Close observers of XWord Info will notice some recent changes in how puzzles are displayed, particularly around how debut words are reported. Now that we have the entire NYT history, we can more accurately display results.
Debut words are now truly debut words
The simplest definition of a debut answer word is one that has never before appeared in any NYT puzzle, old or new, daily, Sunday, or Variety. That's what we show now. Each puzzle has an Answer Summary at the bottom of its page where you'll see up to three different counts: number of words unique to that puzzle, number that debuted in that puzzle but were reused later, and number of words that only appeared in pre-Shortz puzzles. Details are provided when you click the Analyze button.
So, for the very first NYT crossword in 1942, every single word is a debut. Nine are still unique and in most cases it's easy to see why they weren't reused.
Interestingly, the very second NYT crossword already reused four answers. ABLE and TEMPO aren't too surprising, but who can explain why TASSO and LASSITUDE repeated immediately?
By comparison, the first Shortz-Era puzzle, Spectral Analysis by Peter Gordon, contains 6 words that are still unique, and an additional 8 words that debuted in that puzzle and were later reused, so Peter Gordon gets credit for introducing 14 new words there.
What words has your favorite constructor debuted?
The Crossword Innovators page sorts constructors by number of words they debuted. Because we count only true debuts now, these counts have dropped, sometimes drastically, but clicking on names to see the list of debuted words can be fascinating. I like to think it's possible to learn something about a constructor's interests and personality. Here are a few examples: Karen Tracey, David Steinberg, Lynn Lempel, and Kameron Austin Collins.
The top three innovators are impressive if only for the sheer number of words they contributed: Patrick Berry, Manny Nosowsky, and Elizabeth C. Gorski.
(The fine print at the bottom of each of those pages notes that the per-constructor stats count only for Shortz Era words debuted by Shortz Era constructors. We had to draw the line there while pre-Shortz credits are still unreliable and incomplete.)
Who cares?
If we're going to report stats like debut words, they might as well be as accurate as we can make them. While this may or may not be of interest to you, Jeff and I are always surprised at the mail we get whenever we remove or change a feature. Even Freshness Factor generates a lot of feedback. Speaking of which, Freshness Factor still considers only other Shortz Era puzzles (since older words are less likely to be relevant) but it does now consider Variety puzzles which led to some erosion. Percentiles mostly stay the same and that's what's most useful to look at anyway.
Doing all this work required big changes to the underlying software, so we also took the opportunity to do some performance tuning and most pages should load faster, sometimes significantly faster, than before. Any big code change is likely to introduce bugs so please let us know if you find anything amiss.
Thanks, y'all! This stuff is fascinating!
Bug: The intriguing words on Elizabeth Gorski's innovator page that contain '#' marks don't actually work as links.
Posted by: Pavel Curtis | 12/01/2015 at 02:53 PM
Ah, good catch. Thanks, Pavel. It's fixed now.
Posted by: Jim Horne | 12/01/2015 at 03:04 PM