Monday, March 02, 2009

More crosswords and statistics

So my scans are up, and I found my other error in puzzle 5 (I had incompletely removed one governmental organization to replace it with another governmental organization).

Anyway, the tab-delimited file only goes to puzzle 6 (I don't know why), but I think I figured out a way for Excel to tell me an estimate of the number of people who got a puzzle right (they must have scored at least 10w points, where w is the number of words, and the excess over that value must be a multiple of 25). This formula will (incorrectly) count people who have 5n words wrong, for any multiple of n and who had enough of a time bonus to deal with the at least 125-point hit (you can get three letters wrong and miss five words, so 50 points for the words and 75 points for the lost time on letters). I can't say how often that happens. This is my data:
Puzzle 1549
Puzzle 2303
Puzzle 3362
Puzzle 4636
Puzzle 5111
Puzzle 6500
Puzzle 7449

Remembering that there were 675 participants, that means that I estimate 81.3% of all solvers correctly solved puzzle 1, while a whopping 94.2% of solvers correctly solved puzzle 4 (the high), and 16.4% solved puzzle 5 correctly (the low). I wonder how often puzzle 4 outperforms puzzle 1 like that, and by how much.

There were clear winners on puzzles 2 and 4 and 7.

Because it's my blog, statistics about me:
Puzzle1234567
Median>10659751297.5121048015951855
Me1165147015251310115519452375


These have been UPDATED now that data for puzzle 7 is available. I have checked that the formulas are counting exactly the numbers I want them to count; again, I cannot say quite how many people "luck into" a score that could be a correct score.

No comments: