Digits, and where they fall

Every character in each address is tested against the ten digit characters, and the matches are counted. Across all three addresses there are 25 digits and 143 letters. The bar tile splits the digit count by address, and the gauge expresses the total as a share of the 168 characters in the set.

Only six digits are actually available in this alphabet, so the share is lower than it would be in a general purpose string. Four digit characters cannot appear at all, and the ones that cannot are the ones easily mistaken for letters.

The third tile cuts the first published address into seven equal blocks and counts the digits in each. The line joins those seven counts.

It is a lumpy line. Some blocks carry several digits, some carry none. This is the shape randomness produces, and it is also the shape that most invites a story. People look at a run of digits in one part of a string and feel that part is somehow denser or more important. It is not. Move to the next address and the lumps are somewhere else entirely.

Try the same reading on the other two addresses using the same rule and you will get three different lines with nothing in common. If any pattern held across all three, that would be worth reporting, and there is none to report.

In a string that is mostly letters, digits stand out. The eye lands on them, uses them as landmarks, and starts navigating by them: the bit after the seven, the part with two numbers together. That is a genuinely useful reading strategy for finding your place, and a genuinely bad one for checking.

The problem is that landmarks make you confident about the parts you can see and vague about the parts between them. A string that matched on every digit and differed in the letters between them would pass a landmark check comfortably, because the landmarks are the only thing being compared.

• Nothing about identity. Two different addresses routinely carry the same number of digits.
• A weak sanity check at most. An address with no digits at all or an implausible number of them might be worth a second look, and even then only as a prompt to compare properly.
• Understanding your own reading. If you know you navigate by digits, you know which parts of the string you are skipping.

The digit metrics on this panel are the easiest numbers on the whole board to produce and among the least useful once produced. They are here because the board publishes what it can compute and then says what the computation is worth, rather than only publishing the flattering ones.

If you want the one number on this panel that changes behaviour, it is not any of the counts. It is the observation that digits act as landmarks, and that landmark-based checking feels much more thorough than it is.

Why can only six digits appear?

The encoding uses twenty six letters and six digits. The four excluded digits are the ones that look most like letters, which removes a whole class of transcription confusion.

Do more digits make an address harder to read?

For most people slightly easier, because digits break up long letter runs and give the eye somewhere to rest. Easier to read is not the same as easier to check.