Live data from Hacker News

Here's Waldo

slate.com

31–40 of 56 posts

Re: Here's Waldo

#31
post #25
post #5

This kind of spatial scan statistics is really easy to misuse without careful multiple-comparison corrections. Would they have reported a 52% effect? a 51% effect? What about 1.4 inch bands? or 1.2 inch bands? That "remarkably thin" 0.3 percent will disappear pretty quickly. Sorry that this is dismissive, but this piece is just a slightly more complicated case of http://xkcd.com/882/

It's not a statistical model for future predictions. The whole point here is to over-fit as much as possible, because you only care about a single dataset.

Sorry, but if that were the case, then just showing the list of pages and their locations would get 100% of them, with 0% probability of being "by chance".

Then, you say "that's a complicated rule: no one can memorize the list", and that's fine. But finding the simplest rule that has the best fit can be shown, via the Minimum Description Length principle, to be equivalent to models with deep connections to statistical learning.

The gist of the matter is: even if you have a single dataset, statistical considerations are very important to make sure that the statements you make are sensible. That does not appear to be the case in this Slate piece.

Re: Here's Waldo

#32
post #18
post #13

Earlier quoted context omitted.

I've always been told that it's dumbing down, usually the line is "American children wouldn't understand it". I suppose it's the same as the Harry Potter books, "Philosopher" changed to "Sorcerer" in the title. It's very prevalent to change "foreign" names to Americanised ones.

Perhaps philosopher has a different connotation in UK English. Sorcerer just seems more appropriate in a book about magic.

The term "Philosopher's stone" is from medieval european alchemical lore.

https://en.wikipedia.org/wiki/Philosopher's_stone

I don't know if they imagined british kids all knowing about that, but I can assure you the term would be just as fantastical to them as to americans.

Re: Here's Waldo

#35
post #6

He's not at the edges of the page partially due to the print process. You need a large border of disposable material that can be cut from the edges (bleed) should the alignment of the pages be slightly out when the book is bound and trimmed. It wouldn't be much finding half of Wally, so they don't put him where he might be cut in two with a guillotine. (He's still better as Wally than Waldo.)

I think the margin of error for this kind of printing is usually only 3mm though, so they could still put him pretty close to the edge if they wanted to

And do, on some of them according to the map.

Re: Here's Waldo

#36

Slightly off topic, but I've always wondered... Is there any particular reason why they changed his name for the US editions? Is there something about "Wally" that sounds weird to Americans?

I don't believe (note: no statistics were examined in the making of this answer) "Waldo" is a particularly American name. "Wally" has at least some traction, if not particularly recently.

Re: Here's Waldo

#37
post #27

Earlier quoted context omitted.

I'm not sure if that applies in this case... There's nothing to 'understand' about the name Wally or Waldo. They're both just random, invented names, AFAICT.

"Wally" would be a shortening of "Wallace", both are quite old names.

Or "Walter", which seems more common in the U.S.

(This comment brought to you by the American Society for the Prevention of Cruelty to Statistics.)

Re: Here's Waldo

#39
C:\TAD\Text\Words\AUGUST.TXT

eed inappropriately called heaven and earth; and so, that the universal creation, which God made in His Wisdom, that is, in the Beginning, was comprehended under those two words? Notwithstanding, since all things be made not of the substance of God, but out of nothing (because they are not the same that God is, and there is a mutable nature in them all, whether they abide, as doth the eternal house of God, or be changed, as the soul and body of man are): therefore the common matter of all things vis

Re: Here's Waldo

#40
post #31
post #25

Earlier quoted context omitted.

It's not a statistical model for future predictions. The whole point here is to over-fit as much as possible, because you only care about a single dataset.

Sorry, but if that were the case, then just showing the list of pages and their locations would get 100% of them, with 0% probability of being "by chance". Then, you say "that's a complicated rule: no one can memorize the list", and that's fine. But finding the simplest rule that has the best fit can be shown, via the Minimum Description Length principle, to be equivalent to models with deep connections to statistica…

He made no claim to minimal or maximal anything. His technique does give you an edge, though.
Post reply on HN