. Treat it as one shared game and seven turns is enough, but only when most of your guesses earn progress on both boards.

The solver's ranking encodes exactly that math. A candidate word gets scored by how much information it reveals across both candidate lists combined, never on a single board, so the suggestions naturally favor words that pull weight in both puzzles.

Early in the game the two boards are basically identical, both are full five-letter dictionaries, and that's when shared guesses are cheapest and most valuable. It's also why the opening matters more in Dordle than it does in Wordle: every early guess is doing double duty whether I like it or not.

The same opening logic that works in Wordle works in Dordle, with one extra requirement: I want the opening word to be a plausible answer on either board, so its feedback is useful in both columns.

Classic five-letter openers like CRANE, SLATE, or ADIEU are strong because they cover common vowels and consonants without repeating letters. Any green or yellow I get applies to a word that could show up on either board.

A two-word opening, CRANE first and then a word built from whatever letters it uncovered, usually leaves me with a decent shape on both boards by turn two, five turns left, and enough information to start making real choices.

I know ADIEU is controversial among Wordle players, but in Dordle it has a real use: it burns four vowels in one shot, which tells me a lot about both boards at once even when it rarely is the answer itself. On days when I want information fast, it's my opener; on days when I want to actually solve, I fall back to CRANE.

The hard skill in Dordle is reading two grids at once. One guess produces two feedback rows, one per board, and they almost never agree. A letter that's green on board one can be gray on board two, and the moment I let one board's feedback bleed into my mental model of the other, I start building candidates that can't possibly be right.

The solver removes all of that load. I tap the colors for each board and it keeps two completely separate candidate lists. My only job is to enter what I actually saw, and the solver handles the bookkeeping of what's true on which board.

The discipline I've had to drill into myself is simple: a letter's color on board one has zero bearing on board two. The players who run out of turns are almost always the ones who merged the two boards in their head.

Once board one is solved, every remaining guess is a single-board game with a shrinking budget, and that's where I used to lose. The fix was counterintuitive: use each guess to eliminate as many candidates as possible, even if the word I type isn't the answer. The solver ranks candidates by elimination power in exactly this situation.

If I have two guesses left and several candidates still alive, I look for a word that could be the answer itself rather than a pure elimination play. The solver balances both options and tells me which is safer, which matters most right when my nerves are the worst.

The other endgame habit I've built: when a board is solved and the other still has a few candidates, I stop looking for the exact word and start looking for the guess that splits the survivors in half. Two turns of that beats three turns of guessing the word directly, and the solver's ranking shows me which split is cleanest.

Behind the scenes the solver scores each candidate against both remaining lists and reports a combined value. A word that's a plausible answer on board one and reveals strong letters on board two scores far higher than a word that only solves one board. That's why some suggestions look odd at first: a word that isn't the answer to either board can still be the best guess because of what it reveals across both.

It also explains the endgame behavior. When the two boards share almost no candidates, the ranking quietly switches to single-board mode, which is exactly what I'd do if I were thinking clearly under pressure.

One honest limitation: the solver can't read the board for me. I type in the colors I saw, and if I fat-finger a gray into a green, both candidate lists drift. It's also not magic on a board that's still mostly gray after the opener; below two or three colored tiles there often isn't enough information for any tool to do more than guess.

Dordle releases one two-word puzzle a day, and the answer pairs are a small but revealing dataset: the two words rarely share letters, which is exactly what a well-designed pair looks like, two words that force you to sweep a wide letter set. I've started tracking the daily Dordle answers in a note, and the pattern that keeps showing up is the pair's independence, not any single word.

The solver doesn't care what day it is. It maintains both boards, filters both lists, and ranks the shared guesses the same way for today's puzzle or an archived one. The daily cadence only changes which words are in play, never the arithmetic.

The solver also handles the longer variants Dordle players sometimes switch to. Two candidate lists, filtered in parallel and ranked by combined value, works the same whether the words are five letters or more, because longer words change the pool, not the arithmetic. For archived puzzles it works on any date; I just log each board's feedback as I saw it and the two lists stay separate.

Dordle sits exactly between Wordle and the multi-board monsters: one extra board, one extra guess, and the shared-guess mechanic that makes it interesting without being overwhelming. Players who get the two-board discipline down find Quordle and Octordle far less intimidating afterward, and I've watched that progression happen in my own play.

The solver bridges the same gap. It teaches the combined-value ranking that the bigger games need, on a scale where I can actually follow what it's doing. Learn Dordle with the solver and the eight-board game stops being a wall.

I keep Dordle in my rotation specifically because it's the smallest step up from Wordle. The shared-guess idea clicks in two boards in a way that's hard to feel in eight, and I think it's the best place to learn the combined-value thinking that carries over to every bigger game.