OMG, we were the only ones to solve Museum?? Ok, that makes me feel good. We called it in *as* the hunt was ending (we found out on the phone it was over), if anyone's curious. I liked it a lot, fun puzzle.

It's also good to know that only a handful of puzzles were solved before we got the email and actually got puzzles - although I think by Friday night we'd figured out as much.

(Reply to this) (Thread)

Your team was the one who sent somebody down to play Tichu wearing the "hackito ergo sum" guy on a T-shirt -- at a time when I knew you hadn't seen the Museum puzzle, so I kept quiet. I figured somebody on your team would know where the mural of that guy was, even if the T-shirt wearer didn't. It is a little tricky since it is NOT in EC where the caption suggests.

(Reply to this) (Parent) (Thread)

That mural is pretty well known; most MIT students (or at least those who have some connection to the east side of campus) probably know where it is.

(Reply to this) (Parent) (Thread)

So, our team (same as porcupine8) does in fact contain the bulk of the people who have maintained the mural over the years. When the puzzle came out, I believe someone immediately commented "Hey, when was that exclamation mark added?"

(Reply to this) (Parent) (Thread)

Heh. No wonder you solved this puzzle.

In test solving, we had noticed the changes on that wall first (the ! and the stars in the Florey mural beside it, in comparison with This photo) and I pored over pictures of the lobby 2 and 6 murals we had taken for another puzzle and found the added mushroom, and I turned it over to people who would be on campus to find the rest.

Well, not before staring at the pictures closely and guessing what else might have been added. During test-solving the pictures in this puzzle were large digital-camera-sized images coded to that common width on the web page, and I was able to detect that the ashtray in the faculty lounge picture (while being entirely believable for a 1957 photo) was clearer than the rest of that scene, and hence must be phony. The author deemed this was outside the bounds of how the puzzle was meant to be solved, and reduced the image files to the size actually used in the puzzle.

(Reply to this) (Parent) (Thread)

What is the event at T=37.5?

(Reply to this) (Thread)

Which of those beige lines is IIF?

(Reply to this) (Thread)

The graph is probably easier to read when you realize that the legend reads across, and then down, ordered by total number of puzzles finished. For instance, Groovytron is the third red line from the top at the end, and IIF is the topmost beige line, finishing with the 6th most number of puzzles solved.

(Reply to this) (Parent) (Thread)

Aha, thanks. I thought so because of the blip at the end (we solved two puzzles in quick succession about five minutes before word came through that the coin had been found), but I wasn't sure.

(Reply to this) (Parent) (Thread)

The beige line that ends at 73 puzzles solved is IIF. I don't know what the event is at T=37.5; in the plots of individual teams there are lines like that when they opened a round, which means most of them are at the same times for most teams, but none of the round openings is at that moment so I don't know why it is there.

(Reply to this) (Parent) (Thread)

The vertical line at 37.5 is where the coin was found in last hunt. You probably want to go into all-teams-puzzles.php and adjust the number at the bottom to be the number of minutes that the hunt actually lasted. :)

Speaking of which, since the software that I helped write for SPIES still seems to be being used, it should probably get cleaned up -- particularly because it was originally written in the space of about 4 or 5 days before Hunt.

(Reply to this) (Thread)

The software has a few quirks, but for something written in that short a timespan, it is very impressive.

And as long as you're talking about your software -- is there an easy way for me to extract each team's solving progress into separate HTML files, or should I just do it all by hand? I'm certainly willing to do the latter, but if there's some top-secret hidden way to do the former, that would be vastly preferable.

(Reply to this) (Parent) (Thread)

There's no way to do that right now without either writing some code or doing it by hand. I actually was poking at getting the software running locally the other night, so I may be able to just test out and hand you some code to do it. Drop me an email at (my LJ username) at mit dot edu, and I'll see if I can cook something up for you today or tomorrow.

(Reply to this) (Parent) (Thread)

Be aware that I fiddled with the data model slightly, and I think they're basing this on my mods, so there may be issues with that.

(Reply to this) (Parent) (Thread)

Actually, a lot of that code *has* been cleaned up. I ended up doing a lot of cleanliness tasks when I was re-writing for Hell, and will probably end up doing a lot more this time around.

(Reply to this) (Parent) (Thread)

Glad to hear it. I'm glad the teams stopped having to rewrite it from scratch every year, but a bit embarrassed that it had to start from such a mess, and that it was partially my fault. I'd be glad to help clear my name by giving a hand, if need be. It's also the sort of thing that is more-or-less running-team independent, and could use some love to make it more friendly to teams without dedicated computer people (things like the line at 37.5 hours, for instance :)

(Reply to this) (Parent) (Thread)

BTW, how super-sekkrit is the super-sekkrit hunt software? I'm running a hunt with some people from CMU (the CMU KGB), several of whom are also Manic Sages, and we were wondering if we could use it. Last year we used a series of terrible hacks in mixed Perl and PHP, backed by a random directory in AFS for the "database"...

(Reply to this) (Parent) (Thread)

I agree. I did *try* to do that, and tried to do a reasonable job of giving Palindrome a turnkey solution. But there were bits that I failed on (and clearly I didn't bother doing so with the graph generator. :-)

(Reply to this) (Parent) (Thread)

As I did last year, I request the raw CSV data behind the Big Graph be made available, so I can play with it. :-D In particular I'd like to differentiate it so I can see what our solving rate did over time -- that's harder, though not impossible, to see in this format.

-- Manic Sage

(Reply to this) (Thread)

Seconded.

--Benevolent Dictator in Exile, II&F

(Reply to this) (Parent) (Thread)

Thirded.

(Also, if it's still around, the 2006 data would be nice. We were trying to compare our solving data between years, and just having a line on the 2006 graph doesn't quite cut it.)

(Reply to this) (Parent) (Thread)

Do you still have the 07 data?

I know I have it somewhere, but it was only my second year playing so I wasn't being all that methodical, and I'm not sure what I did with it.

(Reply to this) (Parent) (Thread)

I have the '06 data. What information do you want? The database has more information in it that you can shake a stick at. Would tuples of (team name, time solved, puzzle name) be good enough? It also has records of all of the information for answers called in, puzzles unlocked, and the like.

(Reply to this) (Parent) (Thread)

Holy crap. :-)

Well, being the information-packrat I am, I'd _like_ all of it. But the most convenient thing containing at least the mentioned tuples would be awesome.

(Reply to this) (Parent) (Thread)

BTW, note that it was the person I responded to who actually asked for the '06 data, although I would like it too.

(Reply to this) (Parent) (Thread)

I should have a copy of the '07 data somewhere; let me double-check.

(Reply to this) (Parent) (Thread)

That would be fantastic -- I had a copy at one point, but have misplaced it.

(Reply to this) (Parent) (Thread)

gwillen -- If you're still interested, I found a copy of the SQL dump of the DB and put it up at http://www.portnoy.org/~portnoy/Hunt2007-data/

(Reply to this) (Parent) (Thread)

Thanks! Downloaded.

(Reply to this) (Parent) (Thread)

Big Musical Number - solved by Manic Sages

Hah, I feel so proud! :-D

I spent multiple hours forcing everyone on the team to listen to clips from it over and over, and searching songlyrics.com for every song containing the words and starting with the letters we had left. Ultimately, of the 5 songs we didn't get quickly, I got one from songlyrics.com, we got one from an onsite solver in whose face I shoved a clip, and we got one from a remote solver; at that point someone brute-forced the answer without the other two, which were both in the denominator of things.

(Reply to this) (Thread)

In test-solving, it took us ages to find that "table" song, which was critical to the solution, and (judging by the number of times its title/artist appears on the internet) far more obscure than any of the other songs used. I urged the editors to find a way to make that one easier, but I don't think anything changed.

The first 5 of us to look at the puzzle got 11 or 12 songs quickly, and the other ones slowly fell to a combination of lyrics searches and getting other people to listen to the words. By the time we got that one critical song, the only unsolved one was the song originally represented by "same" which was changed to "crumbcake" (which nobody could understand). What ended up being used for that one? It was R.

(Reply to this) (Parent) (Thread)

We went back to "same" because that word at least had more musical background behind it. I tried almost every word in that song, and "same" was the only one that was spread out long enough that you could really hear it. Not Googleable, really, but I felt like a fan of '80s R&B might be able to get it from the snippet of music. "Crumbcake" would have worked on Googleable, but as you say, it was just too hard to understand.

Toonhead found "table" quickly with Tunatic, which was the suggestion we gave to teams later. Unfortunately, we didn't test any of the other songs on Tunatic (we were in a panicky rush at that point), and "table" was in fact the only one that worked. Sigh.

(Reply to this) (Parent) (Thread)

Yeah I guess I shoulda emphasized that Tunatic can't do much when the clip is a few mere seconds long.

(Reply to this) (Parent) (Thread)

Foggy was one of the (if not the only) solvers on Confession. Dayum.

This one was in our last batch of puzzles remaining to be solved in testing, but only because we didn't seem to have any comics/cartoon geeks. It's got a great little a-ha that seems obvious in retrospect.

(Reply to this) (Thread)

Oh goodness no! Confession fell so quickly because it was a pretty big teamsolve when everyone decided it looked fun. With a few "generations" of comic/cartoon fans looking at it, we were able to spot everything pretty quickly. And for what it's worth, I was the one who asked, "Do all of these guys have something on their chest?"

(Reply to this) (Parent) (Thread)

Love it! :) I always like those group solves. We did that with the "7 Days" puzzle of a few years ago at the beginning of that Hunt and it was a fun start to the weekend.

(Reply to this) (Parent) (Thread)

Nobody solved "Son of the Realm...", but FYI, two of my teammates are now IMing me in Chaotian. They are scary.

(Reply to this) (Thread)

Speaking of which... has anyone heard anything about the Chaotian T-shirt since wrap-up?

(Reply to this) (Parent) (Thread)

What was said about it at wrapup?

-One of the teammates IMing oxeador in Chaotian.

(Reply to this) (Parent) (Thread)

Just that someone was planning to get T-shirts put up at a website (not Cafe Press but I didn't catch where) saying "Nea fi 'pazosto-sulda' ot, py nea galaksia lokula-pazosto-sulda at."

Presumably, they'll arrange for it to be printed on a pazosto sulda.

(Reply to this) (Parent) (Thread)

Presumably, they'll arrange for it to be printed on a pazosto sulda.

I believe that was the idea, from what I heard at wrap-up. :-)

(Reply to this) (Parent) (Thread)

I suddenly feel much better about my performance on this Hunt. Character Witness and Campsite were both puzzles on which I cracked the aha.

(Reply to this) (Thread)

Any idea when the puzzles will be posted online? Those of us who couldn't participate really want to see what MIT came up with this year.

(Reply to this) (Thread)

Puzzles will be posted online this weekend. I'll make an announcement in this comm when they're ready.

(Reply to this) (Parent) (Thread)

For the benefit of us trichromats, could you use more variation in colors? On this monitor, I can only see 7 distinct colors in that graph, which makes it hard to figure out which of the blues belongs to Manic Sages. (Actually, even looking at the RGB values I can't distinguish them, so being a tetrachromat wouldn't help.)

(Reply to this) (Thread)

The colors deliberately repeat. Just follow the order -- the left-right, top-down order in the key matches the top-bottom order in the graph, of the colors. So Manic Sages are the topmost blue line.

(Reply to this) (Parent) (Thread)