Log in

23 January 2008 @ 08:08 pm
Mystery Hunt, 2008 Edition  
I hadn't started my puzzle blog when I competed in my first Mystery Hunt in 2006, but my best memories of that event (and my biggest contributions) were probably doing Surely You're Hexing by myself when we were really stuck in Buenos Aires with only a couple unlocked puzzles including a runaround we couldn't do and an event puzzle that was hours and hours away, and, late in the event, identifying the gimmick and eventually identifying all five constellations in A Bunch Of Dots to help open up more puzzles in the final Paris round. One of those later puzzles which may have been opened then (all for one, and one for all) was my second favorite SPIES puzzle after 783658 and had the best Evil Midnight team solve of all puzzles in that whole hunt. I did contribute some to an ante meta or round meta here and there (I specifically got to Chocolate Moose long before anyone else), but other powerplayers on our team did a lot more with metas than I did. I definitely stayed up for that entire hunt although parts of Saturday afternoon were really unproductive for me, and I had an excellent time solving an excellent hunt with an excellent team.

So, with a construction year behind me, which was also a learning experience, 2008 came with a second chance for evil midnight to once again go boom baby boom. I brought my own celebratory mechanism this year (basically, a fist pound turning into an explosion - although I think I forced it more on other people working on puzzles with me than it being accepted by them) and a determination to again crack puzzles that others couldn't.

(This write up is long, and likely has lots of typos, some temporal inaccuracies, and other problems. I will not heavily edit this until more data is available from Palindrome to clear up things and restore my puzzle memory better.)

My Mystery Hunt started in a way on the red eye to Boston as I began the sleep deprivation experience well before the start of the actual event. To wake myself up before the Hunt, I spent an hour or so solving backlogged Nikoli.com puzzles. At least one arriving team member tried to stop me from spending my energy, but I really needed to both kill time and get sharp. As we waited in Lobby 7 for the start, we solved a quick puzzle 0 that someone else had spread around the room that got people to call a caltech admissions line. This was a fun little exercise but reminded me that checking data entry is always important as small errors are easy in the rush of a competition. We had a 666 area code for awhile which seemed appropriate given our hunt last year, but 626 makes a lot more sense when we got the right number.

Eventually, Trip took the stage and warned us of lots and lots of puzzles before Dr. Awkward came out to reveal the coin only to get murdered. Back to HQ to get started. This year, a documentary crew would be filming Evil Midnight during the Hunt which, in addition to an earlier project before the Mystery Hunt, would mark the second time that week I'd been under the watch of cameras for extended portions of the day. I must be a fascinating person who keeps equally fascinating company.

(Times from here on are approximate - data from palindrome on when we phoned things in would clear up a lot).

Friday 1 pm:

We started the Mystery Mystery Hunt with an open black book page, an open dossier page, and little concept of the structure. I'll comment later on how I feel starting with an open dossier was a very poor structural decision, but for now, just as I thought then, onto the puzzles. Taking a quick look at the starting puzzles, I first joined a group of others on One Odd Out. We caught a couple dangling letters that would eventually be correct - seeing "New York" and the end of "Television" for example, but did not have a confident reading of any line yet to make a significant break. I decided to make a spaced out version of the text which I then printed for our small group to solve. As we struggled with this, I took a quick look at the apparent stereogram/domino puzzle which we never got into. Finally, someone got a correct reading of the whole first line of One Odd Out and we were off. We got the television channels pretty well - I contributed E Entertainment Television for example as a better answer than the one we were using as it didn't space right. We had our first black book answer. As I searched for another puzzle to do, I got my first call of: "Thomas, we have a logic puzzle"

Friday 2-4 pm:
Monopoles, in dossier 1, had been unlocked. This puzzle contained 9 monopole magnets puzzles which would eventually, seemingly, need to be stacked. As sometime between 2006 and 2008 I became recognized as the best US logic puzzle solver, I now get called over to do these puzzles whenever they are spotted although I have many other talented teammates who work with me here. We had 9 puzzles, and Dan and Francis and I got started on them. The page I was working on looked rather hard, until Dan reminded me that they were 9x9, not 10x10, so 5's on the outsides were strong constraints. One of the others felt stuck on their page, but I looked at their project and saw they weren't using the monopole constraint at all. Anyway, this puzzle started out this way with some very slow progress, but we gradually got 1, then 3, then 5, then seemingly all 9 done. I probably did the big part on about 6 of them. As we colored in the magnets for the overlaying step, I picked out one of the most uniformly colored grids and then confirmed to myself that it could not pair with anything. It turned out we had an incorrect answer to that puzzle which was quickly confirmed when I checked all the constraints on it. Giving that out to someone else to fix, Evan Davis and I worked to align the rest. When the fixed ninth puzzle came back in, we could do the whole assembly. Some excel entry and letter shifting later and we had the final answer. I rather enjoyed this puzzle, with an appropriate level of difficulty (yes, I found it hard, but it was still doable unlike other things I'd look at this hunt), and the stacking of magnets puzzles was a somewhat inspired construction idea.

Friday 4-5 pm: During monopoles, when people caught onto a 24-themed puzzle, I jumped out a bit to identify people there. I've seen every episode of this increasingly embarrasingly plotted series, but I don't have the best memory for people and names and even those chosen for Facebook weren't the easiest to id. As I was looking for a new puzzle to tackle, I probably went back to facebook for a bit. I took a fresh look at let's ask the dead guy but it seemed it needed a lot more data processing than I wanted to do. The Paint By Number portion of Do You Know Where Your Needles Are? was discovered and while I wasn't really needed, again the "logic puzzle = Thomas" metric applied and I was called over to do a 20 second clean-up of a mostly solved PBN to give us another answer word.

Friday 5-7 pm: My training in science is in biology and chemistry. This made me look a bit into Memo From Agent C with some other people on our team. I really think the DNA puzzles are inelegant for what they are. Yes, codons can encode letters, but somehow the biological context seems missing most of the time I run into these. I always find myself wanting to BLAST things and that is rarely if ever useful, except maybe during the most recent Microsoft Puzzle Hunt I tested with the Burninators and friends when it identified the answer word. After going through a couple of the layers here, we had a read on a lot of three letter answer phrases but no sense of an answer method and a lot of paths we could take without anything seeming best. As others peaked back at special ops - wondering if it could be divided uniquely into dominos - I quickly showed that wasn't the case. Oh well, that puzzle would still be impossible to progress on.

Friday 7-11 pm: Hunt continued, and progress seemed slow. I found myself for a bit of time working with a bunch of others on treasure map. We'd misplaced some pieces on the top line when I first assembled the cut-out grid, which made the rebus harder to solve, but Foggy had backed out what the answer phrase must be so we made it fit by rearranging the top correctly. We then worked for awhile on sharp turns and while I made some observations that were useful, most of the progress was completed by the more crossword strong part of our team.

As Black Book Page 4 was opened, just before our dinner break, I heard another call of "Thomas". A kakuro-looking thing had unlocked. Zack and I sat down to stare at it. We quickly agreed it was a puzzle in another base, which was proven correct quickly, with hexadecimal seeming best and fitting with the flavortext. It also was obviously a product kakuro. While this puzzle required some matlab factoring assistance, I had a fun hour-long solve of a puzzle with Zack as we would get a possible multiplication, place some forced ones, pin some other cells as possible pairs, then speculate what needed to be true. I ran every placement through a "must be true" check with Zack and he agreed most of the time and caught my errors the rest of the time. It was a very effective partnership and we got to our answer word. Proving my devotion to never giving up on a puzzle to take a break even when other necessities call, I was solving parts of this while going to the other room for a pizza break. Sure, I'd lost the matlab assistance, but I can deal with that for a short period of time it seems.

About this time, we also hit one of the coolest "go outside the bounds of MIT" puzzles in the hunt, Nationwide Hunt. This involved calling up people you knew, or didn't know, to find a particular word written in a particular place on either a monument or an apartment occupant listing or several other places in 6 other cities: NYC, SF, Chicago, LA, Miami, DC. This often meant begging friends to go somewhere distant from where they were, but Sheffi got some key people (strangers) nearby the Miami and NYC sites to get the info we needed and with 3 words we cracked the puzzle. It was a real nice concept and has some of the aspects I like most about the hunt - getting outside of just the standard tasks and outside of MIT.

I believe I was then handed Cult of the Helios to look at, but at that hour of night, seeing lots of decimals was intimidating to me even though it ended up being a nice logic puzzle. Someone else would solve that one.

Saturday - into the wee morning:
Around this time, our team had begun to catch onto the groupings of black book people. We had the five letter first names "almost boxers" together, and the brady characters. The depth of answers was not promising on either one but we were starting to think about what to do with them. Someone mentioned the "SOLVE NOW" and "ASHITORI" parts, but only outloud, and after a description of ashitori as a japanese wrestling maneuver, I did not further think of why this seemed odd to me.

Three of us took a look at Common Thread and I quickly saw the alphabetical arrangement gimmick of this basic puzzle type. We got all three columns on the first page done and were trying to figure out what to do on the second page when I heard that Greg L. had found a Caltech hack aspect to the Propaganda puzzle and I had to go help out on that (being an alum and all). This was a real key puzzle to solve, at least meta wise, for that round, and after a half hour of finding answers, finding the first layer, and then the answer layer, that was done. Common Thread had been by then and I still don't know how but I'll read online later. Sean was a good ways into European Cruise and I did my work-horse assistant duty, filling out the answer words we could, and also using my fast eyes and shape identification to spot (at least) 2 errors he had made in transposing the shapes into letters for a word search program. I quickly suggested what I called "non-landlocked countries" since a word like "coastal" is outside my vocabulary after a lack of sleep but we figured out the grouping we needed and then got the final message in due course. It was a nice puzzle with good use of the rotating grid to change letters temporarily between the grids, even if we just solved it like four cryptograms without using straight letters until the end.

I noticed a poker puzzle open, but no progress, and spent an hour that would prove useless putting data into a spreadsheet. It was mindless, but I needed a simple task for a little bit of time. Others were busy identifying beers which was certainly not my specialty.

We had by this time finished a dossier meta, which did nothing for us release-wise, but had not yet made any progress with black book metas. Foggy noticed the street number to answer word for the Brady meta and I think many of us (Dan certainly and myself) thought of a sudoku metric for it. I managed to miswrite the obvious grid up the first time (CHECK YOUR DATA, ALWAYS!) but once we sorted out my transposition that put two identical letters in the same column, we got a part of the grid set in obvious sudoku fashion. We actually were missing the whole right column, but some Scanraid vs. Snyder battles/collaborations gave us potential answers for the missing puzzles and, more importantly, an obvious use of the two diagonals. We backed out the final answers and finished that meta, making the aforementioned beer puzzle and poker puzzle unnecessary.

While other teammates went out to stack some cards, I went back to looking at many of the puzzles we had open but had no progress on. I recall making no new progress. The meta finally done, we opened up another dossier round which contained a super long cryptic. While this isn't my specialty, I made 4-5 identifications and then moved onto other things. I'd return to this impressive puzzle much later, and multiple times, to help with the fill-in once all the words were there.

We had however at this point seemingly reached a real sticking point with our boxing meta group. There were 8 answer words to the meta and we only had 4 with the other four of them to puzzles we were REALLY stuck on (like Special Ops, and Sphere - which I looked at a lot to try to crack once we had the mountain data). Stuck in the "we won't likely get any of these remaining answers" state, Dan stressed the urgency of doing something with this meta. I was still confused about the whole japanese wrestling move thing, so I chose to write out the two column message. SOLVENOW ASHITORA I wrote on the board. Someone corrected me that it was an I at the end (I was going by memory, and I had never seen the word written). After a quick A/SHIT/OR/I misparsing, I got "AS HITORI!!!!!!". Another logic puzzle meta! Woo hoo! While this identification might be enough for a normal mortal to rest on their laurels, waiting for more answer words, I was determined to crack the entire meta with only what we had. I first got all the information we had up on the board, and even though Dan would question how I'd make some markings and so fast (I'd somewhat rudely push him back to the cryptic to leave the hitori to me), I got a start of a message *TA**SE*LESS (asterisks are unknown numbers of letters, but with more expected between the A and S). Looking at the columns gave some possibilities for how letters could cancel in other words and after trying SEAMLESS for a little while, SENSELESS seemed a much better fit taking an additional letter from SKELETON as seems likely. With another bomber giving the key insight word BAREBACK which could get us the start BEAT, we suddenly had gotten to BEAT ANTON SENSELESS. The meta solved, I continued to pound the hitori and got out all 4 answers we were missing, including the ottomans and cozening in the middle that should have been impossible. Needless to say, I'm particularly proud of the work on this meta and it seems others, including Mark Halpin who constructed this beautiful meta, shared that view. I really really liked how this meta worked, particulary how easy letters like the L between the E's in SKELETON being clearly a white square, and canceling an L lower, working much better than any hitori I've seen from another source. Dan and I went to apply some muscle, and thinking better of our team's robot building skills, laid a quick two minute smackdown that opened up another dossier for us to look at. We were early in the morning and understaffed, but the returning teammates who had slept would find fresh puzzles to try.

Saturday - morning:

While we'd been handed some black book pages to this point, we'd now opened two dossiers that had not been mass-unlocked. I should comment that it was very difficult to know during this hunt how we were doing. There was only one event, Friday afternoon, which gave us the sense it was hard for everyone, but from then on we were alone to struggle and believe we were behind. A frustrating thing happened which is we might make a meta assignment and then see it emailed to everyone minutes later. I was turning at this time to trying to make other assignments and grouping data in a spreadsheet to make this easier. Earlier, I'd found the NJ reststop grouping even though we did not have enough to do more with it. Foggy had found the taxonomy grouping. As the morning rolled on, by collating black book data, I noticed a letter bank quality to first names grouped with samantha. I asked some of the fresh, waking members (after I gave them a hunt update) to take a look. They caught onto the symbology and we'd found the auspicious symbols. It would only be later when I went back to google searching that I'd find the important Sanskrit phrase that indeed contained the letter bank for the first names and was critical for this meta. Unfortunately, as hinted above, ten minutes after we found the auspicious symbols, HQ told all teams that those names worked together. Crap. We must be really behind.

With the morning came the release of black book page 7 and, as puzzle were printing, a familiar "Thomas" would call me over to see Chimera. I did the easy nurikabe part and then made a quick check that LAMM BENTO NATIONALISE did not mean anything with the only food hits for lamm bento being web mispellings of a more logical route to it. I did a quick glance at trying a cryptogram fences with the many repeated letters as I'd seen Mark Halpin use before in a labor day extravaganza with the very repeated E,R,I,B in the puzzle but say too many possibilities. I think I'd quickly gone to help out with cryptic crossword filling in when Zack came to ask if I'd done the corral yet. Oh, a corral - not a fences. That's simple too. We had the answer minutes later.

Over the nighttime, when we were short staffed, we'd unlocked two HQ event puzzles I wanted people to do. The Tichu game was something I thought of learning and then figured others could get at better. Case Study, however, was a quick puzzle for someone to do and as we learned after our first failure what was going on, we brainstormed as a group how to crack it. Dan came real close his next time except for completely reshuffling an arranged answer word apparently. We had all the letters when he came back, but it took a third trip by Jenn to get the answer word. We got the Tichu done after Zack got back as well and progress was ongoing.

Functions was a puzzle I joined into after some others (Foggy and Dan I think) had made a lot of progress. Zack and I suggested a couple other functions that proved to work and we tried to power the final answer given the usage/nonusage constraints that seem to fit with the puzzle. I was not around when we got to an answer (I think we got to an answer), but I thought the puzzle - while challenging - had some nice character to it.

One other interesting moment came in the morning which brought a lot of energy into HQ. Apparently, a dodecahedron snowball was outside our building. It was a trick of the eye from our particular vantage point, but we thought it was super cool. Maybe some other team was thinking of a clever way to beat our Hunt from last year, but a dodecahedron snowball is a very evil midnight bombers type symbol.

Saturday: Into the afternoon

I also looked a bit at the Last of Sheila, and we got a lot of the cryptic clues before (after another long period of time) someone made an AHA into how those words meant something else. Again, there was one AHA too far to still go and I couldn't see how to get an answer.

A Set puzzle was somewhere in the mix, but with lots of seemingly complex math. I'd keep thinking I should be able to do it, then get intimidated, then try again, then repeat. This went on for the next 24 hours from time to time. Like many other puzzles, it did nothing to restore spirits when we never got any step towards breaking it.

I'd peaked a lot at where we were stuck in puzzles for awhile and was not getting us over any hurdles. Input/Output was a painful one for me, as we had the whole network filled out - but no answer mechanism. I tried desperately to get the word INPUT into the word OUTPUT by going through the grid but it didn't work. Looking at the data also gave nothing popping out. I did similar glances at other puzzles but nothing came out. At some point, Sheffi mentioned NPR for No Police Record and with Sheffi we got wait wait don't tell me to fit the flavor text way too well. A quick NPR search, with the correct targeted archives, gave us the names we needed to index and an answer. Another clean-up job for the Terminator (my new nickname, I guess, although it's come about since the Hunt ended).

Sometime around the afternoon, we finally had enough NJ turnpike names that the chemistry subtheme to that meta came out. For the rest of the day, I'd be explaining the periodic table and Hydrogen's dual roles - matched to Lombardi's dual rest stop roles - to people looking at the meta. I came up with a lot of messages I liked, even with only 4 answers, to work into the meta. We had HANDSFREE and ALIAS and OUTFIT and CIRCLE I believe, and on the left hand side I like a fit like HANDS - AR - GE - A - ??????, The Ar would match with Cs for Arcsecant or a word like that and Ge would match with Rb for Gerbils or something else. I loved the match that much more because Molly Pitcher, one of the names behind the reststop, was often called Sargeant Molly and police fit with the turnpike theme well. This of course never proved to be the correct message for this meta, but I spent a lot of time trying to find something here.

I jumped over to the other open meta, the buddhist symbols, where I'd found the letterbank of their sanskrit name, and tried again to figure out how to use our answer words (about 4 at that time). Not all of them letterbanked nicely at all. The lengths weren't uniform but ABOVEREPROACH stood out as rather long, as did LAMENTATIONS just one shorter. Someone commented on the unique first letters of the answers matching the letter bank framing the meta, and my brain quickly latched onto using those letters to index into the answer words to read from ashtamangala. Having ABOVEREPROACH already, which gives 5 of the 12 letters, we could get BANKRUPT NASH within minutes. I was proud to take this meta over the several humps with help, obviously, from my super-talented teammates.

The result of solving this meta was one of my favorite activities - to stack a deck so that one could win eight straight hands of blackjack with the deck able to be cut (not shuffled) anywhere by the dealer beforehand. Having cracked the meta, people felt I should be involved in the activity, but it was a group of several of us that talked through how we could do it. I was dying of hunger, so I took a break with Mike and Evan and went out to Mass Ave. to get some Thai food. We came back and our team had figured out how to arrange the deck and had bankrupted Nash once and for all. It was an hour break from puzzles, but good for the morale and we came back refreshed for another assault on puzzles. That meta seemed to only get us arrested, so we didn't get a whole new set of things to work on, but we had other stuff to do anyway.

I wanted something new or different to work on, and I saw Ecolocation being tackled by some others. They were working on the latin-ish id's but I was considering the library map and reading the flavortext and getting some critical parts of what we needed to do. As we mapped answers into the grid, and saw gobbledygook coming out in the message below, I suggested breaking through the walls in the letter selection and that proved the insight I brought to the puzzle but it was important. As others worked on that, I took a peak at Elmo's First Wife, saw a quick Jackson JR placement within a second or so, and went on with Dan to finish that puzzle in about 15 minutes. This hunt needed more of that. Much more of that. I spent the next hour or so with Francis going through Picture Puzzle, with him doing more of the crosswordy tasks and me looking at seemingly forced tilings to give additional letters. It was slow, but we made progress and after I left the puzzle, others were able to work it to its conclusion. Ecolocation eventually met its end as well.

At some point we put on a hilarious scavenger hunt solution. I was deep into filling out cryptic crossword to infinity grids so I likely tuned out a lot of the clever contributions. One of my favorites though, when we had no Aztec Warriors, was Greg and Graham (I believe), putting on a newspaper hat and making a newspaper sword and presenting themselves "As Tech Warriors".

As Saturday turned into Sunday, my tank was starting to run dry. I had spent a long time looking at the taxonomy meta, drawn to the starting ACGT letters in the kingdom through species names. I remembered Joe DeVincentis looking into words spelled in codons read base to base from a mathpuzzle.com post long ago, and knowing his place on palindrome considered if those letters matched to anything with a similar encoding, forwards/backwards/reverse strand, .... We also noticed our answer protective tarrif had the start of protista, a kingdom of life, but that seemed another false lead. We needed more answers. We also needed more answers on the chemistry meta. We had a chip to cash for an answer word, but I was hesitant to use it and perhaps because of that it did not get spent.

At some point in the early hours of Sunday, some magic combination of Foggy's 10x10 board game brainstorming, the spoken name "chutes and ladders", and my storing and constant rechecking of first and last letters in the answer words for that dossier brought together another meta answer as the PLAYCHUTES ANDLADDERS grid was organized within a minute and the meta done a dozen or so minutes later.. It was fun to work out this on the black board as we used different colors of chalk to mark all the chutes and ladders and then "played" the set moves, backsolving HENRYFONDA as well to rescue all those trapped on the horribly broken Underpants Gnomes puzzle.

Metacracking on the blackboards (Foggy, Christy, Zack, myself)- Photo courtesy of Lorinne

Sunday: wee hours -
As head banging on the wall continued, Francis clued the anagramming to get into the first layer of Safari Park which again seemed somehow like a puzzle I should do. I took some real down time though to do little work and just watch solving happen. I made basically no contributions as Department Store was being solved but I was there for the process. I watched At the Beach as well and then did data-mining for the Give Me an M puzzle, which was probably my last real contribution before I decided I desperately needed to crash. Ski Slopes seemed to be falling at the same time my consciousness was. About 6 am on sunday I headed to the quiet room and lay on the floor with my pillow. About 2 hours later, the sun had reawoken some part of me, but not my brain. I know I was very cold and very confused. This lasted for the next couple hours. My few coherent thoughts focused on how I was going to be unable to fit visiting cambridge friends and football into Sunday when the Hunt was nowhere near ending. I was really depressed, likely as a result of this I think. I'd planned hunt until midnight and then Sunday to meet with people before a 6:45 pm flight and suddenly none of that seemed possible. The hunt was kicking our butts and we did not know we were actually doing quite well, all things considering.

After this miserable of set of "waking" hours in which I could not concentrate on anything, could not make progress on anything, could do nothing puzzle related as my head was seriously spinning, I think I finally woke up. I certainly found myself in new clothes with some washed hair and brushed teeth. It is still a surreal experience for me to try to remember (and I have a fair bit of difficulty here) much about 8-10 am Sunday morning except I remember feeling very unsettled. It was a bizarre start to my 28th year of life.

The waking up really happened as I got down to Six Easy Pieces, a nice puzzle left rather untouched. My mind jumped to Feynman for some reason, and I found a lot of ways to get up/down/top/bottom/strange/charm out of chess moves in that grid. UP was uniquely a bishop capture, but from many places. However, nothing was just knights moves from one letter to others, or from one point. Eventually, Zack and Francis would look at this again with me and once Zack or Francis said - well, look at how we can get PAWN - the puzzle fell. I was rejuvenated.

I needed to sort out how to see through the end of hunt, and how to visit friends, and decided to push my american flight back a day and incur the extra expense. I called my friends to fill them in on my delayed sunday schedule due to the interminable hunt. This helped me calm down and focus as well. I got back to solving puzzles, although I was going between desks so often it felt like musical chairs. Over the next several hours, I'd work on Safari Park, or At Sea (where my logic diagram mapping with Zack seemed to get us a good deal into it before we ran out of crazy words to place), and throughout Saturday and Sunday I'd been fretting with Steve on the fill steps of the big cryptic crossword. We'd gotten to many wrong answers, but not yet finished it.

Sunday morning brought, in addition to a second wind, the opening of hinting. We were unsure how much hinting to take so we put in one request at a time and did not add another until we needed it. An early hint got us through Talk to Me which gave us another answer word in the taxonomy meta. The phylum answer seemed to give us a new point of attack. The old PROT.... theory from PROTECTIVE TARIFF and PROTISTA failed when there was no phylum starting APPO (the closest being apicomplexa). A new theory quickly emerged from me - all the letters but the S in protista were in protective tariff. All the letters but the second O in ACTINOPODA were in the answer we had for the phylum. Maybe this held up for the other answers. We had 5 of the 7, but needed to jump down to the order for our next answer, so we searched taxonomy databases for awhile to try to get farther.

While this proved fruitless, we didn't spend too much time here as we gained clarification from HQ that asking for hints was a good idea (and we should do more of it) and that dossiers were more important. Dan and I were field generals of a sort in terms of finding out what puzzles could use a clue to quickly get us to an answer (not puzzles that needed a lot of clues and a lot of work to get to an answer) and for a couple hours we got some more work towards answers but really the biggest hints were things like: You're cryptic dots and dashes and just completely wrong. This made us redo some broken work and get a right answer.

I was jumping between a lot of puzzles I felt I could help on, and sort of speaking people through progress on many things. I wanted to do safari park and at sea, but was doing other things as well. The Dungeon was one I looked at for a bit, caught some of the character of it (each move used only once), but none of us saw the final Aha. The description at write-up was cool.

It got to a critical time though when I felt, if there was anything to do, it was crack some more metas. I'd been fascinated by the will, and wanted board space devoted to info in the will so we could start to solve that round. (Others had begun the work on Dossier 6 which was our other outstanding meta at the time). I put up the 6x10 grid, and marked the numbers in the columns, and marked the x's where the words favorite had been circled in our copy. There were 12 circled favorites already, matching the total expected number, so I just had those x's down on the board. Sean and Foggy were looking at this too, and I think Sean asked several times if there were any particular kind of logic puzzles that could fit onto this kind of grid. We decided pentomino packing made a lot of sense. However, we were quickly defeated on a couple of the tracks we tried (packing all pieces so they contained exactly two numbers for example. Thinking of how I'd construct a pentomino meta, I said that you'd have answer words that told you tiles used in particular places, so that as you got answers, you had more pieces that were forced. This would mean a talented solver might not need a lot of answers to solve the whole grid, but you definitely needed some answers to get started. This idea was fine, and felt fine, except our placement of IOCANE POWDER could not work in the grid as the pentomino for "favorite items" in the upper left needed two of the first row squares, and the I pentomino couldn't go straight down either. I said - well, this assignment must be right though as it feels so right for how you'd construct the meta. I took all our non-iocane powder answers and fit them into the grid. The bottom tiled nicely, with two ways to put an F with one seeming to fail early, but both gave some appropriate looking indexed letters for ???????KED ???????ILL which could have WILL at the end for example. The top though could not fit as we had it, even with the I tile ignored for the moment. I kept repeating the N with the layout we had. This was very frustrating, as it seemed we had to have the right idea. The Round 6 people were making a lot of progress and Dan kept offering for us to possibly take a free answer here (we still had that option) but I'd say no - we don't need it until we figure out what's going on here. The critical moment came almost simultaneously as Sean was rechecking the will info and someone got the answer VOLKSWAGEN. This could not fit with the ones we had. We had not circled Favorite car in the will at all! Favorite items was not meant to be used. The busted grid was instantly fixed. If we just had the proper data, that meta would have been done an hour or more earlier. Minutes later, the meta fell with the obviously framework we knew but just with bad piece data. We were slightly behind the Round 6 dossier which had also just fallen, but I imagine Palindrome got real excited when we called both in within minutes of each other.

Sunday: 5 pm

Soon, Trip ... ahem ... Alan Caster was in our headquarters explaining how suspects needed the means, the motive, and the opportunity to do the crime. We had a set of puzzles and a letter to sort out, as well as a fill-in. I captained the fill-in effort, getting an answer word list from the back of the room by reading the posters, then commanding the resorting of the excel list as I placed the long and short and medium length answers into the grid. While others helped, I can do fill-ins (even with missing words) real well and this came together fast. We had ~2 of the missing people clearly there and we figured out that the positions in the grid matched to dossier positions and could identify the third reasonably well as a possibility from there. The other group was not certain on their three - but together we both knew Alafontaine was in both sets. We convinced Trip ... ahem ... Alan of this, and the endgame was on.

My favorite moment of the hunt, maybe, came as Dan was organizing the group to leave for the endgame. I was copying everything into the black book (we'd been sloppy and not had the 9th group written in there, nor the answers to the others), so I went to do this while administration was going on. I asked for, and got Zack to be a sort of pit crew for me. He ran to my suitcase and found my sneakers, and I slipped these on as I finished copying data. Even puzzle solvers need pit crews!

Soon, we were off towards building 26. We first went into an incorrect lecture room, and found lots of nothing there, until it was reminded that wasn't exactly the right place on the map. Getting to that place, we searched a lower hall and got fascinated by a key in an electrical box (didn't someone take Jan's key?) only to realize it was in all of them. Someone remembered its supposed to be upstairs near a door marked keep so we ran up many flights to the appropriate hall. While we saw the door with keep, it was locked, so instead we tore apart the hall. There was a library filing cabinet and it was fun watching the team open and close every single drawer in this massive cabinet in a 5 person group. Other cabinets in the hall got similarly rifled through. Eventually, we just broke into the door with keep on it as we were meant to to discover two hidden palindromers in a room with a hidden copy of the original runaround. Unfortunately, the runaround went from A to B, we were at point B, and as Dr. Awkward was at point A, we had to do the whole runaround in reverse to get back to the coin. This sounds as hard as it was, as clues had to be applied somehow before you got to a place. Language often gave you many choices - you knew to get to the floor you were at, to confirm you were at the right place, but the clue before just said to take the elevator. This meant it could be any other floor the elevator could have come from. We got caught up first with finding so many different floors with lasers that it seemed it was the last floor (the 4th?) Zack and I searched, well after exploring what someone was sure was the basement, to continue. It should be mentioned, we needed to manage to stay together during this search process but similarly scout out different directions to find promising possibilities. We eventually got to the strata center which, even in a forwards runaround, is difficult to navigate. Breaking through some classrooms backwards, we got to the first level reasonably, then eventually figured out where from outside we had come from, and then made a wrong extension all the way into building 16 to a cheering Chaotic team (even though we were struggling with the runaround, they were very excited to smell a team on the endgame possibly ending the hunt). After seemingly forever, we got back on the right track and continued through the runaround. It went on like this. In a hunt with too many 3+ hour puzzles, this runaround was just another. Fittingly, at the same moment Jenn was finding the coin, someone was questioning in a very concerned way if we were in the right place from a reading of the first runaround clue. Hunt was dead. 20+ hours late, we went boom baby boom.

The bombers who survived to the end - photo courtesy of someone using Lorinne's camera

Link to flicker set slideshow from the Palindrome organizers

Some quick perspective:

The hunt had many good puzzles, and some fun metas (I love logic puzzle metas, and the pentomino/hitori/sudoku metas were well done and helped me step up more than other kinds of metas would normally allow). However, it was crippled by having too many hard puzzles which forced not only round release to all teams, but eventually complete circumvention of the black book meta structure.

I really liked the black book meta structure, and liked solving how to group things. I did not like being told by HQ how to group things ten minutes after I'd worked hard to figure such a grouping out. I also think a critical structural flaw, which led to a lot of confusion, was starting teams with both a black book page and a dossier. We solved the dossier 1 puzzles reasonably fast, and cracked that meta first overall. This gave us nothing at that time, except more words on that dossier page. We did not know you were meant to solve black book puzzles first (to open more black book pages) and then solve black book metas to open rounds. We'd learn this from observation about 20 hours later, but the structural confusion came from starting in both the first and second tier of the hunt. When I imagine a Mystery Mystery Hunt structure that begins with just black book pages and no dossier, the structure becomes cool and functional from the start, not just confusing. In this mythical world, of course, the puzzle difficulty has to allow teams to get through the black book metas on their own and not be round-released the dossiers, but it still seems an improvement I wish Palindrome had made.

Batting 2 for 2 in hunt winning, I hope we can continue a solid effort in construction this next year as well. We learned a lot in writing the Hell hunt, and we can take a lot of lessons from the challenges faced by Palindrome in this hunt. Look forward to something magical in 2009. As I blogged on the 1st of January, "I hope (and simultaneously fear) my winning streak continues" at the Hunt. Well, now the dream (and nightmare) has arrived. Another year of lots of puzzles, and lots of work.
devjoedevjoe on January 24th, 2008 05:35 pm (UTC)
Thanks for the detailed writeup, Thomas. I am glad you enjoyed the puzzles, and sorry that the structure was not better. Originally, Abigail was supposed to be a free answer that you would call in ("The answer to this puzzle is ASK ABIGAIL") but somehow (perhaps associated with the fact that you would already have the names and addresses from the back side of page 1 in the printed Little Black Book and the desire to also start you that way in the electronic one) this was changed to the puzzle already being solved. In hindsight, this was one of the decisions that led to the structure not being so great.

Another issue was that the Hunt you solved was a stripped-down version of our original plan. After it became apparent that the Hunt was too long, the parts most in need of work were sacrificed and the remaining parts duct-taped together. Without spoiling the unreleased puzzles (since we might figure out how to make them work for a future event), there was supposed to be one more normal dossier full of puzzles (which would have contained Instant Replay from the You! dossier) and 3 more events like the search warrant, which would have cast suspicion on a total of 11 suspects, leading to similar but expanded versions of the Means and Opportunity endgame puzzles. (Wouldn't it have felt better to have five puzzles in the first round of the Hunt Dr. Awkward was writing, instead of three?)
devjoedevjoe on January 24th, 2008 06:07 pm (UTC)
Another part of the problem was that, in my opinion, we did not do a good enough job of making sure the Little Black Book puzzles were easier, and therefore easy enough to get through to open the pages and let you solve those first-layer meta puzzles in a reasonable time. At one point, when most of the puzzles were written and their proposed grouping in the LBB decided by the editors, we were asked to check that the order made sense and that there weren't logjams that were going to get people stuck in the LBB. My response was that it seemed like too many of the hard puzzles were in the LBB and I didn't see a way to order them that wouldn't get people stuck somewhere. This (unlike some of our beliefs about puzzle difficulty) turned out to be exactly right, and was one of the factors that led to our choice of an accelerated round-opening schedule.

Early on in the development of this Hunt structure, the Little Black Book was supposed to have been full of what one team member called "puzzlets" -- puzzles of normal non-Mystery-Hunt difficulty which you were expected to solve quickly as an appetizer to the real puzzles in the Dossiers. There were supposed to be 60 of those and 90 Dossier puzzles under that plan. This was probably a better idea than what we actually ended up doing, but it was poisoned by the proposer's comparison of the puzzlets to the puzzles on Tanga.com, which while short (just one medium-sized image) and sometimes easy, were sometimes horribly broken and unsolvable. Tahnan wrote about one of those recently.
Jammermeeko713 on January 27th, 2008 06:55 am (UTC)
two things:
1) i've been working on the sudoku's from: http://www.akiloyunlari.com/tst08/tst2008eleme%20.pdf

there is a slight language barrier, but i can't seem to #5 for the life of me. do the directions say that you have to superimpose the rules for all three puzzle types, and then solve? because if so, i'm having a fairly difficult time with this. but the cube variation is pretty awesome.

2) i don't know if you'd be interested, but if you would deign to play a game that me and my friends are organizing, we would love for you to play:
motrismotris on January 27th, 2008 07:21 am (UTC)
Re: two things:
1) Having tested this puzzle for Ferhat before the Turkish competition, I can confirm that you are meant to solve a cubic sudoku such that the digits 1-8 appear in each "normal" cubic sudoku region, as on the left, and each jigsaw region, as on the far right, while simultaneously satisfying the killer constraints in the middle. It is a very impressive, if difficult, construction. You can find more "sane"-looking cubic sudoku if you search around but it is one of my favorite, if simple, variants.

2) I'd heard rumors of this, but not seen the actual site. I'm very excited and will definitely compete. I've already sent out some feelers to friends at Stanford.
zundevilzundevil on January 28th, 2008 08:34 am (UTC)
Re: two things:
The Stanford Game looks awesome! Suffice to say, nothing of the sort goes on down here, and would probably be of interest to about a half-dozen people in town anyhow.

BTW -- solving the digital sum thing...I get two solutions. My middle-middle digit and the one above it can be swapped. Have I gone wrong?

I wouldn't want to fly Oceanic Airlines anyhow -- I hear their track record is spotty.