Solution: Race for the Galaxy

Written by Jon Schneider

In this puzzle, solvers must successfully complete a randomly-generated mini-hunt comprised of ten short (mostly standard) puzzles and a meta. Ordinarily this would be simple, but there’s a catch – if solvers do not complete the puzzles quickly enough, they must start over at the beginning. Each hunt (or as the puzzle refers to it, “race”) starts with 3 minutes on the clock; solving each puzzle adds an additional 3 minutes. Whenever the clock runs out, solvers must restart at the beginning of a new race. Completing all ten puzzles and the meta reveals the answer of this puzzle, which is STRONGER.

Some general tips for solving these puzzles speedily:

  • There are many online tools to help with standard wordplay tasks (anagramming, finding words with missing letters, solving cryptograms/wordsearches, etc). We mention these below in specific subpuzzles where they are useful (but some tools are more generally useful).
  • Some subpuzzles (e.g. subpuzzles 1, 2, 6, and 9) can be easily decomposed into elements (“subsubpuzzles”) which each give one letter. It is not necessary to solve each of these elements! It’s often much faster to solve a carefully selected few, and then find the unique word which fits those letters.
  • Likewise, subpuzzles which decompose into many elements are easy to parallelize with a team of people.

Here is a description of all of the subpuzzles in each mini-hunt.

Subpuzzle 1: Encodings

Each letter of the answer is encoded (one per line) in one of four classic puzzlehunt encodings: Braille, Morse code, Pigpen, or Semaphore.

Subpuzzle 2: Anagrams

Each line of this puzzle contains the sorted letters in a word with a unique anagram (ideally – since the wordlist is incomplete, there are occasionally words with multiple anagrams; usually the more common word is the correct one). The first letters of these words spell out the answer.

Many online tools exist for finding anagrams: Internet Anagram Server, nutrimatic, and Andy’s anagram solver are just a few.

Subpuzzle 3: Dropquote

This puzzle is a dropquote. Solving the dropquote reveals an “answer phrase” disclosing the answer to the puzzle (e.g. “GREAT WORK RAINBOW IS THE ANSWER TO THIS PUZZLE”).

I am not aware of any automated dropquote solvers, but tools that can do regular expressions lookups (e.g. nutrimatic) can be helpful to find individual words in the dropquote. Getting a feel for what the answer phrase can look like is also very helpful for solving this subpuzzle quickly.

Subpuzzle 4: Cryptogram

This puzzle is a cryptogram. Decoding the cryptogram leads to an answer phrase revealing the answer.

Quipqiup is a very good online tool for solving cryptograms (the correct answer is usually somewhere in the top couple candidates quipqiup outputs). While very rare, it is occasionally possible to get a randomly generated cryptogram with an ambiguous answer (if e.g. many of the letters of the answer do not appear in the rest of the answer phrase) – sorry if this happened to you!

This puzzle is a word search. Solving the word search and reading the unused letters reveals an answer phrase.

There exist some automated word search solvers on the internet (e.g. this one on, but I find it is also very quick to copy into a spreadsheet/image editing program of your choice and solve it by hand (especially since the random generator tends to generate relatively easy word searches with many words in the same general directions) – plus, it is good practice for subpuzzle 10.

Subpuzzle 6: Celebrity Identification

This puzzle consists of a sequence of photos of famous actors and actresses, each with a green number in the upper-left corner. Indexing these numbers into the full name of the pictured celebrity reads out the answer.

These celebrities are generally very famous (they were pulled from some collection of “top x actors/actresses” lists on IMDB), but if you do not recognize some of them or cannot remember their names, many online reverse image search services exist (e.g. Google Reverse Image Search and Tineye).

Subpuzzle 7: Spot the Differences

This puzzle contains three 10 by 10 grids, side by side. The first two grids each contain a variety of colored symbols and letters - specifically, they each contain one of the symbols, ‘o’, ‘x’, or ‘+’, are colored either red, blue, green, or yellow, and lie on a background similarly colored either red, blue, green, or yellow. The third grid simply contains 100 letters.

To solve this puzzle, we must note that the first two grids are always almost identical, but some pairs of corresponding cells differ (either in the symbol, the symbol color, or the background color). Taking the locations of these cells and reading the corresponding letters in the third grid spells out the answer.

Subpuzzle 8: Word Square

This puzzle contains a 6 by 6 grid, where each cell contains 3 letters (arranged in alphabetical order). These letters can be divided into 6 words running horizontally (one per row), 6 words running vertically (one per column), and 36 remaining letters which spell out an answer-phrase (reading left to right, top to bottom).

As with subpuzzle 3, tools that can perform regular expressions searches can easily find possible words for each possible row and column (the word in a given row/column might not be unique from just the letters in the row/column, so some logic might be needed to decide which word to use).

Subpuzzle 9: Multiple Choice Logic

This puzzle contains a sequence of multiple choice questions, each with 5 answers. Most of these answers are self-referential, and make reference to the correctness of other answers for the question (e.g. “A) Either answer B or C is correct.”). In each question there is exactly one subset of possible answers which can simultaneously be correct. Interpreting this subset as five-bit binary (where true answers are 1s and false answers are 0s), each subset can be converted into a single letter from A to Z. These letters (for each question) spell out the answer.

Since these multiple-choice questions are randomly generated, there might not always be a clean logical path to arrive at the correct subset of answers. One strategy which seems to work well in practice is just assigning arbitrary truth values to each of the answers at the beginning, and updating individual truth values until everything is consistent.

This puzzle is identical to subpuzzle 5 (Word Search) except the entire word search has been encrypted with a simple substitution cipher (i.e. a cryptogram). The words to find remain unencrypted. As before, the unused letters spell out an (enciphered) answer phrase.

This puzzle tends to be the most time-consuming puzzle for most solvers. One good way to get started is to find a relatively long word (so that it has few possible positions) with a few repeated letters (so you can confirm identical letters). Another good strategy for this subpuzzle is to work on it in a spreadsheet, so that you can find/replace letters with their deciphered letters as you find these correspondences.

Meta Puzzle

All the answers for the previous ten subpuzzles feed into a meta-puzzle. This meta puzzle is randomly chosen from one of three possible metas:

  • Last letters: the last letter of each answer will spell out the meta answer
  • Diagonalization: Diagonalizing all ten answers (i.e. taking the ith letter of subpuzzle i) will spell out the meta answer.
  • Double letters: Each answer has a unique consecutive pair of identical letters (e.g. the two Ds in HIDDEN). These letters spell out the meta answer.

Bonus Challenge

Can you solve this puzzle without any external aids (i.e. online/computerized tools, other people)?