Mission accomplished – DASH 8 described

DASH 8 deck of cardsThis site makes no apology for writing a considerable quantity about DASH with just as considerable delight; it’s always one of the highlights of the year. If you couldn’t attend this year, here’s what you missed… and perhaps, just perhaps, it might make you interested in taking part in a future year. If you played DASH elsewhere and were keen to know how London interpreted this year’s puzzles, you can find out here as well.

Fair warning: now that DASH has finished, we’re into potential spoiler territory. Every previous DASH has had its puzzles posted online reasonably soon afterwards. If you didn’t play DASH, it would still be a lot of fun to get a group of your friends together and try the puzzles for yourself once they’re made available. This post is going to be fairly generic, avoiding the Aha! moments for each puzzle, but the comments may be more specific. Nevertheless, if you want to avoid spoilers altogether, it may be wise to skip this post and it may be very wise to skip the comments. However, if you played and want to relive the experience, if you played elsewhere and want to compare stories or if you know you’ll never play this year’s puzzles and just want to find out what you missed, then to get to the detail you can click on the mission dossier that is the “Continue Reading” button below.

This year’s title was From Puzzles With Love and the theme was, loosely interpreted, espionage: secret agency, spying – the double-zero motif points to a well-known film franchise with the edges filed off and a very welcome gender subversion. The theme lent itself to an easy-to-follow but intriguing, enjoyable plot; there wasn’t a vast degree of narrative. Last year’s specific theme was a real audience-splitter, generating both strongly positive and (a few) strongly negative reactions; this year’s rather more accessible (if generic) story didn’t risk turning people off in the same way.

As has been the case in recent years, the structure was an unscored icebreaker puzzle that required co-operation and information sharing between teams, a string of intermediate puzzles and a metapuzzle to conclude. This year there were seven intermediate puzzles as in DASH 5, though one fewer than in DASHes 6 and 7. The previous statistics post will be better able to inform you than any verbal description as to how voluminous this year’s puzzle load was. The overall time limit for the hunt was a flat ten hours. We used nine and a half of those ten hours, but we were a very relaxed team, enjoying our locations to the full – four and a quarter hours solving, five and a quarter hours not solving. (This wasn’t the longest dawdle time, even among finishing teams, though it was one of the longer ones; for instance, the San Francisco winners, Black Fedora Group, took 2:20 solving and 5:36 dawdling.) Certainly it felt like an unhurried day and the day was all the better for us as a result, though preferences may vary.

The icebreaker puzzle, GHOST in the Machine, saw teams provided with one of three sets of information and required teams to swap information with other teams so that every team might work from a complete set. The information in question concerned twelve spies from film and TV; each spy was associated with a code letter and a fictive birthdate, and each set of information permitted you associate four spies with their code letters and a different four spies with their fictive birthdates. Swapping information, you could complete the line-ups for the dirty dozen. Each team also had a grid of alphanumeric characters that was suspiciously sized 12×26; sorting by the day of the month and cross-referencing birth month with code letter permitted you to pick out the puzzle answer. Fun, and more entertaining as a gentle warm-up puzzle (after the data-sharing!) than the counterparts from previous years.

In London, this puzzle was hosted in a beautiful Thames-side bar just the other side of the Vauxhall Bridge from the real-life MI6 building. One very minor criticism was that the London hunt started perhaps 30-40 minutes later than advertised – and, marginally worse, there was no attempt to keep teams informed as to when the start time really would be. This is only a tiny speck of a criticism when the London Game Control team did so many things right. As well as the scheduled puzzles from the hunt, there was a completely optional sidetrack of entertainments between puzzles to earn Her Majesty’s Appreciation Tokens; we could fill the time of the delay trying to solve (well, short-cut!) the first of those, which involved identifying missing words from plaques that would be encountered on the day’s route… or could be searched for online!

The scored puzzles, in the order in which they were encountered on the Experienced track in London, were as follows:

  • Stacked Deck, played at the same bar, was an algorithm-following puzzle to determine the play and results of baccarat hands played to a given ruleset using a deck of cards in a predetermined order. Baccarat is a game with few decisions, but only one combination of decisions would permit all ten hands of baccarat to be won. The extraction was based on the hands’ values but calculated using some of the Baccarat rules; here, it was one of those occasions when a team-mate extracts the answer in a way that doesn’t appear at first to make sense, proves to be correct but you can then back-derive the logic behind the extraction from the ruleset and confirm that it makes perfect sense. Lovely puzzle; Baccarat is very highly thematic and yet just obscure enough to work properly. Very nice production values here too with a pretty (and unique!) deck of cards as swag. I took an early guess that the transparent film on the card box would be used to show up invisible ink later, which was not to be.
  • The first activity, on the walk to the next puzzle, involved picking up a message from a secret agent and promising to deliver it to his identical twin. Some very non-identical non-twins prompted us for this message later and our decision not to release it to them earned us an appreciation token at the end of the day because we proved that we could keep a secret. Very cute!
  • The second puzzle, courtesy of our own dear (and last year’s London lead GC!) Iain, was Miss’D Connections. Solve pairs of clues and trace out the solutions in grids of hexagonal cells. Observe that these solutions overlap and transfer these overlapping cells to another grid, from which the puzzle’s answer can be traced out. There was some corking British history trivia here, which was an amusing response to the frequent DASH subject matter focus being reasonably designed to suit the vast majority of locations in the US. However, we too had to look the facts up and we were far from alone. Interestingly, despite the mechanic in this puzzle bearing some similarities to the UK TV quiz Hive Minds, teams in London were not unusually quick at solving it (someone else can perform an appropriate significance test here) though London’s Magpie team did get a global sixth place on this puzzle.
  • After a tube ride, Gear Up was the first long puzzle of the day with a par time of 55 minutes. This was a multi-stage word and picture puzzle: solve word clues, identify that the clues’ answers were included within “gears” of letters on the picture, identify that the remaining letters in each gear had an additional commonality, identify that the remaining letters after stripping out this additional commonality left a pattern that made sense in the context of the picture, pick out picture elements based on this pattern and sort them in the order given by the commonality. Lots of steps, but a very clear, logical, straightforward, delightful solution path.
  • This puzzle was solved in a park near the One New Change shopping development. Our team decided against taking lunch at any of its restaurants and instead advanced to the pub which held the next puzzle, eating there. Teams solved upstairs at Ye Olde Watlinge after climbing ye olde irregulare staircase. An entertaining additional challenge while sober, surely a safety hazard after a couple of pints. (The steak sandwich on the menu was well-received and the sides of fries, salad and onion rings were all notably good, but I’d steer you away from the cheese and chutney sandwich I had.)
  • The puzzle at the pub was Paper Trail, with a par time of 75 minutes. This was another multi-stage epic. The meat of the puzzle was folding sheets into pretty origami cranes, with a training video having been released days before the hunt began, teaching the technique. (Did the most competitive teams have all their members practice how to do this? For instance, London’s Misremembered Apple team deserve plaudits for their brilliant global fourth-place finish on this puzzle.) Once folded, messages could be read from the assembled cranes and interpreted according to techniques stated within the puzzle. Happily – or luckily? – we folded and dealt with our cranes in the correct order, possibly being the order in which the sheets were handed out. The second last extraction in the puzzle used a cipher generally considered a little too advanced and time-consuming to be seen frequently in DASH, though one that can be sped up through Puzzle Sidekick at least and possibly other apps as well. The construction of this puzzle was a beautiful piece of craft.
  • Off to the achingly cool The London Stone pub (which had an appreciation token earning opportunity hidden within the men’s toilets… and, though we didn’t check, hopefully the ladies’ as well) for GHOST Protocol, this year’s take on quickie mini-puzzles. There were a set of cryptic clues (deliberately – and wisely – kept very accessible), a rebus puzzle and a picture-identification letter puzzle. The final extraction was very neat. We were thrilled to sail through this in the fastest time in London for our only local first-placed finish. We had been enjoying every puzzle without exception; we enjoyed this more for our facility at it than anything else.
  • Thence to Spurious Information, with a bundle of phone numbers and images. The phone numbers could be translated into country names (though the converter I tried at first was a bit of a dud, slowing us down), the country names into telephone calling codes to extract sequences of letters and these letter sequences could be cross-referenced against the images, with the stray letters giving the final extraction. Extremely neat construction with enjoyable results; congratulations to London’s Magpie team for a global top-five solve here.
  • Off to the final location, rooms within the giant Wetherspoons The Crosse Keys pub, for the last regular puzzle and the meta. SPUR of the Moment, designed by north-west England’s own Puzzle The Cheetah, was a beautiful physical puzzle and a delightful, intricate piece of design. Cut out printed, coloured nets and affix them around 27 provided wooden cubes. Orient them in a 3x3x3 capacity box so that their markings can be decrypted to provide the puzzle’s answer. The markings took the forms of block capital letters, so that an H could be a sideways I, a N could be a sideways Z, and it’s impossible to distinguish between capital E, M or W. The colouring was crucial to determine what went where; a member of our team with some colour vision issues had no problem here, though there are very many different types of colour perception issues. We struggled with this one like we did with no other, but we really enjoyed it; we beat every other par value except one by at least ten minutes, but we crawled past the finish line to claim just two very hard-earned bonus points here.
  • Puddings to celebrate, then to the metapuzzle, Win your SPURs. Assemble a map of sixteen marked tiles into a 4×4 grid so that the internal intersections’ marked corners create eight squares. (The end result looks a bit like a crazy multi-coloured QR code, or perhaps a superposition of QR codes in primary colours, but isn’t one. I tried it!) These eight squares can be matched to eight coloured vehicle tiles in a thematic fashion, using the answers to the eight previous puzzles. The tiles’ colours provide a decryption sequence for the markings on the tiles to generate the final answer to win the hunt! London’s Misremembered Apple team finished with a flourish, gaining global eighth-placed solves on both of the last two puzzles to finish an extremely impressive global eleventh overall.

A pleasantly full day, and yet not overloaded; so much to admire and for which to be grateful, and so many who deserve all our thanks. It would be possible to go overboard by thanking the puzzle designers, global and local co-ordinators, testers, volunteers and people behind the scenes in every single post, but not to do so here would be unduly mean. Thank you; it was a sensational day!

The London experience was rather more than the sum of its puzzles, though. There were other optional activities along the way – if you could find them! – to win additional Appreciation Tokens. A test of marksmanship with a Nerf gun offered a chance to earn a token; another station asked you to crack a Mastermind code (four pegs to guess, five possible colours) within a reasonably exacting five-guess limit for another token. Apparently there was also another physical puzzle elsewhere along the line, possibly a little like the de Bono L game. The production values through the day were very good; while Puzzled Pint sets the bar ever higher, the deck of cards and assembled cube were excellent souvenirs. Last year’s non-puzzle activities felt a little… awry; this year, the activities in London were excellent palate-cleansers and their optional nature (though the prize for the team earning the most Appreciation Tokens was really pretty nice!) felt like a really wise, player-friendly decision. Excellent practice.

The moan department only really has one criticism for a piece of practice that we loved last year and missed this year: in DASH 7, some puzzle locations gave two identical copies of each puzzle to each team, which didn’t seem to happen this year. (Or perhaps there were limited copies and only teams larger than our foursome received two copies?) In any case, the more copies of a puzzle are distributed, the easier it is to parallelise puzzle solving; please budget the event so that multiple copies can be distributed as standard.

There were other omissions from previous years, but not every innovation has to stand the test of time. The Junior track was a lovely idea, but one which created a lot of extra work for evidently very little demand. Credit for, and confirmation of, partial solutions is solver-friendly practice, though not one which would have been called on frequently in this hunt. (I hope it was a bigger factor in the Normal track, though.) Not distributing token food prizes at the end of the hunt – a bowl of chips per team in DASH 5, a plastic lemon of sweets per team in DASH 6 – is definitely not a loss.

ClueKeeper seemed to work really well this year, and seemed to be most conspicuous by absence of its discussion, the ideal referee that doesn’t draw attention to itself; I get the impression that three years of improved practice has been enough to iron the few remaining issues out – or, alternatively, post-Hunt (not Post Hunt!) discussion might bring the remaining issues to light. Many thanks to Rich Bragg and team for this, in what is presumably one of the biggest stress-tests that the servers get all year. Using ClueKeeper to be able to produce global results with such speed as happens here is a massive, massive advantage.

The ClueKeeper did drain our batteries relatively quickly on iOS but that may be inherently difficult to avoid. (We brought a spare battery for this eventuality, and I don’t think our Android-using team member found this nearly such an issue.) The rate of distribution of free hints was finely and successfully judged; our middling-to-pretty-good (barely top 40%) team stayed ahead of the knowledge dispensed by the free hints on every puzzle except the metapuzzle, which helped us feel good about our progress.

The London locations were easily fine and often notably good. We were fortunate with the weather once again; showers were threatened, but at the slow pace at which we proceeded, only one outdoor puzzle had a few spots of rain, if that. Apparently London is lucky to be able to have so many indoor puzzles at all. The Brit-picking was generally excellent with relatively few cultural references that required investigation; Internet access cannot be guaranteed as wifi access is spotty and pubs – especially basement bars – often inadvertently block radio signals. The Game Control team were tremendous, being friendly throughout a long day; many thanks to you all for giving your time up for us players.

The par times were notably generous compared to previous years. My views are pretty much unchanged from those of last year. A principle often stated on the much-loved, much-missed Snoutcast was “Everybody likes solving puzzles, nobody likes not solving puzzles”; my corollary to this runs “Everybody likes solving puzzles, everybody likes solving puzzles and earning bonus points from doing so even more”. It’s a supposition, but it’s one I stand by. Certainly these par times have much in common with the Generous Average Solving Time principle explicitly stated in the past; it would be possible to measure generosity of the par time for a particular puzzle as a function of how many teams beat the quoted value and/or by how much. Without having done so, the top-level view points to these par times being relatively generous.

Is it possible that they might be too generous? There is an additional use to par times in setting people’s expectations. A set of par times that were four hours per puzzle across the board would do well at getting people bonus points (and would punish teams who chose to pay to take hints relatively hard, which might or might not be intentional) but would do poorly at giving people an indication of whether the puzzle they were about to face could be expected to be relatively long or relatively short. A set of par times that varied from, say, two hours and twenty minutes to two hours and fifty-five minutes would seem to have all the desirable properties but it would be hard to take them seriously.

What else did this hunt have to admire? Its players inspired more great team names; I’m a big fan of Octopuzzly. Yet perhaps the hunt saved its best for last with the message that came up once the hunt was finished. It seems that there is a global co-ordinator for DASH 9 already! Excellent news; Chris Willmore, this year’s San Francisco co-ordinator, you have a lot to live up to! The message also suggests that the web site might have a hidden puzzle. Perhaps my note thatthe logo on the T-shirts (and the other tote bag) hides analogues to morse and braille at the very least” did sell the logo short after all!


  1. Congratulations for staying ahead of the free hint release. I feel like we benefited from it twice during the hunt.

    You asked about paper crane preparation… Most of us had practiced before, but we had one member who proved particularly proficient. So we parrallelised the easy steps between the rest of us, saving the fiddly bits for the expert… We enjoyed most of that puzzle, although thought it was a little pointless including a more complex cipher if the intended solution (according to the hints) was just to write your input into a web form and read off the answer! Otherwise, a great little puzzle.

    The psychology of score is interesting. I tend to think about it as loosing points for each minute spent. Consequently I feel less pressure on low par times, as we have much less to lose if we completely fail to spot something… Having said that, feeling like we’d solved a puzzle quickly and not beating pat would be frustrating – so there are limits.

    • We had the same experience with premature free hints, which is really the only suggestion for improvement of any sort I would raise. On some puzzles we were given the solution to steps we hadn’t reached, despite being ahead of ‘par’ on every one, and on puzzles with practical elements they were very optimistic about how fast teams would complete tasks. For example, we got the hint with the transcribed information from the cranes just a few seconds (literally) after we finished folding the things, which was actually quite annoying having struggled to make them all accurately. And on the very first puzzle, we were given the full extraction key information before we had got anywhere near to that part.

      Perhaps the hints to successive parts of puzzles should only start their release timer when you enter a partial answer to a previous part. In that way, you wouldn’t be spoiled for bits you haven’t yet got to. The puzzles and Cluekeeper prompts could be designed to encourage you to answer these part answers in order to release the mid-way hints.

      Given that the only real point of the day is to have fun solving puzzles, it does to an extent spoil it to be given the solutions when you don’t want them. I realise you don’t have to actually open the free hints, of course, but I think it’s hard not also be at least a little competitive and not want to penalise yourself – plus until you read them, you can’t tell if they are spoilers or actually just extra instructions (like the one clarifying the orientation of a digit on a playing card).

      As an aside, I actually have no idea what the penalty for taking a hint was, since there was no explanation that I saw of how this would work this year. No doubt this is online somewhere, but in previous years we’ve been given explanations of how the scoring would work. I don’t think this is a problem since it’s just for fun, but it might also be good to include information on arrival, or mail out about these things in advance – and I had no idea there was ‘training’ online, until it was mentioned several times in clues. Perhaps there should be a central set of emails that go to teams signed up to DASH, with standard advance material?

      But really, I thought all the puzzles were great, and the UK event was really well run throughout (once it eventually started, anyway). It even had good (not uber-embarrassing) activities along the way – even though we couldn’t find some of them! So well done and thank you to all concerned!

  2. This was tremendous fun this year and felt like a lot of criticisms from last year had been considered. Enjoyed the bonus token aspect, although I think we missed most of the ones Cluekeeper didn’t point out which may suggest agents in the field might have been a bit *too* inconspicuous (really interested to find out what we may have missed). Great stuff!

  3. I’m curious — would most British people know offhand the trivia referenced in Miss’d Connections? (Not that looking it up is hard.)

  4. Thanks for the ClueKeeper shout out! We are very happy to hear that it seems to have gone well for most people. We have, as you suspected, been trying to squash most of the issues that have been raised in the past, and it seems to have paid off. Our numbers show that right around 2000 different people used the app at some point during the hunt, and we had a grand total of only 3 crashes that all recovered immediately, so “congrats” to those lucky few! 🙂 Also of interest is that this is the first year I had enough confidence in the system to actually play in the event, and had an absolute blast in doing so!

    Interestingly, the feature we implemented for this year’s DASH that we were most excited about was the ability to continue to work on unsolved puzzles and submit answers even if you timed out on the event. This seemed to be a relatively major complaint in the past couple years, particularly with the UK crowd, and we were hoping to showcase it this year, but it seems to not have been noticed much because most people just finished everything in time!

    Regarding the par times, that is a topic of particularly interest to me. We actually do recommend that par times be set especially generously, as that allows teams to have their finish order almost entirely dictated by their overall solve time. If you set par times too low, then you can end up with some clues that simply don’t differentiate teams at all, or even worse, with the difference between two team’s scores being determined more based on which clues were insufficiently calibrated than on their overall performance. It certainly does raise the question of the naming, and thus perhaps “par” isn’t the right connotation. Really these times are meant to be upper reasonable bounds that cap the time a team might possibly want to spend on a clue, and with that in mind, I do think the DASH 8 GC got this right.

    Finally, I do understand the frustration about hints coming too soon, and I agree that having some hints that are entirely gated (or essentially gated, by using a longer release time) on partials might be a great idea for some clues. The current ClueKeeper system is already capable of this, actually, but in GC’s defense, this is also the sort of thing that it a lot easier to identify in retrospect than it might seem. I will take from this the recommendation that for multi-step clues in the future, to recommend longer release times (or no release time) for hints that might spoil something, especially in cases where you can provide clear instructions about submitting the partial that will get you past that hint when you’re ready.

    • Rich, thanks again for letting DASH use ClueKeeper! I hadn’t noticed until a week or so ago that you had added the AR feature, but I look forward to trying that out.

      Using ClueKeeper on a very old iOS 7.1 iPhone 4 in read-only mode when another device was the master for our team, sometimes pressing “sync now” during the hunt was very slow or even would not appear to respond at all. This was towards the end of the UK hunt, so pretty much peak time for server demand considering that the East Coast hunts would be a couple of puzzles in and even the West Coast hunts would have just about started. DASH must be an amazing stress-test for your servers, possibly one of the biggest ones you face!

      I love having the amazing, detailed scoreboard pages, but I wonder whether (especially during times of heavy load!) it might be better to return a rather shorter, faster-to-load, less computationally-and-markup intensive page by default and having the full data available to download as (perhaps) a .csv file instead?

      • Indeed, I just heard from someone else as well today that the “Sync now” was especially slow, which is disheartening and I will look further into it. The system is designed to scale well (it is built on Google App Engine), but sounds like there may have been some choke points that I wasn’t aware of. The unfortunate part is that it is pretty difficult to simulate a DASH-sized event to reproduce such issues, as you’re right, it is our biggest event of the year. 🙂

        I agree that a faster-to-load scoreboard would also be nice to have and I’ll put that on the list, too. In fact, we may even try to get a scoreboard you can view in the app itself by then, as that too has been on the wish list for a while. 🙂

        Also, in the interest of full disclosure, while I’m very grateful and happy to let DASH use ClueKeeper (as you put it), we did actually change our policy this year for community, not-for-profit events from completely free to “pay what it is worth to you” in an effort to keep our company going while still providing a cost-conscious opportunity for labor-of-love style events. Hopefully, once ClueKeeper becomes the international sensation we all expect it to be :), we can again afford to donate the full cost of our services to events like DASH. So really, in my opinion, this time more thanks are owed from us to DASH GC and all the paying players than the other way around!


Leave a Comment.