meta 2:30
hello and welcome to episode #652 of matt gaffney’s weekly crossword contest, “Solved in Full”. for this week 4 puzzle, the instructions tell us that the answer is a familiar 5-letter proper noun. what are the theme answers? they’re not the longest answers in the grid, but there are five different answers of the form [letter]-[word]:
- {Amazon’s business} E-COMMERCE.
- {“Fun, Fun, Fun” Ford} T-BIRD.
- {Hoops} B-BALL.
- {Weapon used in WWII} A-BOMB.
- {Genre for BTS} K-POP.
the title suggests the next step: we’re interested in the full word shortened to a single letter at the start of each of these theme answers. so ELECTRONIC commerce, THUNDERbird, BASKETball, ATOMIC bomb, and KOREAN pop. those five are hiding in disguised form in five across answers. disguised how? well, they’re anagrammed with one letter missing (what some people call a transdeletion). from top to bottom in the grid, we have:
- {Outperforms} BEATS. that’s BASKET without the K.
- {Creature sometimes spotted at Mayan ruins} COATI. ATOMIC without the M.
- {Bodø dough} KRONE. KOREAN without the A.
- {Went looking for a buck, maybe} HUNTED. THUNDER without the R.
- {Patch things up} RECONCILE. most spectacularly, this one is ELECTRONIC without the T, and it’s the one i noticed first, since i went looking for ELECTRONIC first. i suspect the E-COMMERCE/RECONCILE pairing was the genesis of this puzzle.
taken in this order, the missing letters spell out K-MART, which is a familiar 5-letter proper noun, and also itself fits the form of the original five theme answers. so that is the meta answer.
this is an excellent meta, cleanly executed and with several satisfying aha moments. although there were no huge intuitive leaps necessary, finding each of the transdeletions was a mini-thrill, and the final answer was a nice extra click.
it was also, i think, much easier than last week’s meta. once you locate those five theme answers, you almost can’t help but solve it, because the next step is always strongly suggested. and while it’s true that it’s not obvious at first glance that those five theme answers are the theme answers, the first two of them are easy to identify as such based on their location (17- and 19-across) and parallel formation, and from there it’s a simple matter of noticing the other three. the instructions suggest looking for five things, so you basically know when you’ve found them, and then the title tells you what to do next. so i think this is basically the week 3 meta that we didn’t get last week, and that was the week 4/5 meta we should have had this week. works for me! although i didn’t get last week’s, i was grateful for a relatively easy one over the holiday weekend, since i was away from my puzzles until sunday. overall, it was a satisfying slate of november puzzles even though the difficulty ordering was maybe a bit wonky.
that’s all i’ve got. how’d you like this one?
Thanks, Joon. 316 right answers this week.
You’re right, wonky difficult levels this month: 1.75, 1.5, 5.0, and 3.0
No no no. I solved a week four puzzle and you can’t take that away from me!
full credit restored, my apologies
Very beautifully executed, but i was really thrown by the long answers not being germane to the meta. More my failing than anything, but learning to focus on “theme” material hurt here.
+1
A-BOMB can also be ATOM BOMB, so I used TOM at 4D to extract an A, leading to the letter set KAART. I first anagrammed this to KARAT (not a proper noun) but then found A-TRAK, which was not familiar to me but seemed notable enough to be an answer, and the hyphen was a strong enough click that I submitted it. Matt accepted it, but I see now that K-MART is a much better answer, being more familiar and not needing anagramming.
Actually we had some leaderboard glitches over the weekend and I did not meant to accept A-TRAK. But I will send it to the panel this afternoon. Two other solvers besides you submitted A-TRAK. That was an unintentional but truly vicious red herring!
IANAP, but to my mind, the unqualified description “a familiar [5-letter proper noun]”, implies the expectation that almost every solver will already be familiar with it, a bar I rather doubt A-TRAK meets. I’d think that ought to cast enough doubt on this solution for one who has already considered both ATOM and ATOMIC as valid expansions of the A in A-BOMB to at least scan the grid for “transdeletions” of the latter.
I agree that A-TRAK shouldn’t qualify as “familiar”, but he does have considerable enough Wikipedia entry where one could perhaps make an argument. I’d argue that if it weren’t hyphenated, then it definitely shouldn’t be accepted; but the 1-4 pattern was a very strong click, and in my biased opinion I think it’s a valid alternate answer.
That said, if the committee chooses not to accept, I’ll begrudgingly accept their ruling.
Fellow A-TRAK’er here. :). Figured it was wrong with the “familiar” requirement, but I somehow missed COATI despite scanning the grid numerous times. Doh. Googling “a bomb” only showed atom bomb, further pushing me down the wrong path. There’s always next week?
I did find out that A-TRAK was part of the DJ duo that recorded “Barbra Streisand”, and now that useless bit of trivia will be forever logged in my brain. https://www.youtube.com/watch?v=wWhtcU4-xAM
I spent a long time trying various approaches to handling the five theme answers. The 7th try looked promising. I took 1st letters of the eliminated parts of the words represented by a single letter – eLectronic, tHunder, bAsket, aTomic, and kOrean. This produced LHATO, which gave me a possible answer of HATLO. He was a beloved cartoonist for my generation, but I doubted Matt would use him, as many younger solvers are likely unfamiliar with his work. Furthermore, Matt almost never uses anagrams. So I kept digging until I hit the right answer.
joon wrote: “i suspect the E-COMMERCE/RECONCILE pairing was the genesis of this puzzle.”
That was the *last* part, believe it or not. Turns out to be a very small set of usable entries. Was looking to use D-MARK as the meta answer for a while but didn’t work. Wanted K-MART but had no usable T. Was about to give up when I decided to make one last search. Desperation mode led me to see if ELECTRONIC (from e-mail etc.) could lose the T and make a 9-letter word and — holy cow, RECONCILE. Lucky.
the odds against this are so long as to strain credulity. “does this random 9-letter string have a common english anagram?” no way!
To be fair, the odds of a random 9-letter string having a common English anagram are much, much lower than the odds of a 9-letter string formed by deleting a letter from a 10-letter English word having such an anagram, as the distribution of letters in English words is very far from random. I will now put this to an informal test by assaying the first five OneLook results for “??????????”, sorted by commonness. (I don’t know what the results will be in advance, and I will post them in this comment either way).
1. ABSORPTION = PROBATION S, ABORTIONS P
2. RESOLUTION = N/A
3. RESISTANCE = CANISTERS/SCENARIST E(x2), INCREASES/SCENARISE T, SINCEREST A, SECTARIES N, IRATENESS C.
4. METABOLISM = IMMOLATES B
5. DEPRESSION = PERSONISE D, DISPENSER O.
All anagrams found by The Internet Anagram Server.
By my count, 11 of 50 possible deletions from these five 10-letter words yields a string of letters anagrammable to a nine-letter English word. For this purpose, I counted each letter of the 10-letter words as unique (so that deletions of repeated letters were counted multiple times, for better or worse), but did not distinguish between strings that had multiple anagrams and those with only one. You may prefer to count a little differently, in which case, you’re welcome to take the data above and recount it as you prefer.
Regardless of such fine details, the results of this experiment suggest to me that 20% is a reasonable rough estimate of these odds (which is the only sort of estimate I’d feel entitled to base on such a small sampling of words, in any case).
But it had to leave the T, not just any letter
Unless you have a reason to think that deleting the T from a 10-letter word is less likely to result in an anagrammable string than deleting a different letter, the particular letter you needed to work should be immaterial to the odds that it would work. If about one in five deletions of any letter from a 10-letter word yields such an anagram, then about one in five deletions of T from a 10-letter word should do likewise.
To put it slightly differently: once you realized that you needed to find an anagram for ELECTRONIC – T, I’d say your odds of finding what you wanted were statistically ~ 1 in 5. On average, around two of the 10 letters in ELECTRONIC should work this way; if you take that as a given, then T has about a 2 in 10 chance of being one of the 2/10 letters that works this way, for this word. Does that make sense?
You still got lucky, mind you!
Lots of ways you could test this more formally, but using basically the widest possible definition of “word” and only using the letter “t” it looks like the probability is a bit less than 10%. This drops way down in my testing (around 3-4%) if you only use common words (for some definition of “common”).
https://pastebin.com/eEqHdy3h
Nice. That’s obviously a more accurate figure than the one yielded by my very rough estimate above. What source[s] did you use for ‘common’ words? Edit: never mind the question: I looked at the directory of your source URL and saw the lists there.
@Alex, I altered your code to use the word list called “count_1w.txt” (described, oddly, as ” The 1/3 million most frequent words [in the Google Web Trillion Word Corpus“, though it only appears to contain 200,000 words) and got a similar result to yours with the ENABLE list: 1352/14322, or about 9.5%, so that seems to be a reasonable estimate. What list of ‘common’ words did you use that yielded only 3-4%?
It’s worth noting that, while paring your root list of 10-letter words shouldn’t particularly skew the outcome one way or the other, paring the list of 9-letter words will tend to reduce the number of hits you get. Thus, for instance, when I used the SOWPODS corpus of some 268,000 odd words (many of them admittedly rare), the hit ratio rose to 13.5%, when I tried with a file of the most common 100,000 words (from the same site), it plummeted to 4.1%, and if you hypothetically imagine trying it with a tiny list of, say, the five commonest words in the language, it’s easy to see that it would most likely go to 0%.
Thus, for the most accurate estimate, we would need a word list that really does contain all admissible 9-letter words. The smaller the subset of that list we use, the smaller we should expect the percentage to fall.
For what it’s worth, using the 100,000-commonest words list for 10-letter words together with one of the larger lists of words admissible in Scrabble-like word games for 9-letter words. I got 11%. Of course, these lists include many unfamiliar words that would not make good crossword entries, but, on the other hand, they also exclude all proper nouns, many of which make perfectly fine crossword entries. If Matt is willing and able to run code like yours on his master list, we can get a more or less definitive number, but short of that or a close approximation to it, ~10% looks to me like a reasonable estimate.
Am I the only one that was led astray by 65a? I mean, I get the double meaning, but as a phrase, “Apple pieces” is barely in the language, so it seems like pretty shaky ground to build on just for a pun. (Google says 452,000 results, compared to say “Lego pieces” at 1,360,000 results.) I figured it had to be thematic.
This is two weeks in a row that I failed to solve the meta due to my mis-/over-interpretation of the word “piece”.
End of rant. I’ve said my piece.
This answer also had me befuddled, as you’ll see later. I assumed that the clue was referring to the Apple company’s inventory, or ‘pieces’. So the answer made sense, but barely.
Strangely, “Apple pieces” helped me, insofar as I thought it might have something to do with iTunes – which worked with the answer (TUNES) above it. That’s pushed me to look at [letter]+[word], even though it *wasn’t* part of the solution itself.
BTW, another sister city of Muscat, Oman? Tunis. Way too much like TUNES also.
There was a puzzle a few months ago that must have lodged into my head…I remember Target and Subway were 2 big companies mentioned in oddly worded clues. This one looked similar, and there were 5 big companies in the clues for a 5-letter answer. Hmmm. I associated…
Ford – LYFT
Ikea – KRONE
Apple – BEATS
but couldn’t match Staples or Amazon with anything in the grid.
I also (thanks, WSJ!) obsessed over the 28-down RADII since it had “segment” in the clue, which is like a “piece”. I nearly Hail Mary’d KEITH, which radiates, well, I guess *to* the center (not *from* it) from the bottom right corner.
Note that the currency being clued at 38A is the Norwegian krone, while the Swedish currency is the krona, so the Ikea connection is tenuous.
https://youtu.be/ZYgCAid1JJo?t=7
It takes me longer than most people to solve these later month metas. I had found the 5 ‘themers’ and I finally made the breakthrough to the next step when I spotted RECONCILE looking close to electronic. Then I had found the M-A-R-T and was convinced the answer was MARTA or MARTY, common proper nouns. STOCK was the only answer with a K to make from BASKET, so I was confused. The answer was specious, but the crosses were solid.
I finally got out of that rabbit hole and found BEATS, and I realized what a great meta this was!
Great job, Matt!
The rabbit hole I got stuck in was with Lisa Loeb. I’m curious if cluing her with a “Fuller House” reference was an intentional red herring– there would certainly be other ways to clue her without including one of the title words.
Once I got past it I was off to the races and loved the puzzle!
+1
Did anyone else notice TURNME diagonally starting from the “T” in UPTO? I thought there was a connection with “Full” in the title and TILT (46A). TILT + TURNME had me rotating the grid for a while looking for diagonal words. I also thought maybe the O right in the middle of the grid might double as the center of an LP but I didn’t find anything as I turned the “record.”
I spent the weekend playing Twister, with each hand/foot in a different rabbit hole.
1) There seemed to be lots of names in the puzzle, so I wondered if “Full” was meant to refer to “Full names”; went off looking up a bunch of celebrities’ middle names.
2) I noticed TILT. “Full TILT”? I then noticed BED. “Full [size] BED”??
3) I noticed SATED. Hey– that *means* “Full”!
4) I noticed B-BALL. “Full court”? I then noticed RADII. “Full circle”??
And — like one other commenter mentioned above — one of the long Across answers has “Fuller” in its clue; that’s the hole in which I buried my defeated head…
A very fair meta! Too bad I was no match for it!
So I’m the only one who looked to the 5 letter “hidden capitals” in the clues (Petty, Apple, Tours)? I never got past that.
SUCRE is neatly hidden in ACCUSERS
It’s nice to see I was in good company , and many of the potholes others got stuck in I drove into after them.
One I did not see mentioned is that there were 5 “fill in the blank” clues as well. I spent a good bit of time trying to use parts of the clue to come up with the 5 letters of the word. I also fell for the TEEM and SATED – both synonyms for full, as well as the FULL TILT, FULL BED (thankfully “Nickname for the British Field Marshal in command of Allied Ground Forces in Normandy” wasn’t there). Even after locating the 5 letters and seeing TOM, which was the natural “completion chunk” for A-TOM I got terribly lost since I thought I had to add EACH of the letters to a word to get what I needed. BASKET was easy, RECONCILE was amazing but then KOREAN…already used the A, now what? FINALLY I saw COATI and realized what was going on (yes I’m a little slow).
What I find VERY interesting, and I assume was part of the prestidigitation of the creator, is to have a term like KPOP, from which we get a K, and then to have to find Korean WITHOUT using the K…if that was not a “quick look over here” while the magician hides the coin in his collar then it was just an amazing bit of chance.
I was lucky to get this one pretty quickly but I was momentarily hung up on BREXIT – British Exit. I thought that fit with “Solved in Full” similar to the others that had only one letter to fill in. Once I abandoned that one I had it.
I got this one fairly quickly, and it was a fine puzzle – just wanted to comment on 46A: [Rarity for a pinball wizard] = TILT. Counterintuitively, this is not true. “Pinball wizards” nudge and shake the machine at every opportunity to keep the ball in play. The common saying is “If you’re not tilting, you’re not trying.”