On June 9-10, the gf and I went to sunny CA for PiratesBATH. Although she'd played in past Puzzle Hunts, this was the gf's first Game. She didn't play with Briny Deep, however, but with The Bonny Wenches, a new team comprised of various lady friends of Briny Deep. I'll not comment on the Wenches' experience, since I wasn't in their van, except to say that at least two of them, including the gf, had enough fun that they're now talking about playing in the next Shinteki event in August.

This Game distinguished itself from others in three main ways. First, most of the main puzzles were contributed by the teams themselves, each of whom was invited to submit one for inclusion. The advantage for doing so, aside from having an opportunity to impress fellow teams, was that when you encountered your own puzzle you'd get to skip ahead to the next clue immediately. Second, clues were provided not via live phone support from GC or a PDA, but an ultra-low-tech scratch-off and envelope system wherein teams purchased pre-canned hints for points. Third, there was no overnight leg; instead, teams roughed it at a campground (we were told in advance to bring tents and sleeping bags).

Things have been so crazy busy since returning from the Game that it's taken me this long to find time to write about it. So please forgive me for resorting to bullet lists.

What I liked

  • The majority of clues were fun, well-constructed, and highly thematic. When each team only needs to create one puzzle, they're able to focus all of their piratey ideas into one concentrated burst of freebooting goodness. We had puzzles themed around messages in a bottle, walking the plank, skulls and daggers, lovelorn pirates, sea chanties, sea battles, pieces of eight, treasure maps, and more.
  • Many of the locations, especially the coastal spots on day two, were spectacular-- and quite appropriate for a pirate-themed event. The weather was perfect, allowing us to enjoy stunning clifftop vistas and gorgeous sandy coves. If you're going to have a Game where most locations are just clue drops, this is the way to do it.
  • Our site at the campground was just a few paces away from the edge of a cliff overlooking a beach, so the gf and I were able to fall asleep to the sound of waves crashing to the shore-- one of my favorite things in the world.
  • The mini-puzzles. In addition to the main puzzles, we got a total of 36 mini-puzzles throughout the event. These minis were designed to be solved in the van between clue stops, in our tents at the campground, or whenever we wanted to work on them. Some of them were so trivial they almost solved themselves, others involved a little more thought. For our team they provided a welcome drive-time diversion, and I hope other Games pick up on this concept as a new area worth developing further.

    Things I Didn't Like

  • The hint system. In the past, the idea of purchasing hints has come up regarding the Microsoft Puzzle Hunt, and I've always been passionately against them. This event underscored why. With pre-canned hints of increasing cost for increasing information, it's really frustrating to purchase a hint that tells you something you already know. Often you'll get all the way through the puzzle and be stuck on the final step, but there's no way to get a hint for just that step-- you have to pay for everything leading up to it. Shinteki ameliorates that problem by letting you enter partial answers to demonstrate progress. A scratch-off system can't do that.
  • The skull economy. Teams were given a purse of skulls, each worth a point, and were free to barter them with other teams. Some teams scratched off hints and sold them for a profit. Some teams banded together when stuck and used skulls to spread the cost of hints evenly among them. The skull economy worked perfectly well, and gave teams license to collaborate when normally that kind of thing is (informally) frowned upon. But I found it to be a distraction, and that kind of negotiation wasn't why I was there. Rather ironic, really, since the trading game in Mooncursers added a similar element, but this is the first time I've played in a Game with such a system. As with so many things, it comes down to managing expectations properly. I think I'd have been fine with it if I'd known about it in advance. I'd have come to the Game with a different mindset and gotten myself psyched up to make a profit and embrace the economy (or, perhaps, to ignore it completely). There's a tendency to keep secret as many details about the event as possible until it begins. While that's certainly desirable for clues and locations, I'm not sure I agree when it comes to structural elements. Experimentation is good, but tell us in advance so we can be ready for it.
  • One broken clue and one horribly, disastrously located clue. The broken clue wasn't a big deal-- we were told before beginning that it had a problem, and so we were able to adjust accordingly. The other clue cost us about an hour and was completely avoidable (see below).
  • Timing. The Game got off to a late start and as a result teams arrived at the campground fairly late, leaving many with no time to both get dinner and participate in the activities GC had planned at the campground. Dinner wasn't as well-executed as it could have been. GC kindly provided food in the form of burgers, hot dogs, chips, and potato salad, but the burgers and dogs were uncooked and the grills were cool by the time people arrived-- so hungry players had to load 'em with charcoal and wait for the coals to heat. Cheers for providing dinner, jeers for not having it ready to go when players arrived.
  • Many of the sites on day one had also been used three weeks earlier in No More Secrets. This was a very unfortunate coincidence, completely out of GC's control. We've played in every Game for the past three years, and none had used those locations. Then, WHAM!-- back-to-back Games overlap. It was an odd sense of deja vu.
  • Starting day two with a completely pointless three-legged race on the beach. Note to future GCs (including the Shinteki folks): when you're about to start a long, steamy day of being in a cramped van with a bunch of other people, the last thing you want to do is get hot and sweaty right out of the gate. When that exertion is gratuitous, that's just adding insult to injury. I have nothing against three-legged-races per se, but activities need to fit the event and venue. Some kind of beachfront treasure-hunting activity would have been perfect. But a sack race? This was PiratesBATH, not a company picnic.

    The bumps, however, were minor and easily overshadowed by the positives. Great locations, many great clues, and terrific people all around. We had a fantastic time. Many thanks to Captain Bloodbath and crew for all their effort in staging the event!

    And now, the clue-by-clue rundown. Apologies in advance if I say horrible, mean things about your baby. Kudos to EVERY team who created a puzzle for this event, even the ones I hated. I appreciate the time, effort, and creativity that went into them. But I nevertheless offer my honest opinion, because I believe honest feedback is the only kind that's of any value. If my words are too blunt, I hope you'll forgive me.

  • Single Pirate Seeks Same: The event began with a fun scramble through a playground to find 26 pirate portraits and associated words, each of which paired with another based on similar meanings of their words. Once you had the pairs, you had to notice that each member of a pair was wearing their eyepatch on a different eye, telling you how to arrange them so you could read their shirt buttons as Braille. We did pretty well on this clue-- we knew immediately that we had to pair the pirates up, but there were so many other potential data points-- hat/hatless, weapon/weaponless, shirt/shirtless, male/female-- that it took a while to zero in on the words. I think it was a poor design choice to have an equal number of male and female pirates, but not have all the pairs be mixed gender-- when we noticed it was an even split, it seemed a clear signal to put them in male/female pairs. Still, we left the site only a few minutes behind the leaders and way ahead of most other teams, and the puzzle had plenty of internal cluing to lead teams to the pairing and Braille steps. A solid kick-off to the event.

    Turtle Island: Ugh. Six photographs of pirates burying treasure, and a big map of an island with lots of spots circled and labelled with letters. This puzzle had two problems. Fundamentally, it wasn't especially fun to solve. Even when we knew exactly what to do and what the answer was, some of the pieces still didn't seem to jibe-- associating each photo with where it seemed to be on the map was inexact at best. But the bigger sin was made by GC when a) they told us to bring a shovel, and b) they positioned this puzzle at a large, water-encircled park. Those things together made us believe that the map correlated to the real location and we'd need to dig something up somewhere. I mean, come on-- six photos of pirates with shovels!. Digging! So two of us traipsed around the entire island looking for a landmark from the photos. An hour later, we found the rest of the teams in the parking lot solving from their vans. It was absolutely, 100% foreseeable that some teams would think the park and map went together. All GC had to do was distribute this puzzle almost anywhere else-- somewhere not park-like, not encircled by water, and otherwise not resembling the hilly, grassy island depicted on the map. This fiasco sucked all the momentum we'd gathered from rocking the first clue, put us in a foul mood, and made us not trust GC for a long time.

    iPatch: A set of words with their letters arranged alphabetically except for the final letter, which was out of sequence, all presented on the screen of an iMac. The last letters, read in sequence, told us to apply an eye patch to the words. The eyepatch/iPatch pun was immediately obvious, but we were stuck for far longer than we should have been because all the ways we tried to apply the iPatch weren't working. We tried inserting an I. We tried inserting an I and anagramming. We tried deleting an I. We tried deleting an I and anagramming. What failed to occur to us was replacing an existing letter with an I. Immensely aggravating at the time, since we felt we had solved all but the last step of the puzzle, so we knew hints weren't going to be useful, but entirely our fault-- the right approach was just sitting in our blind spot. Many thanks to lowkey for breaking the conceptual logjam for us and XX-Rated.

    Lego Battleship: In retrospect this puzzle, which involved reconstructing naval signal flags on a Battleship-like grid using narrative accounts of sea battles between four warring fleets (each with their own Lego color), was a bit overwrought. The puzzle was straightforward and mostly a matter of just following somewhat obfuscated directions. But it was well-suited to parallelization by four solvers, and... Legos!

    Pieces of Eight: The Bonny Wenches' puzzle. They spent literally hundreds of woman-hours on this puzzle over the course of multiple months, much of which went to actual physical production. The lunatics decided to fabricate it themselves, which entailed cutting out 216 circles (9 x 24 teams) from foam board, further cutting each of those circles into three pieces, and gluing content to the top of each piece (which also had to be cut out...). Not to mention putting a complete set of the resulting pieces into a lovely sachet. In contrast, Briny Deep spent less than ten man-hours on our puzzle, leaving the production to Captain Bloodbath's galley slaves. But back to Pieces of Eight. Each properly-assembled circle spelled out an 8-letter word, but of course the words were tricky and were split among the pieces such that none of the first letters were actually first on a piece-- so finding them was a fun puzzle. Ultimately we were aided by a production glitch which effectively split the pieces into two sets, one for six of the words and one for the remaining three. Noticing that glitch enabled us to focus on the smaller set and isolate the correct words quickly, which in turn helped us gain a foothold in the other set. Once assembled and sorted alphabetically, highlighted letters spelled out SHIFT VIII. Caesar-shifting the words eight places revealed the answer reading down one column of letters. Placing the puzzle at an outdoor statue of Caesar at the Rosicrucian Egyptian Museum in San Jose was a nice touch, although one whose significance went unnoticed until after the fact. We felt like we did very well on this puzzle, gaining back some time, momentum, and good spirits.

    Letter from Home: The first of two tile-matching puzzles in the event. The creators of this one were kind enough to cut up the tiles for us (are you listening, Golden Booty?). Numbers on some edges mapped to elements, the abbreviations of which also happened to be same as the abbreviations for U.S. states on other edges. Assembling yielded a morse message of directions and values, which told us how to navigate through the letters of a fixed-font message and generate the answer. The pieces were a bit disjointed-- why combine Morse with states and elements?-- but it flowed pretty naturally. I would have liked the start point on the message to be clearer. Our suspicion of where to start was correct, but we were more certain that we'd wind up ending at the X so we opted to start there instead and work backwards.

    Knots: We were given dozens of tiny pictures illustrating steps in the creation of certain kinds of knots, each associated with a letter. We had to follow instructions to tie knots, labelling each step along the way with the correct picture/letter to generate a message. I didn't much care for this one, because there wasn't much puzzle solving involved until the final step. It was mainly an exercise in following directions and not letting tiny slips of paper blow away in the afternoon breeze. A much more entertaining version of the same puzzle would have had us making knots of each other's bodies instead, Twister-style. Future GCs: feel free to steal that.

    Daggers: The second tile-matching puzzle. This time, we had to cut up the tiles ourselves (boooooo!). The first step was similar to my tile puzzle from Mooncurser's, but with words instead of pictures and with flimsy paper tiles we had to tape together instead of sturdy foam tiles that were easy to manipulate. Did I mention we had to cut them out ourselves? Each tile showed two halves of a dagger, with the blade pointing in from one edge and the hilt pointing out from another. When property matched up, the tiles formed a skull shape (very nice). Each hilt had one to three dots on it, and apparently there was a message to be read by indexing 1, 2, or 3 letters into the word for that dagger. We, however, immediately thought of ternary, so we skipped that intermediate message. We used the right ternary decoding (subtract 1 from the dots) and somehow (thanks, Excel!) found the message reading backwards-- which is what the intermediate message would have told us to do. Since we didn't know about the indexed message the final step felt clunky to us, but now that we know how it really worked it was rather nice-- and we feel even smarter. A win-win all around, then.

    Boardwalk: This puzzle, located at the Santa Cruz Boardwalk, involved matching Photoshopped images of signs in the area to the real signs to extract a message. This was the broken puzzle, which wasn't a big deal. Especially because along with the puzzle came a bonus challenge, to win any prize at the boardwalk and modify it to make it more piratey. So what does a Jersey boy do to win a prize at the boardwalk? Skee-Ball, baby! I was pleased to see I hadn't lost my touch. I was psyched to be on the boardwalk, and only wish we'd had more time to spend there-- definitely one of the most fun locations I've been to in a Game.

    Crates: The puzzle itself was very simple: tip over crates of various heights to create a path from start to finish. Too simple, really-- it only took a couple of minutes to solve. Lowkey took it to the next level by constructing a life-size version of it for us to solve on the beach. A+ for production values-- I just wish it had been a bit more challenging.

    Chantily Clad: Day two began with the aforementioned sack race, followed by a delightful clue involving a CD full of sea chanties accompanied by a deck of image cards. The title of each chanty could be formed by phonetically combining three of the cards. Arrange the cards for each chanty in order created a 3x15(?) grid. The back of each card depicted a pirate, but inspection revealed that the pirates weren't all the same. Some had patches, some had tattoos, some had hats, some had hooks, and so forth. Isolating each trait and looking at only those cards which had that trait revealed 3x5 pixel letters forming the final message. This was totally fun, thematic, solvable while driving, and very satisfying. Bravo!

    Walk the Plank: Our puzzle. 24 dowels and a wooden plank with 12 rows. Each dowel had a word and 0-2 stripes on it. Each row of the plank had two blank lines (one flush left, the other flush right), 0-2 nails driven in between the blanks, and a hole for a dowel on each side. The word THE was written down the center of the plank, and the bottom two blanks were filled in with the words WALK and PLANK. This was a word association puzzle. Each dowel was half of a compound phrase. Each compound phrase matched another, such that the non-dowel words formed a 3-word phrase of the form VERB THE NOUN. For instance, the dowels SKI and FIN lead to SKI JUMP and SHARK FIN, or ski JUMP THE SHARK fin. So the SKI dowel got inserted on the left side of the plank, and the FIN dowel on the right side in the same row. Solvers were instructed to fill in their answers in alphabetical order, and when they had them all each row could be read as ternary using the bands on the dowels and the nails. We revised the puzzle multiple times to remove as many ambiguous pairs as possible, but the capacity for people to make up phrases and feel strongly about them is boundless. The ones that tripped the most people up seemed to be SODA POP THE QUESTION MARK and STAR BURST THE BUBBLE GUM. People really wanted to make POP THE BUBBLE instead, but BURST THE BUBBLE is the stronger idiom (and the completely flawed approach of Googling the two phrases bears me out, 73,800 to 29,100). The puzzle was originally created with constraints that were ultimately removed, so I'll take the blame for not returning to the puzzle after the constraints went away to remove the weakest pairs (DIFFERENCE ENGINE) and eliminate all ambiguity. Still, multiple test groups solved the puzzle as-is. I liked this puzzle, especially for this kind of event, because it's a great team activity. Even if you're driving, you can still participate and shout out phrases. It's hard to create Game clues that can get all team members involved.

    Message in a Bottle: The production work on this was the real star-- a real FUZE drink bottle had been precision etched with custom content, hidden beneath the obviously-detachable label. Inside the bottle was a stick with a rubber tip. The label had to be put inside the bottle and positioned with the stick so that its contents lined up with etchings on the bottle. We completely overlooked a key element of the bottle that bound two pieces of content together, and so we were stuck on the final step for entirely too long. But what really bothered me about the puzzle was that one of its two messages was completely unnecessary, effectively just telling you to read the other message. I suppose in a way it's no different from the Daggers puzzle in that regard, except in this case we didn't skip directly to the second message. But the second message was obvious to us, we just chose to decode the other one first-- which turned out to be a waste of time. A nice idea with great production, though.

    S un ken Ship: Two easy crosswords fastened together with a vertical barrier between them, obviously evoking Battleship. One side of the barrier showed a deaf pirate, the other a blind one. Each puzzle contained coordinates hidden inside the completed grid. On the blind pirate's side, the coordinates were phonetic (BEFORE = B4); on the deaf pirate's side the coordinates were spelled out (HEIGHT = H 8). The coordinates on each side pinpointed locations on the other representing ships 1, 2, 3, and 4 spaces long (as hinted by the title of the puzzle). When read in size order, those letters spelled the final clue. We completely rocked this puzzle, which our morale sorely needed after our flubbing of the bottle puzzle. I thought it all worked together beautifully.

    Origami: A surreal origami puzzle, presented in an enormous origami container. I am not very good at construction puzzles so I mentally checked out on this one and focused on mini-puzzles instead, but it all seemed to work nicely as intended. Again, though, most of the puzzle was just following directions-- the actual puzzle part was pretty small.

    + Marks the Spot: A set of popsicle sticks with velcro pieces and numbers on them, which had to be assembled such that the value shown on one stick was formed by summing the values shown on all sticks touching it. Just to make things harder, the sticks were double-sided. Harder, but certainly not more fun. The correct approach to this puzzle seemed completely arbitrary to us. In fact, even after scratching off major hints for this one, we STILL had trouble making it work out properly. Maybe we were just grumpy at this point, but by the end of this puzzle the bile had risen in our throats. Suffice it to say that this is the only puzzle in Briny Deep's history to be cast out of the van in disgust and literally peed on by two pirates.

    Little Things: Cards with three attributes-- color, time period, and subject matter. The instructions strongly suggested that the right approach was to create sets of four cards such that no attribute's value was repeated within the set. But we couldn't make it work. We tried, and tried, and tried again. We thought we might have gotten our time periods for some images wrong, so we adjusted and tried again. And again. Instead of thinking, "Maybe this is the wrong approach," we were certain we were doing the right thing incorrectly. The instructions could absolutely be read to confirm it. But no. That wasn't the idea at all. Instead, you just had to deal out the cards in a certain way and read the corresponding letters from the grids on each card. I can't really judge the puzzle fairly because the instructions led us so convincingly astray that I can't imagine how our interpretation hadn't been spotted in playtesting.

    Follow the Directions: A meta-puzzle involving all 36 mini-puzzles and the 36 wooden tiles we'd also gotten throughout the event. The tiles, each with a letter or number, assembled into a treasure map. But following the path on the map just produced a "It's not THAT easy" message. Here we were sunk by our own cleverness. On day one, we'd noticed that the answers to the mini puzzles formed pairs-- HEART and SOUL, SALT and VINEGAR, FIRE and ICE, and so forth. Each tile corresponded to a mini-puzzle, and each mini-puzzle had a compass rose on the back. When the path didn't make sense, we were sure the pairings came into play. So we spent forever trying things like: if the path took us south off the SALT tile, read the southern letter from the compass on the VINEGAR tile instead. We tried many variations on this theme, but nothing worked. What we missed was that each answer had some combination of the letters N, S, E, and W. So when the path took us to the VINEGAR tile, we needed to read the northern and eastern letters off the VINEGAR compass rose. This could have been a great meta if GC hadn't decided to be too clever for their own good. The pairs, you see, were part of the hidden meta-meta, which we didn't know existed (in fact, no teams did until the end). So when we hit a meta-puzzle that looked like the finale, nothing on the planet was going to move us away from utilizing the obviously-intentional pairings of the mini-puzzle answers. The meta-meta effectively ruined the mini-meta.

    Meta-Meta: The meta-meta itself was actually quite elegant. There were 36 minis and 18 main puzzles. Each pair of mini-answers could be anagrammed into one of the main puzzle answers plus two extra letters. Over the course of the event we also received eight doubloons, each of which depicted an icon. These, too, were paired-- CUT and DRIED, OPEN and CLOSED, LAND and SEA, and YES and NO. If you take the leftover letters from the anagrams and remove all the letters in the doubloon pairs, you were left with the final password to Captain Bloodbath's treasure chest: BARNACLE. I solved this on the flight home with much of the data missing-- I didn't remember all the main puzzle answers and didn't know many of the minis, but once I figured out what was going on I was able to backsolve all of my missing information. It was a fun meta, but only one team got to solve it. Let's think about that for a minute. The entire structure of the event-- the answers teams were assigned to build puzzles around, the very existance of 36 mini-puzzles, let alone the answers to those puzzles-- was designed for the express purpose of making a meta-puzzle, and only ONE TEAM got the opportunity to discover that puzzle during the event. One team out of 24. We would have LOVED to have worked on this puzzle. We'd have marveled at its elegance, at the pieces that had been waved under our noses. But the event hadn't been structured in such a way as to allow us to do so. That's a crushing disappointment for everyone involved. I feel bad for the people on GC who were obviously excited enough about the concept to craft an entire event around it. When you make something with so many precision-crafted parts, you want everyone to wind it up and watch it go. For only one team to have that opportunity must have been a let-down.

    In Puzzle Hunt, we've learned not to backload our most impressive stuff. As tempting as it is to build the event around an intricately-devised meta structure, the cold reality is that only a small fraction of teams will ever get far enough to see it. And while every aesthetic, puzzle-designer sense in me burns to create a whiz-bang finale that brings all the pieces together, the sad truth is that all that effort is better spent elsewhere, where more players will actually see and appreciate it. In the case of PirateBATH, considering all the planning that went into crafting the meta, I wish a little more planning had gone into making sure everyone got to it. I'm sure the organizers wish the same.


    *shrug* GC kept telling us to make the puzzle harder. So we did.

    "As tempting as it is to build the event around an intricately-devised meta structure, the cold reality is that only a small fraction of teams will ever get far enough to see it."

    Although it seems like a Game is where a meta would work the best, because you can guarantee that everyone gets all the pieces—when a team gets skipped ahead due to time, you give them the answer they missed. Seems like it's then just a matter of being ruthless with the timing, so that if there are (say) 18 main puzzles to solve in 18 hours to leave enough time for travel and solving the meta, then teams get skipped after 1 hour no matter what.

    Why didn't this event work that way? Were they just not skipping puzzles, or not doing so aggressively enough?

    Heya Peter,

    Alexandra from GC here. Interesting writeup, as always.

    Aha, YOU were the team who peed on the popsicle sticks clue - hee! John Owens told me some team did, but wouldn't tell me who :-)

    So, here's a peek behind the GC curtain.

    In this comment I'll use our internal terminology which referred to the "surface" meta and the "hidden" meta.

    Here's the thinking behind the two metas.

    We wanted to have a game that would stand on its own as a complete game on the surface, with clues, a meta-puzzle incorporating the answers to the clues, hints, scoring, and rankings.

    But - it's a pirate game, and what's a pirate game without a buried treasure at the end? And does a pirate TELL PEOPLE that he has buried his treasure? Er, NO. Usually, he leaves behind some cryptic map with X marking the spot. The first man to find the map, interpret it and dig up the treasure, wins it all for himself. So, we wanted to set up that very scenario.

    We discussed, very early on, whether it would be OK if *no* team found or solved the hidden meta, and we agreed that we would be absolutely fine with that.

    Of course, we were hoping that several teams would find it, as that would add to the excitement at the end party. We were actually hoping for a last-minute duel, team to team, which didn't quite materialize.

    BTW two teams - not one - solved it (Golden Booty and Get on a Raft with Taft).

    We decided that in keeping true to the pirate theme, the hidden treasure/meta would NEVER explicitly be referenced, that we would re-use elements from the surface game to clue it, and that every element necessary to make the hidden meta work would have a surface purpose.

    That's kind of how the skull economy and carnival came about - the whole purpose of that was to get those eight doubloons with the paired images on them into teams' hands.

    I thought the metapuzzle, with three different solutions embedded in the same data set, was gorgeous, and have to give Greg deBeer major props for thinking it up in the first place. It was pretty much the first element of the game that we designed - then we tweaked it for months. We scoured the internet and our own brains for related word pairs, slammed against a huge dictionary to find major words that incorporated those pairs, plus exactly two letters more, then had to find messages that would use 18 of those meta-words, with letters left over that would spell something meaningful in the hidden meta. And we wanted it to be a meta-meta, so it had to spell four major words, that themselves were made of 4 pairs of mini words, with 8 letters left over that could anagram to a piratey-sounding word (in this case, BARNACLE).

    Then when we decided every mini word had to have at least one of N,E,S and W in it, we rewrote it, then when the battleship/signal flags clue turned out to have too much of one color, we had to give them a different word, and tweak both the surface and hidden metas again...it went on forever!

    Then, Greg decided one message and a hidden meta aren't enough - we can cram ANOTHER message in there - and so the double surface messages came about - in addition to the meta meta that got you to BARNACLE.

    I must admit, when teams test-solved the surface meta, several of them did notice the (mini word + related mini word + 2 letters = major word) angle. Several of them (and several teams in the actual game) backsolved the solution to the surface meta - TELEPHONE - by noticing that there was no major clue answer for the mini answers HEEL and TOE. So they added two letters and anagrammed to TELEPHONE.

    Unfortunately, this did have the effect of directing teams away from the hidden meta, since they used the hidden gimmick to solve the surface meta.

    However...those eight doubloons were NOT used in the solution to the surface meta, and we were counting on some eagle-eyed teams to wonder why not. I mean, you got these doubloons, they had four matching pairs of words represented on them....cmon, doesn't that make you suspect something else is going on? And indeed, two teams did notice this.

    As for making sure that teams got all of the elements of the meta...that was entirely possible. First, a team could take a 100% hint and scratch off the answer and write it on its answer sheet, for any mini or major it hadn't solved. You don't get credit for that solve, but you don't lose points on it either. So, theoretically, ALL of the information necessary to solve the surface and hidden metas was available to teams - essentially for free - regardless of skipping. All, that is, except the images on the doubloons - which, we hoped, most teams would have - which is why we made the carnival ridiculously easy - or we hoped they'd trade for it, and besides, you didn't need all of them to solve the hidden meta. Six, maybe, but not all eight.

    We did fall down somewhat (translation: ran out of time to think through and execute a plan) in terms of getting the 36 mini answers and 18 major answers into teams' hands at the final party. We ended up lamely giving teams back their answer sheets and suggesting they scratch off the silver stickers to see what the answers would have been. While this seemed like a big fat hint that SOMETHING WAS GOING ON, most teams kinda went "uh, ok" (politely, rather than enthusiastically) and walked off with them, only to put them down and go get some food :-)

    We also counted on teams having some down time at the campground to think about all of this, and I think maybe we crammed too much content into the game to make this feasible.

    What we should have done is just prepare answer sheets and hand them out at the end. Again, kind of a spoiler hint, but it might have made it more likely that teams would actually think about it if they saw all the words in one place. (That's also why we gave five free mini answers, because we wanted teams to get as many of the mini answers as possible in order to be able to solve the surface/hidden metas).

    About the hints system - we initially were going to go with a Shinteki-style hints system on Palms (or the equivalent) with time-released free hints. That was absolutely our first preference. Then, we decided to program it ourselves...for laptops. Thus avoid renting or buying Palms and having to program using the Palm OS.

    Then we ran into the problem of - how do we prevent teams from tweaking their laptops' system clocks to get hints early? Or, if they do that, how do we detect it and correct for it so they don't get a hint early? We thought about this for a long time, and eventually decided it was just too complicated to do a time-release system on laptops, because we couldn't expect teams to keep the program up and running continuously. If they had, we would have been able to track any sudden changes in the system clock.

    So, we reluctantly went with paid hints. We did develop an electronic system for laptops, but felt we didn't have enough time to thoroughly playtest it, so we abandoned it in favor of the low-tech analog that we knew wouldn't break - the silver scratchoff stickers on paper. Which seemed more in keeping with a pirate game anyway.

    I don't think - and I don't think anyone is saying - that the silver stickers or the low-tech approach was the problem. Just that people hate paid hints, and they particularly hate paid hints that tell them something they already know!

    I totally agree. And I think, one of the most challenging aspects of producing a game in the modern era (= 21st century) is that Game Control really has to nail the hint/scoring system in order for the game to be a success. I think the Shinteki folks have mastered this, and the rest of us are lagging somewhat. Of course, with enough manpower, I think manned telephone hint lines are the best way to make sure you can give personalized hints that give teams useful information, but then you get into a couple of issues:

    - teams who call in early and often for hints can get an unfair advantage

    - you can only place clues in locations with cell phone coverage

    As for Stephen's question about why teams didn't get skipped after an hour? Well, they sorta did. We had approximately an hour budgeted for each clue's solve time.

    In the "more information than you want" department, I'll tell you how the skipping worked.

    Taking Saturday as an example: First, we budgeted 12 hours for the game (8:30 am to 8:30 pm, or 720 minutes). And no, it didn't start late, Peter! Then we drove the route - twice - kept track of driving times between clue sites, and subtracted the total drive time from the 720 minutes. The resulting time was total solving time. We then divided that by 10 clues to get the solve time per clue. Say, 59 minutes. To avoid teams getting skipped early if they hit a bad patch right out of the gate, meaning they'd never be in contention to win, we added a 50% cushion to that (29 minutes) to give you 88 minutes to solve clue 1 before you were skipped past clue 2 to clue 3. If you had a clue in the game, we gave you an additional 59 minutes = 147 minutes before you were skipped to clue 3. Then, we took back that 29 minute cushion proportionatey over the rest of the day. So, rather than having 59 minutes to solve clue X>1, you had 59 - (29/9) ~ 56 minutes. Plus, of course, whatever was left of your cushion, if any. We left it to teams to manage their cushion - we realized they could blow it all on the first clue and be playing catch-up later in the day, but it made the site staffing schedule easier if everybody - teams who wrote clues and teams who didn't - were on the same schedule after a team hit its own clue.

    The skipping instructions in the envelopes were based on the above algorithm. Plus, because clue 9 on Saturday, and clue 16 on Sunday, were not written by teams who were playing, we skewed the skipping schedule so that you had to have AT LEAST 75 minutes left to solve those clues, PLUS driving time, PLUS solve time for clues 10 (or 17 and 18) before we would route you to clues 9 and 16. Your team's bad luck, you were ahead of the curve and got sent to #9 - which was broken - and #16, which you peed on :-)

    But anyway, this comment is probably longer than your whole blog entry so I'll stop here!

    Alexandra: Sorry, I assumed the Game started late because of the starting location switch (the park didn't open until 30 minutes after the original start time).

    Briny Deep got to the end party sufficiently late that we never had time to look at the meta-meta. My understanding is that Golden Booty showed up and was denied access to the giant chest, which gave them the heads-up that there was something they were missing. Had we had a similar experience, we'd have "found" the meta-meta, too. But once an end party is underway, I think it's unrealistic to expect anyone to find a hidden meta-puzzle without any additional prompting from GC.

    My point about the meta-meta is that GC spent what sounds like hundreds of man-hours devising and revising all its components, ultimately for the benefit of just two teams. If I was the project manager for that event, I'd have said that GC either needed to find a way to get that payoff into all the teams' experiences or else devote that time in other places where it could have a bigger positive impact. We face the same issues in Puzzle Hunt, which is why over time the big payoffs have migrated to the front of the event instead of the end. For 22(?) teams, that clever, gorgeous meta was completely wasted effort. I appreciate it after the fact, but it wasn't part of my Game experience (and, in fact, hindered our solving of the surface meta). I guess I just have a mental disconnect with the idea of GC deciding a priori they were fine with the even worse outcome of NOBODY solving the meta-meta. If I'd created that thing, I'd have damn well wanted everyone to find it.

    Your X-marks-the-spot cryptic map analogy isn't accurate. Usually a pirate brags about his treasure and how cleverly he's hidden it. The cryptic treasure map is essentially a dare to other people to try to find it! Usually, a seeker knows he has the map. Teams didn't. Nobody's going to look for a treasure for which they have no map. If anything, we thought the map was the set of tiles we were collecting. Expecting teams to make the unclued leap that there was an implied meta beyond the surface meta was, perhaps, unrealistic.

    As for the doubloons, we figured they were part of something that had been cut, or perhaps were a step (like the indexed message in Daggers) that we just short-circuited.

    I think scratch-offs are a great low-tech solution to Game hints (likewise invisible ink, which might have been even more thematic)-- it's the price tag that people hate. If the hints were free, it would have been terrific-- just scratch until you get beyond your problem, then stop.

    Not to point fingers at my teammates, but I feel compelled to note that I did not participate in that particular exercise of freedom of expression-- I was too busy holding my head in my hands in the van, mortified. While it's a funny story, I also think it was totally uncool.

    "Then we ran into the problem of - how do we prevent teams from tweaking their laptops' system clocks to get hints early? Or, if they do that, how do we detect it and correct for it so they don't get a hint early?"

    I think a laptop app is a great idea. As a resource for any future GCs that might read this, let me point out SoftwarePassport. This is a lightweight DRM package for shareware developers, and one of the things it does is track and detect system clock rollbacks (to secure time-based free trials). Another feature is that it encrypts the executable, so that you wouldn't have to worry about someone scanning for the hint text using a hex editor. A free trial version of the package is available that would provide all the security a hint app would need.

    Stephen, that sounds cool! I understand that on the Shinteki Palms that they use in their games for hints/confirmation/skipping, they have the ability to detect system clock rollbacks - and if one is detected, their program just automatically advances the hint timer by the amount of the rollback, so teams can't game the system. You don't see anything that says "busted!" - you just never can get your hints early. Sounds like SoftwarePassport would accomplish the same thing.

    At some point, you've got to question how much time it's worth spending on trying to prevent cheating in an event where there's no prize for winning...

    Monthly Archives