captcha have gotten human unreadable.

captcha have gotten human unreadable.

unless Randall himself actually verifies this, basically the only argument I have is my reasoning, so here it is.

When my mom, who, incidentally, introduced me to xkcd, sent me to wordcount.org, it was immediately my goal to think of a word that I knew that wouldn’t be on the list. their algorithm seems like it wouldn’t be terribly effective after 10000 words, since the signal-to-noise ratio is so low with word frequency. As for why the list is written down, my bet is that Mr.Science was being more methodical than I and keeping a record of his trials.

The obvious thing, is to pluralize words, find less tenses of common verbs or whatever, and try to break this algorithm that way. so evening becomes evenings, and it’s really more of a benchmark test anyway. next comes something obscure, once again pluralized, scores much higher (ranked 66358), then a proper noun again pluralized, which is surprisingly common (johns is 17608 most common word in all the manuscripts that wordcount has surveyed), parrot is a bit of a ridiculous try, it’s a common word, and it pirate themed, so naturally is going to be popular amongst the internet. Lithographer, whilst scoring very high, is not really in common usage, and therefore doesn’t count (you may as well enter zloty). legends is also a silly attempt, but it is an expected one because it stinks of arcaneness and therefore rarity, also Randall might have thought the algorithm was gathering data from a specific type of document that was very unlikely to contain the word legends. resupply scored surprisingly high (not resupplying, not yet), followed by bunting, which was lower than expected, and then randall realized that he could add an -ing to resupply (as in bunt) and drastically lower the hit-count, but not really decreasing the common-ness of the word. leg is the beginning of something arcane that begins with leg or perhaps an attempt at adding to the word legend.

since resupplying fails to be in the list, and thus satisfies the condition, it earns a check mark. the x’s could be a methodical part of this process which I have missed, they could be mental scratchwork manifesting itself as stupid doodles, which I find more likely, or my entire theory could be complete bullshit. If so, I intend to spend no extra time on this problem, and look forward to the answer.

Boggle. Someone commented on similar words within words, so Boggle would be a relatively apt explanation.

However, there’s also that there are very few smaller words on it.