Puzzle Hunt 123


The Microsoft Puzzle Hunt, which I've been working on for about 15 long months, is now over. This event broke me. It is probably the last puzzle event I will ever run with a volunteer committee.

We tried some big, risky things in this event, and I'm very happy about that. I'd rather fail spectacularly for trying something different than do something safe that doesn't push the envelope. At least one of our innovations-- timed puzzles that teams were encouraged to solve as a team as an in-conference-room event-- was a resounding success. This concept was born from my feeling that out-of-conference-room events represent a tremendous amount of overhead for something that a small percentage of players ever see. From a cost/benefit perspective, they're a horrible investment. I wanted to find a way to create special moments the entire team could partake in. Timed puzzles, specifically constructed to be conducive to group solves, were a great low-cost, high-impact solution, and they seem to have been universally adored.

Some experiments work, some don't. In retrospect, it's clear how different decisions would have made the event better. I take the blame for all the problems that didn't get corrected. There was no one leader-- the hunt was essentially run by committee. That doesn't excuse me from responsibility for poor design or execution. We had a chance during the event to correct the biggest problem-- players being blocked from accessing more puzzles-- and I pushed the wrong priorities. Instead of looking at the evidence that teams just didn't want to use our existing release valve of moving from the Competitive to the Recreational division, I stood by it under the belief that any change of course at 3 AM would represent a breach of trust to the Competitive teams that had moved beyond the blockage, and to the teams that had already switched to Recreational to get around it. I still believe that to be true, but breaching that trust and unblocking players may have been the lesser evil.

I feel deeply disappointed that, after 15 months of planning, the event we ran was not the event people wanted to play. I grossly misjudged what people wanted from Puzzle Hunt. Competition is deeply ingrained in the DNA of its players, and they accepted enormous amounts of frustration rather than give that up. Some people on the organizing committee thought that might happen, but I didn't believe it. I was wrong. I accept the blame. I deeply apologize to all the players whose fun was compromised as a result. I also feel terrible for all the puzzle authors whose work got less exposure because of it.

The event was created when two teams, each planning a Hunt, ran out of steam on their own and merged (the events merged; almost all of my original team simply bailed). That was reflected in many ways in the event, and usually not for the better. Elements conflicted with each other. Problems compounded each other. And mostly, the creators were just tired and ready to be done. It frustrated me to be the front man for an event that I didn't entirely believe in, and it depresses me to feel so defeated by the experience. I don't intend to put myself in that position again.


Ummm, Peter, from where I sit, you're being _incredibly_ hard on yourself over nothing. I had a lot of fun with the event. The harshest criticism I can lob at it is that I'll remember the travel brochure from Puzzle Hunt 8 before I remember any of the puzzles from this event, but that's hardly defeat. I expect mind-blowing stunts like that to be the exception, not the cost of entry.

The puzzles were very high quality as any event I've ever seen run, and as you point out, teams who had as their goal "see all the puzzles" had the opportunity in a stronger way than ever before. If their stubbornness blocked them from seeing that, it really _is_ there fault. Are you on the hook for their realized fun, or their potential fun? I'd argue that it's the latter.

It would be a tragedy if the world lost you as a puzzle constructor.

To clarify, when I say this is the event that broke me, it has nothing whatsoever to do with the outcome of the event itself. If it had run perfectly, with all teams finishing and thinking it was the best event ever and praising me as a veritable genius among men, it wouldn't change anything. Well, of course I'd be happier with that outcome. But I wrote a version of this blog entry before the hunt even began. The problem lies in the process, not the outcome.

As a member of a team (Wrong Ideas) that forgot to read the directions to Meta 2, the ultimate result of which was we unlocked the "You went to sleep" achievement without actually going to sleep, I agree that the structure had a problem.

I suspect that we were hit by that issue more severely than many teams. But you know what? We had a lot of fun, and saw a lot of great puzzles. We were stuck for a good while, and that did detract from the experience, but not to the point where we weren't all thrilled to have gotten to participate.

Thanks for putting what was clearly a huge amount effort into making this happen. I hope you're not permanently burnt out on all GC type activities, but instead can apply the lessons here to another game later.

(I got a http 500 response when I POSTed that comment, it looks like it went through though)

Dude. That was an amazing event. Of course you should try new things. And we're all adults; we knew what we were getting into when we signed up for the competitive division. I can't say that I had perfect insight into what all teams experienced, but from our point of view, the fact you guys wrote both a great set of puzzles AND a set of puzzles that was fair meant that any frustration was at a minimum. And no different than any other puzzle event!

Usually when we puzzle, we get what GC gives us in terms of hints. This time, we got what you gave us - and getting no hints is not any different than many other events I've done - AND we had an additional choice as well.

Again, it's GREAT you tried new things. I think what you tried was important to try and worked out fine for at least this team; and if it doesn't work for all teams, the next GC can try something different. For what it's worth, this was one of the **very best** puzzling events I've ever done, and you know that my resume in this area is long. As far as I'm concerned you guys nailed it, and I wouldn't say this if I didn't really mean it. You guys are rockstars, and ran a tremendous event.

This was a great event overall. The puzzles were very well written and well tested. The daily doubles (as you know) were an awesome innovation that I hope will be replicated in future events. Even the recreational/competitive split was, IMHO, a good thing with perhaps a little tweaking (but we had to start somewhere) and also some getting used to.

The only thing that went wrong was the one, disastrous decision to have the metas rounds run in serial rather than parallel. Obviously my team was less affected than others, so maybe my perspective is skewed. But I can't fault you for the decision(s) you made at 3 AM - it's hard to think straight at that time of night, there's fairness issues, and who's to say what the response would have been if you had decided differently? By that time you were already between a rock and a hard place. At least we (as a puzzle community) can all learn from the experience, right?

Honestly, Peter, are you talking about the same event that we played? Your post makes little sense to me--we had a great time, and everyone I spoke with had a great time as well. It was tons of fun, and I actually donated *more* money to the cause afterwards as a result of my very positive feelings afterwards.

Just get some rest. You'll feel much better!

I took time off work and traveled all the way across the country for the event, so my expectations were pretty high, and you certainly delivered on all fronts.

When you look back on this I hope you're able to focus on all the things that were great about PH 123. The great puzzles. The daily doubles. The pretty much flawless execution of the software and the infrastructure of the event. The virtually typo-free puzzles.

Successfully hosting the largest puzzle hunt in history is a huge accomplishment and you and all your team deserve the highest of praise.

I really hope that people truly appreciate the time and effort that you put into this and I hope that you continue to run events in the future as having you "retire" would be a big loss to the puzzling community.

For now take a long, well-deserved break.

I'm very happy that you guys had a great time! The experience of the top teams, however, was very different from that of teams lower down. So in one sense, no, I'm not talking about the same event that you played.

The thrust of my post was less about the way the event turned out than it was about the way it got created and the resulting unnecessary stresses on my (and other peoples') lives.

One of my Grey Goo teammates forwarded this link to the team, with the comment that it felt spot-on to him, and I have to agree. We were one of those teams that fell into the awkward place in the middle, and I've been refraining from saying too much about the event, because immediately afterward we were frustrated and demoralized. I think we're one of the teams you're talking about.

In some regards, it was a great hunt - we repeatedly found ourselves cheering at individual puzzles, and grinning like kids as a group of us pulled a solve out of a particularly elegant puzzle. (Mai Tai, Houses of the Stars and Walk of Fame all got this reaction - others too, but those three I clearly remember working on.) Solving the Daily Doubles was absolutely incredible for our team, with the pressure and intensity of collaboration magnified by the interactivity and the timer. (There was both cheering and hugging when we completed one in under 10 minutes and got the achievement.) And, for all that it was a source of endless frustration to our team, when we pulled the first round meta together and solved it (about six hours later than we really should have), we were cheering loudly enough to be heard down the halls.

The hunt's failing was the demoralizing stretch in the middle, when we were totally blocked on the meta and a set of four puzzles we'd simply been looking at for too long, and were no longer making any headway on. We reached the state where people who wanted to solve were going to sleep instead, and our team never really regained momentum. I think your assessment that "breaking trust" with the competitive teams at the top would have helped a lot of the teams in the middle was spot-on; I know there was a lot of table-talk on our team that we felt the hunt had been geared for the teams in the top few places, and that the rest of us were expected to "forfeit" by going recreational. We discussed it a few times in the wee hours of the morning, but we were bouncing between fifteenth and twentieth place at the time, and we simply weren't willing to give up a placement that was, for us, far higher than we were expecting to achieve.

In the end, those last few hours dominated our immediate memories of the hunt. When we got to the wrap and discovered there was an entire round we hadn't seen, we got bitter. We griped about the pacing - some of it loud enough that I'm sure members of the producing group heard. What was foremost in our minds was that a lot of those puzzles looked really fun, and we felt deprived of the chance to see them and solve them.

But, with a bit of time to reflect, one thing stands out. The event still succeeded, and succeeded tremendously. My team had five first-time puzzlers on it - five people who had never done any puzzle event at all. At the end of the event, all but one of them wants to play in the next puzzle event in the area. (The lone holdout has given a firm "maybe," citing lack-of-sleep.) Everyone on my team is still talking about favorite puzzles. And this from a team that spent almost six hours badly blocked on the meta, and were at each others (and the puzzles') throats for much of that.

I think most of the big experiments of this hunt succeeded. The ones that didn't (the blocking round-metas, the exact structure of the competitive / recreational split) still asked some important questions about how hunts are run, and have prompted serious reflection, at least on my team, about what we'll do in the events we're running to do a good job of supporting the different play-styles in the puzzling community.

So, take some well-deserved time off to play in events instead of run them. (We're looking forward to having you in our SNAP, certainly!) And, once you've had some more time and distance to reflect, I hope you'll share more of what you learned from this hunt, because it sounds like there's a lot to be learned for future event-running.

First off, thanks to all those that put on this event. I know first hand how much of a toll this can take on both your free time, your ego, and your relationships. And when I did it, I think I paid 300 dollars for the privilege.

There were several successes for this hunt, and those have been talked about a lot. The meta stall has also been talked about a lot. But what I see in the mails getting thrown around a lot is a bunch of talk from the people that finished the event about how great it was. I see very little from people who didn't finish or came in below 10th place or so. Last time I tried to critically look at puzzles, I got no real response or was put in the light of a hater and forced to argue from this position. I won't repeat that, so I'll just state my opinion about this last hunt and leave it at that.

This might very well be the last hunt I participate in. Similar to what Jason says, I really felt that if you weren't in the top 6 or 7 spots, this hunt wasn't for you. If you aren't in the top, you were expected to go recreational and in effect 'forfeit'. I don't really like being made feel like that, so I have to decide what to do about it.

In hunts 5-7, I very much felt like an outsider to the puzzle hunt crowd. I thought some of the explanations after the event bordered on sheer lunacy and didn't understand why I was supposed to make the leaps described. After PH 7, I also was considering quitting. Puzzlehunt 8 and 11 (especially) struck very different tones, and seemed very inclusive of mid-tier teams and up. While we didn't get every puzzle, I felt the puzzle's were fairly presented. I didn't want to hit anyone in the mouth during the closing event. This made me think the puzzle hunt was moving to a more polished presentation, and less craziness in the leaps expected. I thought it was impressive that 15 teams finished PH11.

I believe that this last puzzlehunt made a move towards the older style and puzzle types, to the exclusion of all but the top 7 teams. I also see many of these same people talk about how they really thought this was a great hunt, which makes me think it will be emulated in the future. I don't want to advocate changing on the behalf of the public, because I'm not sure that's the goal. If people want to have an event that caters to 70-80 people that they probably know at least as nodding aquiantences, that's fine. I just wish that people would be more honest about it. It would at least let me know if I should join up an enjoy myself. If told at the outset what kind of puzzles we'd be looking at and who the hunt was really for, I might have different ideas on how to spend my time.

I'm not upset by thinking of skipping on PH in particular. I do other puzzle events that seem to be more friendly for mass consumption, and I for the most part enjoy myself. Maybe the lesson is that PH as envisioned by some, isn't for everyone.

Oh, and by the way Peter, my favorite puzzle of the hunt was your walk of fame puzzle. That was the full package, and probably was the highlight of my 36 hours. So don't think I'm a complete hater

So, for the folks who were on self-admitted mid-tier teams: can you speak a little on what would make this hunt a better one for you? It seems like one thing is that getting stuck behind a meta is really frustrating since the number of live puzzles is (a) small and (b) hard to crack. What would make this better for you? Would it have worked for you if the second wave was released at a fixed time, say, 10 pm, no matter if the first meta was complete or not? Or is the issue that you'd like metas that might be solvable with fewer puzzles? Or is easier puzzles the answer? The teams that run these events are generally teams near the top of the leaderboard, so it's natural that they would create a hunt that they themselves would enjoy; so it's really valuable to hear what your teams would change to make it the most fun for you.

I would have had a little more fun without some guy named 'banana' running around and shooting me with a rocket launcher every 5 seconds. But it's hard to quantify how much. :)

John: I think the fixed-release second wave would have been a huge boon to our team; we needed time to work on the meta (and, in one case, to notice that we'd written one word on the wrong card, which took a terribly embarrassing call to Puzzle Control to figure out). While a few of us were working on that, it was frustrating to have eight teammates getting very annoyed at puzzles they'd been looking at for twelve hours.

The difficulty of this event was, by and large, what we bargained for. A few of the puzzles seemed a bit aggressively un-clued for our team's taste, but almost none seemed unfair. (Real People earned that "almost," though we actually solved that by judicious application of backsolvery and brute force.) We'd have liked to see a bit more music and literature in with the TV this hunt, but that's a factor of puzzle authors, and of the fact that nobody on our team watches mainstream TV. (We're mostly DVD buffs...)

Really, in the end, it was two things frustrating us: One was the fact that the meta was blocking for the whole team while a few of us worked hard at it, and a lot of awesome puzzles seemed to be locked away in late rounds where we never saw them. (Some of this is, of course, a grass-is-greener effect, I'm sure; we extracted every ounce of fun from those puzzles that we had, in the 24 hours we spent in their company. We only got a glimpse at the shiny new late-game puzzles, and never spent time banging our heads against them.) The other was, ironically enough, the ability to choose to end our frustration. We knew we weren't going to win - we knew that weeks before the hunt ran; we were a team where only four members had ever done a hunt before. But we weren't in the bottom of the pack, and we had something to lose if we chose to exercise the recreational option. So we never exercised it, and it became a point of pride for our team - but it also yielded a lot of fretting, because it was far from unanimous, and we had to keep telling ourselves that we wanted the frustration we were going through, because if we didn't, we'd have called Puzzle Central and converted. (To be honest, the fact that we had a reputation on the line is probably responsible for five of our solves - knowing we had a team to beat to claw back to fifteenth place kept us working until the buzzer. If we'd gone recreational, we probably would have looked at the puzzles, solved the easy ones, and given a collective shrug and gone to sleep. The competition is what motivates our team, even though we haven't got a chance in hell of placing in the top ten, much less winning.)

So, given that, what do we really want out of a hunt?

We want to see all or almost all of the puzzles over the course of the event.
We want to have a meaningful competition with the other teams at our level, without necessarily knowing in advance what that level is. (If you ask me what team I want to play against, I'm always going to say I want to play in the same league as Cracking Good Toast and TLA and Buzz Lime Pi and the others. I can't beat them, but they make for a hell of a good ride when we try to follow.)
We want to have a reasonable number of puzzles to work on at any time. We feel like the puzzle-stream shouldn't dry up unless we're nearing the end of the event, and it's time to hammer everything home. When people tune out because they're no longer contributing, we've found they end up staying tuned out, which hurt our morale.
We want to be challenged. This year's puzzles were hard, and that made it sweet when they finally fell. (Picture Page was memorable for this, in particular. When Dice fell, we were thoroughly happy, anagrams and all. And after the event, the player who single-handedly assaulted Rank and File said that it was probably the single hardest chess problem he had ever solved, and certainly the single most rewarding.) Don't lower the level of the event; just make sure that less-experienced teams don't get stuck hard early on.

As I said before, our team actually enjoyed the event. We were just frustrated at the same time, and some of the sources of that frustration can be avoided in future years. (Some frustration is inherent to puzzling. If the puzzles were easy, the event would be very different, and we wouldn't be happy.)

Gee, I had a really fun time at this event. (Thank you for your part of running it!) But I have other sources of fun, ways that don't involve you becoming miserable. So don't do that.

"It is probably the last puzzle event I will ever run with a volunteer committee."

Was the problem volunteers who un-volunteered? Or was it the committee decision-making style?

If it's the former, then I'd think you might want to run another game with the folks who stuck around to finish creating this game. Maybe by the end they were "ready to be done"--but they demonstrated their ability to finish a project.

If it's the latter... well, maybe there are ways to speed things up. E.g., put someone "in charge" of each piece; they're encouraged to talk to other folks in the group, but they have decision-making power--for their piece. That cuts out a lot of the committee-ish dithering.

Or maybe I'm misunderstanding what you're saying the root problem was. But I sure hope you can find a way to keep creating events like these... without driving yourself unnecessarily nuts.

After reading Jason's comments, I think it's quite possible that if there had been more available puzzles available throughout, my opinion of the hunt might have been much different. It's likely that had I the ability to switch what I was looking at more often I could have self adjusted away from puzzles that I found extremely frustrating and wouldn't have my experiences so colored.

What I want from a puzzle hunt is to see tough puzzles, but more asking how I can solve problems presented rather than ask what the problem is that is. I would like it if puzzles had clear steps to solving them.

Since it was brought up, I felt that the chess puzzle was just completely insane. I completely understood what needed to be done, but the permutations were terrible. Asking someone to do the 'reasonable' thing I didn't find fair at all. And the solution copped out and gave no definitive that the puzzle could have been solved. My gut reaction to seeing that puzzle was to write code. I should have followed my gut, maybe I wouldn't be so bitter. Grrrr.....

As someone who didn't attend, but loves reading about this stuff, could someone please explain what the "blocking" was all about?

And maybe descibe a few of the better puzzles.

Or just direct me to a link elsewhere that does this, as I'm not sure we're gonna get the standard Hunt "session report" from Peter any time soon :)

@Dave: The first stage of the hunt was presented as a Jeopardy! board. Each of the 30 spaces was linked to a trivia question and then to a puzzle. There was a meta-puzzle that used solutions to these 30 puzzles, although it could be solved without solving all 30 puzzles. (I think we had 27 when we solved it.) Solving that meta-puzzle unlocked another (smaller) Jeopardy! board full of puzzles, and was the only way to get to these puzzles. A third round was unlocked in a similar way. So twice during the hunt there were more puzzle available that were only accessible if you solved a particular meta-puzzle.

David: In very brief, the blocking we're talking about is this structure:

At the start of the hunt you get up six puzzles (in the top six squares of a standard Jeopardy! board). When you solve one of those, it opens up all of the puzzles immediately adjacent to it, until all the puzzles on the board are available. These thirty puzzles each provide an answer into the first-round metapuzzle. (By the way, we thought this structure was pretty cool. This is not the blocking we disliked. It was easy to use the adjacency to "solve around" puzzles that were giving us trouble.)

There is a second round of puzzles, but it cannot be accessed without completing the first round metapuzzle. There are twenty puzzles in that round, each feeding into the second-round meta. When you solve the second-round meta, you get access to the third (and more-or-less final) round of twelve puzzles, with its own meta. (And after solving that, there's a "usual" chain of meta-meta to do to actually win.)

So, teams that got stuck on the first-round meta never saw the second-round puzzles. Teams that didn't clear the second-round meta didn't see the third round. That was the "blocking" we're referring to: the metas blocked access to the rounds after them, and restricted the number of puzzles a team could see at any given time.

Dave, as the author of the insane chess puzzle, I want to assure you that I have NEVER written a puzzle that requires a computer to solve, though I have written several puzzles where players have chosen to write programs to do the work.

I make sure that they can be solved by hand by doing it myself - if I can solve it in under 4 hours, then I assume that the advantage gained by writing a solver from scratch is minimal. If people think that writing a solver is fun, then they're free to do that, but nobody should ever feel obligated to do so.

For puzzle concepts like this where I have to prove that a solution by hand is possible, I give my wife the parameters and ask HER to construct the puzzle. Then I have no clue what the answer is. In this case, I gave her a list of chess games that used all 64 squares of the board, asked her to pick one of them at random, and asked her to pick the 64-letter phrase. Because I could solve it by hand in about 3 hours, I put it in the event. Our playtesting time was similar. I think it would take me a lot more than three hours to write a solver for it.

I did thoroughly have a great time at the hunt. It was my first ever Puzzlehunt. Our team had similar experiences to Jason of Grey Goo with one notable difference.: we knew we were going to go recreational eventually. We wanted to have fun and we’d stay competitive as long as that was still the case. For the most part it was. Then the wee hours of the morning hit. It wasn’t that we were stubborn as some have suggested. We lost many members of our team to sleep or departure for the evening. The rest of us were paired up trying to decode things off of existing puzzles we knew how to solve (like B52 and Triplesec) and spinning out wheels on some puzzles we hadn’t advanced an inch on (Algebra II, Chemistry, etc). So by 3 AM, we were down in numbers, we WERE making progress on some and THOUGHT (foolishly) we were making progress on others. Also out brain functionality was diminishing so we weren’t really in our right minds to push the button then. And once we lost our leader to sleep, that was official. By morning we were rejunevated and actually solved 2 more and only finally started making progress on the meta. This also was rejunivating but not enough. We realized we probably wouldn’t get unblocked on anything else and also not solve the meta because we needed more data. So we went recraeational.

I must say I loved the hint system. It made the rest more fun actually. But even with the hint system, some puzzles were just way way way too much. I think that’s fault one. There were a handful of puzzles that should have gone back into the “fix this” pile. Their mechanics didn’t work or they were so ambiguopus or unintuitive that even when you find the secret , you still have hours and hours of work. To have puzzles like this be blockades to have more puzzles released is indeed a problem.

Then we have metas. I myself don’t like metas. I also don’t like extra hard puzzles that are trying to be clever. Metas blocking puzzle releases just seems counterproductive. Having a big event, fine, with a second round starting. I don’t know. I think it’s all a big Reagen trickledown theory. Some broken puzzles = no unlock other first round puzzles = no help for the meta = can’t solve meta = no entry into second round = no more puzzles…

Oh yeah, I had another thought about the hint system (possibly for future events). In keeping with the jeopardy theme it seemed like there could have been a way to merge the concept of hints for competitive teams and the jeopardy convention of deducting points…thus make the puzzle worth fewer points if you take hints. That could have worked and solved the whole blockage problem.

That being said I have been totally inspired by your puzzles specifically, Peter as the ones I saw by you were fantastic. My favorites that I worked on were Houses of the Stars, Hollywood Walk of Fame, MaiTai, Triplesec, Real People (yeah it was broken but we were proud to have the 3rd fastest solve as our big achievement (until the rec button was pushed  )
And of course the 1st daily double, Googolplex was a work of hilarious art (like a Mad Libs).

Thanks for a fun weekend!

Thanks for your comments, Jason (and the other folks), on what you were looking for. I also play the Mystery Hunt on a team that is not at the top of the leaderboard, so I've seen both sides of this.

There's a tension between "I want to see all puzzles" and "I want to concentrate on certain puzzles" (like the metas). The organizers do not want to go 100% toward the former, nor should they, and nor should we; otherwise we'd just want a giant flat puzzlehunt that gives us every puzzle at 10 am on Saturday. There's merit and fun toward concentrating on certain puzzles (metas). I'll say in the last mystery hunt, I'd probably have been happier if we'd done a little MORE concentrating on metas rather than be spread out over the large amount of puzzles that were available to us at the end.

But, spending hours and hours banging your head against a small handful of puzzles is no fun. So the compromise has traditionally been that new rounds are released on a time basis even if the previous meta has not been solved. It sounds like that would have gone a long way for you in this case (that at 10 pm the second wave might have been released no matter what, and the third wave at 8 am, or the like).

As far as the difficulty, I'm not sure that's a solvable problem. Different people want different things here and it's not tractable to write more than one version of a puzzle. For what it's worth, even after seeing the answers of the ones we didn't get, I personally didn't feel like anything was unfairly difficult. There were hard puzzles to be sure, but nothing unfairly hard. But that's my two cents.

Peter, I think you're being too hard on yourself. As someone who has run an event (PH9 for those that don't know me), I know just how soul crushing the burden of being expected to run a hunt can be, especially with insufficient or undedicated help. I also know just how scarring running an event like this can be; this is part of the reason I didn't jump in to help you guys finish this event. I'm only just now, some 3 years later beginning to consider running an event of this magnitude again, but next time I will create the event first and then when it is done, announce that I'm going to run it.

I commend you guys on trying new things, regardless of whether they succeeded or failed. I think this event was successful overall, despite its flaws - every event has these unless they are extremely lucky and the team running the event is extremely experienced. This is a big part of the reason PH8 was so successful, and why we were perfectly comfortable following their formula on PH9. We took few risks with PH9 because we'd never run an event of this magnitude, and limited our innovations to introducing an unlocking structure and a media rich interface for the first time.

The Daily Doubles introduced a new, in-conference room team event, which was a huge success and is now a new tool in the puzzle hunt creators' toolkit. Bravo!

The recreational vs. competitive structure may not have succeeded in exactly the way you hoped, but I don't think it was a bad idea. In fact, I think it was a good idea whose implementation was flawed by a faulty understanding of the motivations of the mid to low tier teams. The good news is, we've now learned a lot more about what makes these teams tick and the next hosting team can make a new attempt at achieving the goal of making the event more fun for everyone without disrupting the competitive standings of the top teams and possibly shifting the victory to another team.

The blocking meta at the end of each round was, in retrospect, probably a bad idea for mid to low tier teams, but I don't know how you could have anticipated this. Essentially, what you did is pulled the endgame experience (few puzzles remaining that had the team stumped and therefor frustrated along with a meta that you don't have enough info to solve) into the realm where it was experienced by the majority of the teams. The upper echelon of teams finds this to be the ultimate challenge and tackles this with aplomb, but you had no way of knowing that this could be a joy kill for teams without this level of expertise. You have done a service to all future organizers by making them aware of this problem so that they can try alternate routes in the future.

The one constructive criticism of the event I can offer to future organizers, since I know you already agree with me, is that making a meta puzzle be a physical puzzle is a Very Bad Idea(tm). The reason for this is because it inherently makes the meta a one-person solve. Only one person can manipulate the pieces at a time, which is the opposite of what you want for a meta. In a meta, you want it to be very parallelizable (is that a word?) so that every team member can be working on the puzzle individually or collaboratively and so that team members can check each other for errors (like the team that wrote an answer on the wrong card). This is what I think had the single biggest impact on slowing the overall pace of the event and increasing frustration among the teams, even if they weren't aware of it themselves. I don't think it was the ruin of the event, but I would not recommend that anyone does this in the future.

Overall, I really enjoyed the event, despite the somewhat surreal fusing of themes. I think you should be proud of the fact that you delivered over 34,000 hours of entertainment to a community of puzzle fans. So what if not everyone enjoyed it equally? Running an event like this is really, really hard (no, I mean really, really, really hard), and anyone that successfully delivers an event, no matter how they do that or what compromises they had to make along the way, should be proud of their accomplishment. I have my own regrets about PH9, but with a few years of hindsight, I think that PH9 was by-and-large a successful event - I think you will find the same to be true for your event.

Cheers to you and everyone that contributed, but especially to those that helped finish it. This was no small accomplishment.


I believe you when you say you solved it in 3 hours, and that play testing also gave the same. However, I don't see any way that this makes it a fair offering. You offered in your solution no step by step manner in which to solve the puzzle. You offered no explanation or justification for the term 'reasonable' in playing chess. Does this mean I should be fairly knowledgable in chess, or do I just need to know how to make the pieces move? When presented with something similar in the future, is there any way for me to prepare for it in advance or is my team just doomed in the face of such a task? Given that I can assemble a 2 stroke engine in a couple of hours from a bunch of pieces after cleaning, is it fair to ask people to do it during a puzzlehunt?

Perhaps it comes down to what people see as a puzzle and how they derive entertainment from it. Just because something is difficult, doesn't make it a puzzle nor entertaining to me. If the answer is to brute force something until a solution presents itself, that's just work to me. Lots of puzzles in puzzlehunt have this wall you need to overcome in order to begin working on the puzzle in any real way. This one (thankfully, in retrospect) put that wall up front and in your face. So at least it wasn't as brutal to me as the dice puzzle which took a chunk of time up front to discover that we couldnt' scale the wall to solve the puzzle.

Hi! Speaking as the Grey Goo guy who actually solved the chess puzzle for our team, I might be able to answer this better than Kenny can.

The puzzle does unquestionably require chess knowledge; a team could probably look up the basics of chess notation (the first hit on Google points to the Wikipedia page on algebraic notation), but that wouldn't necessarily give enough intuition to be able to recognize a few of the erlevant patterns. The single biggest piece of chess knowledge that made a difference to me in solving this puzzle was some opening theory; the gateway in is to recognize the 'pawn move, pawn move; knight move, knight move; bishop move, pawn move; bishop move' opening as being almost unique to the Ruy Lopez (the B-P-B sequence being the biggest giveaway) and knowing a little bit of the theory of that opening (namely, that the black pawn push/white bishop move sequence slightly later is probably ...b5 / Bb3), as that gives nine squares to start with and helps with the analysis. A few other scattered examples from my notes: on move 20, White plays (enciphered) 'kxn1'; the lowercase k means that this is a pawn move, but White hasn't moved a pawn onto any k' square in the transcript to that point, so it must be a capture from a pawn on its original square to a third-rank square. On move 11, White moved a knight to n1'; at that time, his knights were (from the opening) most likely on d5 and f3, so n1' is almost certainly the square e3. Tracking the king walks gives a lot of information -- since they can only move one square per turn, you get a lot of adjacencies between squares that can rule out other possibilities (e.g., r4' -- the square that a knight moves to n1' from -- has to be d5 instead of, e.g., g5 because the black king goes from r4' to e4' (which we know from the opening is a4) in 3 moves.) Checks give a lot of information, as do captures and pawn promotions; and one about 35 squares or so fall, quite a few others topple quickly to backsolves ('MOVIE FEATURES A SERIAL KILLER' fills itself in easily).

The broader question is an interesting one, though. This puzzle unquestionably required above-average levels of chess knowledge; I think it's fair to say it would've been virtually impossible if no one on the team was a chessplayer. But at the same time, I think the same could be said of many of the Scrabble puzzles (and not just this year's); they may be theoretically possible without Scrabble knowledge, but that doesn't mean they're practically feasible. Chemistry lived up to its name, and so did Algebra II -- knowing a little basic group and ring theory helped immeasurably with that puzzle, enough to bring it from 'slog' to 'straightforward'. I think this points up two major points, one for designers and one for teams: designers have to make sure that their metas are at least partly backsolvable -- a team should never have to solve all puzzles to solve a meta (to its credit, of course, this one was, though maybe not quite enough so). But on the other hand, puzzles like Chess and Algebra II and Cannibal Housewives and even Real People point up that it's immensely to teams' benefits to have as much diversity of knowledge and experience as possible.

Ack, sorry for the double post -- I was getting server errors and assumed it hadn't gone through.

As I think about it, incidentally, I'm actually surprised there aren't more chess puzzles involved; they're one of the oldest and most well-established form of 'brain puzzler', certainly much more so than many of the forms that are standards now for events like this, and especially for PH-level events with large teams and higher expectations of knowledge and skill it seems like they'd be a more common tool in the box. Is there an unspoken schism somewhere between chess puzzlers and non-chess puzzlers?

Dave, I completely agree with you that the best place to put the wall is right up front. I don't mind having a wall in the middle if that wall is thematically related to what you were doing previously, where scaling that wall enhances the work that has gone before. For example, Kitchen Sink has a wall in the middle that directly connects to the initial work, and nicely ties the initial, current, and future work together. BTW, that wall was originally much closer to the middle when there were no letter breaks provided in the first phase; we cut that for overall time, but solvers found it very rewarding in its original form too.

On the minus side, I'll call out one of my own puzzles here (Hunt the Wumpus) which had a wall near the end of the puzzle that was trying to make sense of a set of poorly-formed letters - while there's only so well you can draw letters on an icosahedron, the frustration here did not enhance the work that came before and trivialized a lot of the initial fun. That's not good puzzling and I'm sorry I put people through that.

Specifically on the chess puzzle, I think a lot of solvers *like* being confronted with a puzzle that they are completely unprepared for, and which looks impossible. Then, upon discovering that they can in fact make progress, reaching the final solution is enormously rewarding. I understand that this isn't the kind of puzzle you gravitate to, but I think it has a place in the event. And yes, I would personally enjoy being asked to assemble a two-stroke engine even though I have no experience - especially if you had playtested the experience beforehand and found it to be achievable in a reasonable length of time by relative mechanical novices who like puzzles.

And by the way, I am a poor chess player - I know how the pieces move, but I know none of the opening book, mate-in-two puzzles tax my brain, etc. We put the term 'reasonable' in there in order to convince people who liked chess that the game wouldn't be a dissatisfyingly haphazard set of moves designed to cover the board rather than make any sense. Sorry if that threw you off.

The quality of the solution page has zero bearing on whether a puzzle belongs in an event. In this case it's akin to posting the solution process for Japanese-style logic puzzles like Sudokus, Kakuros, Nurikabes, etc - it just isn't generally done.

Ack, "anonymous" was me, I forgot to fill in the field.

This was only my second Puzzle Hunt, but I have to say that I had a great time. I hope you can take encouragement from the fact that a relative newcomer like me enjoyed Puzzle Hunt 123 immensely. We got to work on a lot of cool puzzles and spend the weekend with friends, and what more could you want?

First, I had a great time with PH123, the first Microsoft Hunt I've done, in part because of opening up of a Bay Area simulcast (another great idea for the community). The innovations to the conference room game like the Daily Double were phenomenal and the many great puzzles I ran into have been inspiring, this week, as I begin puzzle construction again, a month after running a Mystery Hunt that left me feeling the same way you describe, never wanting to even approach a task that big again.

The similarities are a little surprising but the lessons learned I think common. My team, over-confident we could get teams through a meta-meta bottleneck after we got all teams through a single meta bottleneck in 2007 for the Hell Hunt, had to make several similar "3AM decisions" not only as to how to get all teams to see the puzzles, but just to get the lead teams to find a coin fairly. Combined with errors in the files we were releasing to teams on metas, which lengthened the Hunt, it was a demoralizing experience. For at least a month, I never entertained a thought of running another Hunt-sized event. I was completely drained both mentally and physically (and sick to boot).

PH123, the first event I've competed in since that Hunt, has gotten me excited about doing such an event again. There was still a lot of quality and polish to Puzzlehaunt, particularly given its history of joining to separate hunts. The puzzles themselves weren't buggy and the interface and achievements were fun. It wasn't without its problems, but I had a tremendous amount of fun with it. We can all take something from the lessons of running such a large hunt. Its good to have a leadership organization and not govern by committee. Its good to think about bottlenecks, particularly if they block puzzle releases for a large number of teams. Somehow, all puzzles should be able to be opened by all teams before an event ends. Not all new ideas succeed, and a recreational league might not resolve the "how to hint" question even though I would also have agreed it sounded like a step in the right direction. All this said, I greatly value what you and others did to construct this event and I hope that you will (with time) come to realize you can be a part of one of these games again.

I didn't think the Chess puzzle was unfair. The key to breaking into it was the opening (which as mentioned was relatively common), the moves where both squares were specified (the pawn captures, as well as the several rook moves when both rooks could move to the location), and the king moves at the end. From that, you could get ~20 squares, and a sense of which squares were in the middle of the board. Playing it out from there was relatively straightforward, especially if you can find words on the board.

Some of the other puzzles were less fair, but's that to be expected; Hunt the Wumpus (letters on the icosahedron?), Triple Sec (not that hard, but took forever), and Eight Bits Make a Dollar (indexing into the script?).

I don't think Peter's being too hard on himself. It's precisely because he has high standards that he's a great puzzlehunt designer, and when something doesn't meet those personal standards, your friends shouldn't give you a free pass. Peter should be commended for this kind of self-reflection.

I received an anguished call from a team trapped in the "fight or flight" dilemma Peter described. It was awful. Based on the standings, I could theorize that a lot of teams were in that position. That's a bad position to be in. I know, because I helped make that possible for MIT Mystery Hunt teams the month earlier. We had a very similar bottleneck, and it damaged team morale.

So I'm now done with bottlenecks forever. If an event has a "you can't get to a major area of the hunt unless you solve this one metapuzzle," it's BROKEN. We don't need any more illustrations than these two hunts. The emotional cost of bottlenecking, especially with the potential requirement of switching competitive status to overcome it, is too high.

So given that lots of hunt designers are reading this, I think it's time we stopped saying it's a good idea. It might be best to focus on writing events where PUZZLES get you advancements, not metas. Puzzles are about progress; metas are most often about capstoning progress. If you have 30 puzzles open and you can't advance, that's your fault. But if you have 3 puzzles and 1 meta open and you can't advance, that's the organizers' fault.

Now here's where I say I didn't actually play this hunt, since I had a family issue that kept me away. So it's possible I missed something. But the descriptions from my teammates and other players tells me Peter is spot on with his analysis. Whether he deserves the disproportionate amount of self-criticism is probably NOT spot on, but I felt like he did last month, and I feel better now. So it'll pass. But we still should learn.


I dont think having META or physical puzzle is a bad idea. In fact, METAs were the biggest reason why I started participating in the puzzlehunt. My best memories of the puzzlehunts are that of the META puzzles.

I think the problem here is that the META was too geared towards not being solvable with fewer answers. It was like the puzzle designer specifically wanted the solver to have all the answer before the META can be attempted. So in a way, he/she achieved their goal.

I, personally, spent roughly 7 hr on the first round META. It was the worse META I have worked on, till date.

That being said, for a META (esp. those which open new puzzle rounds), I think the following should be true.
1) Given ALL the solutions, the solution of the META should be relatively straightforward. In other words, once I have all the clues/solution, META should be the easiest puzzle of the round.
2) Given ZERO solutions, the solution of the META should still be evident. That is, though the solver cant get the answer, the way to solve the META should be obvious or workable.
3) Given ONE additional answer, it should help solve the META further. That is to say, that the answer to the META should become more obvious, or the method more straightforward.
4) Given ALMOST all solutions, META should be hand solvable.

Speaking as a member of a mid-tier team, we did enjoy the hunt and weren't too badly gated by the meta. I ended up solving the meta by writing a program to solve it -- turns out this works with about 1/2-3/4 of the answers, so I did this while others still worked on first round puzzles or tried solving the meta by hand. I wish I had started work on this sooner. After that we really were gated as we took far too long in our sleep deprived state to solve "Intermission."

I guess the lesson is don't put a difficult blocking meta right in the middle of the event. Past hunts have had an easy blocking meta at the beginning, a difficult blocking meta at the end, and multiple round two metas, none of which would block you on making progress toward any of the others. That structure seems to work well. And now that we've tried we know the recreational/competetive league doesn't work. We did like the acheivements idea -- it might be interesting to have some of the available acheivements known before you get them so teams like us who don't expect to complete the event can pick some as goals.

When I first saw the new rules about the Competitive vs. Recreational divison, I thought "Interesting. Makes a lot of sense to me." But when I told one of my fellow team members on the Burninators, his response was "Wow, that sounds like a bad idea." We discussed further. Eventually, I started seeing his point of view. Let me try to state it in my terms:

People participate in events to have fun, be they footraces, board games, or puzzle hunts. When you're organizing events, the fun of the participants involved should be the top priority, and that fun is derived from things you do in the event. Running in a footrace. Making the best move in a board game. Solving a puzzle in a puzzle hunt. But, one needs to be VERY careful with regards to meta-event decisions. What do I mean by "meta-event decision"? Here I don't mean metapuzzles, I mean the decisions and activities that aren't part of the event per se, but relate to how one participates in the event. Spying on your competitors for their running styles and weaknesses before the footrace. Trying to figure out if it's such a good idea to make the best move in a board game if it means that you're going to make your wife angry and that you'll be sleeping on the sofa tonight. And, trying to decide if you're going to move to the Recreational division in the MSPH.

While some people enjoy meta-event decisions, the majority of participants don't. When they participate in an event, they are trusting the organizers to give them a fun time. They don't want all these dials and settings to "control their own fun" with; it's like going to a restaurant and having the waiter show you into the kitchen so that you can cook your own meal. Maybe that's exciting for some, but the casual patron would not find this fun.

But here's the thing. The people at the top, the "power users" if you will, LOVE this sort of stuff. Customization? Full control? Hell yeah! And in a community-created environment such as puzzle hunting, the people who end up organizing these things often are the "power users" to begin with, and well, it's easy to design the parts that work well for people like yourself; not so easy to do it for people who aren't like you.

On to the more specific issue of Recreational vs. Competitive. The problem here is that you're asking teams to have to evaluate THEMSELVES, to not only make a judgment call on whether they're having fun or not, but also to expect them to have the presence of mind to even ask the question in the first place. And guess what, that's a meta-event decision, and most of the teams aren't going to do a very good job of it.

A specific case that I think is worth mentioning. During the Gooooogol Puzzle Hunt, with about 2 hours to go until the end of the hunt, we decided to send out little caregiving missions. I would go around to each of the team conference rooms, ask them how they were doing, give them hints on any puzzles they were stuck on. It turns out that, every single team I went to welcomed the visit. Not a single team tried to shoo me away because they wanted to stay competitive. Every team I visited had many puzzles that they were very happy to get unstuck on. This, frankly, was a big surprise. Why? Because we didn't give any penalties for hints! How come these teams, who could have called for a hint any time, chose to wallow in being stuck when a hint would have made them much happier?

The epiphany I had was this: Except for the teams at the top, most teams don't realize when they can be meta-event-ing to increase their fun. For the teams at the top, the Puzzle Hunt is a competition. If you're at the top, you are a Competitive team, and you love doing meta-event stuff. You're calling GC and trying to perfect the art of "trying to get a hint from GC without them realizing it." Heck, during the actual puzzle hunt, I felt like I "gained the system" by asking for data confirmation in a specific way that allowed me to get a hint in an indirect fashion! For everybody else, this whole gaining-the-system thing is just confusing. You're here to solve puzzles and have fun, and you're trusting GC to make the experience fun. If you do end up gaining the system, you're probably not even doing it on purpose.

What makes this very hard on GC, though, is that this whole thing is a continuous spectrum. The Competitive/Recreational division probably worked very well for the 10 or so best teams, and probably the 50 or so worst teams. For the poor teams in the middle, though, it probably forced them to choose sides when it really is more of a grey area.

Overall, was the Competitive/Recreation experiment a good idea? Absolutely. I would even go so far as to say that with some changes at the boundary, it could even be continued -- it was by no means a failed experiment. But I'm very happy I didn't have to be the one conducting it.

Mike makes a pretty absolute statement here:

> If an event has a "you can't get to a major area of the hunt unless you solve this one metapuzzle," it's BROKEN.

The problem with this absolute, in my opinion, is that, well, it completely rules out linear hunts, such as The Game. In a linear hunt, you don't get puzzle n+1 until you've solved puzzle n. Therefore, I would argue, you can't get to puzzles 2 through k (a major area of the hunt) until you solve puzzle 1. Is it broken?

I say no. It's just how the hunt is designed. But in most linear hunts, all the teams and GC are under the understanding that this is how things work. A team that gets absolutely stuck on puzzle 1 is not going to be told, "sorry, we're not going to let you go on, it wouldn't be fair to the other teams."

Here's how I would amend Mike's statement. If an event has a "you can't get to a major area of the hunt unless you solve this one metapuzzle," then it is almost certain that, no matter how easy you think the metapuzzle is, that some teams are going to get stuck. And, unless GC wants teams to stay stuck, they better have some sort of mechanism for dealing with it. ACTIVELY dealing with it, and not just "oh, the team can press the big panic button we gave them."

In the Gooooogol Puzzle Hunt, we had a reasonably complex system where if you solved the 7 easiest puzzles, you got a piece of a big meta, and when you solved that big meta with all 7 pieces, you unlocked a whole area of the hunt (which had its own meta structure, but I digress). Is it true that, if you didn't solve that first meta, that a big area of the hunt became inaccessible? Yes. Was this a problem? No, because we were aware of it, and we were paying careful attention to teams that seemed like they would need that meta to get further. (It helps a lot by not saying up front that you're going to have a Competitive division and that you're going to be equinanimous with hints.) On the second day of the hunt, we were paying very careful attention to the teams that hadn't solved that meta yet, even coaching them through some steps.

Wei-Hwa said: "Except for the teams at the top, most teams don't realize when they can be meta-event-ing to increase their fun."

I think what's actually happening is that newbie teams, who are less familiar with how puzzle events run, are more likely to adhere to whatever explicit rules have been given. If there are no rules for something--e.g., how often they can ask for hints--they go by their own prejudices or assumptions about the event.

More experienced teams, on the other hand, know more about what happens behind the scenes, and in some cases may even know the event organizers personally. You're much more likely to call your friend for a hint than to call a total stranger whom you know only by his or her imposing title of "Game Control."

Both experienced and newbie teams are following the rules as they understand them; it's just that more experienced teams have a better understanding of the unwritten/unspoken rules. There's no way to eliminate that knowledge gap, but GC can do their best to be explicit with the most important stuff and treat everyone fairly and equally.

And now for something completely different: http://snout.org/512 :)

There's a tension between "I want to see all puzzles" and "I want to concentrate on certain puzzles" (like the metas). The organizers do not want to go 100% toward the former, nor should they, and nor should we; otherwise we'd just want a giant flat puzzlehunt that gives us every puzzle at 10 am on Saturday.

Intern Puzzle Day does this, doesn't it? The one I was in (2007) gave out all the regular puzzles at the beginning, and an easy meta that you needed maybe 40-50% of the puzzles to solve. Once you solved the easy meta, you got the hard meta, which solving would let you win but needed nearly all the answers to solve. (It was trivial if you were missing one, which is what my team, which one, did, but maybe possible with 2-3 missing.) There were events that only opened up at particular times, but solving them just gave you clues to the hard meta.

M moncler jas y wife and daughters happening to return a visit to neighbour Flamborough's, found that family had lately got their pictures drawn by a limner, who travelled the country, and took likenesses for fifteen shillings a head. As this family and ours had long a sort of rivalry in point of taste, our spir

But then mont blanc fountain pen uk he wanted to do something for her, to help her.
"Come," he said. "We'll check you."
She stiffened in his arms.
"No, no," he said quickly. "Don't be afraid. I'm sure we won't find anything. But if we do, I'll cure you. I swear I'll cure you, Ruth."
She was looking at him in the darkness,

  Having for many years felt love in my heart towards the n mont blanc uk atives of this landwho dwell far back in the wilderness, whose ancestors were formerly the owners and possessors of the land where we dwell, and who for a small considerationassigned their inheritance to us, and being at Philadelphia in the

full of pith and idea. "The greatest m montblanc pens onument ever
erected to Napoleon Buonaparte was the British National
debt," said he yesterday. Again, "We must never forget
that the principal export of Great Britain to the United
States IS the United States." Again, speaking of
Christianity, "What is i

warning take a meaning now. Oh, i monte blanc pen f I had only listened
to him then, before I helped by these repeated sittings
to forge the links of the chain which binds me! But I
will see him to-day. I will apologize to him for
having treated his warning so lightly. I will see if
he can advise me.

4 P

.. Site selecti mont blanc fountain pens on & construction: Assisted owners in locating and selecting site in up-and-coming
downtown neighborhood. Oversaw building construction and club layout. Represented owners at
community meetings and before licensing boards.
.. Revenue protection: Established foolproof ticketing s

  Correlation of Growth

  I mean by this expression that t mont blanc uk he whole organisation is so tied together during its growth and development, that when slight variations in any one part occur, and are accumulated through natural selection, other parts become modified. This is a very important subject, m

The laws of science, as we know mont blanc fountain pen uk them at present, contain many fundamental numbers, like the size of the electric charge of the electron and the ratio of the masses of the proton and the electron. We cannot, at the moment at least, predict the values of these numbers from theory - we have to find the

The reason was obvious and disappointing: simply that Sig montblanc pens nor Mantissa himself had been through them all, each booth was a permanent exhibit in memory of some time in his life when there had been a blond seamstress in Lyons, or an abortive plot to smuggle tobacco over the Pyrenees, or a minor ass

But the gene mont blanc pen ral in him was foremost, even in the excitement of battle. The number of his opponents, their next move. These things concerned him seriously.

He searched the line of light with eager eyes. He listened to the sound of movement. These things were all he had to rely on, and on their

It gladly accompanies the godly monk as he climbs the jade mountain.��
monte blanc pen The Tang Emperor then ordered that the cassock be unfolded. On examining it from top to bottom he saw that it was indeed a fine article.
��Elder of the Great Law,�� he said, ��I tell you truthfully that I am now propagati

(technically, subatomic fusion) created mont blanc pens cheap when its beam intersects a shield.
LEGION, IMPERIAL: ten brigades (about 30,000 men).
LIBAN: Fremen liban is spice water infused with yucca flour. Originally a sour
milk drink.
LISAN AL-GAIB: "The Voice from the Outer World." In Fremen messianic legends, a

Monthly Archives