Static Zombie

We felt kinda the same. I guess we're a tier two team now, finishing 9th (1 minute behind 8) but in the early afternoon. This was the first time we've finished the hunt which we were quite happy with, but at the same time, felt like there could/should have been more. It was a very solid hunt. Bug free, ran very smoothly, and I thank Buzz Lime Pi for throwing it. It was great to get out again.

So when is CGT hosting again? :)

"Perhaps the problem is that the top teams are just too good."

I think the top teams are just too big. Just for a start, it's damnably hard to create enough high-quality puzzles to fill 360 solver-hours. But it's actually worse than that, because of network effects. Each puzzle gets attacked by the person out of 12 most interested in and experienced with that puzzle type, and that person gets the benefit of 11 others' intuition.

That's most of why playtesting doesn't reveal these problems. It's rare enough to have a full twelve solvers working in one room; maybe PHA did, but PH9 sure didn't, not ever. It's rarer still to have those twelve solvers be people who've all worked together before, who know each others' strengths and weaknesses, and who feel comfortable swooping in.

Shrink the team size and I think this problem becomes much more manageable. You'd need to increase the number of teams competing, but I don't think the increase would be proportional. If you took the most extreme approach and cut the team size in half, for example, I'm certain that the number of teams would not double, as most teams currently have to struggle to fill their ranks up to 12, and generally do so with people who are something less than hardcore.

I'm going to speak selfishly here... I think the weekend-long events are great for my husband and friends to participate in, but I really wish there was a 12 hour event, so that I could do this sometime! (without leaving our children to be raised by wolves in both my husband's and my absence!)

---Ellen

This was my first hunt, and I quite enjoyed it, although I was also surprised at how soon we finished, around 5:30am. I definitely think it would be more enjoyable with smaller teams. By comparison to the others on my team, I felt like I contributed to a large fraction of the puzzles, but this was still less than half of the ones available.

You'd need to increase the number of teams competing, but I don't think the increase would be proportional.

Maybe not, but increasing the number of teams has the difficulty of increasing the work for the host team. Most physical puzzles (Rubik's cube, Round 1 meta, etc.) take a fixed amount of time to prepare. Add more teams, and you add the need for more prep work.
There's also the problem of gated puzzles, where only a fixed number of teams can participate at once (swimming). Increasing the number of teams means it takes a lot longer to get through those puzzles, and if you have a site with limited availability (like the Pro Club) you may not get all the teams through that want to get through.

One workaround for this would be to cap the number of teams. The hard-core puzzlers would be more likely to sign up early enough to make the cap. But having to turn away interested teams might generate ill will.

The situation is a natural result of the way the puzzle community is growning, no doubt helped by the way puzzle competitions seem to be becoming more mainstream. I don't think there's a simple solution to this.

I too was surprised at how quickly the hunt finished. Had I know it was going to happen that way, I would have paced myself differently than I did.

It's interesting how the number of teams who finished has been used as a metric the past few hunts. If an organizing team wanted to optimize for that metric, they could make a single puzzle consisting of "The final answer is 'foo'."

Anyways, as the number of participating teams grows, it's definitely going to be harder to keep top teams entertained for an entire weekend and still have a significant percentage of the participants get to work on all the puzzles you wrote.

This was the second puzzle hunt I participated in. The first being on CGT in PH2.

So, we finished 33rd this time with many first time puzzle hunters and the rest with very little experience. The time before we won the hunt.

It's interesting. I had a heck of a lot of fun with both experiences. When I was on CGT it was a well oiled machine... Or at least it was compared to what Drunken Robots, not that DR wasn't doing things well...

If we took the top four puzzle solvers from CGT and the top four from DR, CGT would win. No doubt about it. The difference, though, is that if you took the top four from CGT and took the whole of DR it wouldn't be too different a finish. I'm not saying that contributions weren't made by all members of our team.

If you took the top four off CGT and pitted them against the second four off CGT. The top tier teams are basically all star teams competing against the rest of the league.

What do you think?

I was on the top-rated team that did not finish the hunt. We needed about another 30 minutes, maybe not even. But our team is less competitive than yours; we stay all night, have some solvers than we work on between hunts (but not many), but I don't think we're as committed as the top teams are generally. We don't do too much between hunts to stay in practice, for instance.

So basically, demographically, we're pretty much the same as Buzz Lime Pi is in terms of puzzling and where we usually finish. usually hovering around top ten but maybe not quite in it. This year we were 13, which was a little lower than usual for us. And for us, this hunt was designed almost exactly right.

What I think happens is that the organizing team tends to design a hunt that can be finished in the allotted time by teams that most closely match their own in skill and intensity. When you guys are organizing, it's very likely that only a few other teams will get to finish the hunt -- because you're optimizing the hunt for teams like yours. BLP optimized for teams more like theirs. I doubt that happens consciously, but if you look over the last several hunts, there does seem to be a correlation between the competitiveness of the organizing team and the number of teams that finish the hunt.

> If an organizing team wanted to optimize for that metric, they could make a single puzzle consisting of "The final answer is 'foo'."

We didn't optimize for that metric. We tried to optimize for fun. I'm sorry that Peter didn't have an optimal amount of fun, but I'm pretty sure a lot of other players did have a lot of fun.

> BLP optimized for teams more like theirs. I doubt that happens consciously

No, it was conscious (and somewhat controversial within the hosting team, of course). BLP was thrilled when we first started solving metas; even more thrilled to complete a hunt.

I'm really intrigued by the reduced team size idea. I don't think simply reducing the cap will solve the problem -- you'll still have a zipf distribution of smaller teams.

But what about handicapping? Here's a first idea: Enroll any size team you want (max 12). The performance of a completing team is measured not in hours, but person-hours. (Based on roster; no discount for no-show team members. :v) With handicapping, CGT can play a whole weekend with 4 people, and a milder team can complete the same hunt in the same weekend with 12. It's an intriguing idea. (The thematic "endgame" would have to be adjusted a bit, since arrival order doesn't determine finish placement, but that's a solvable problem.)

In his post, Peter prefers the old-skool philosophy: design the hunt to err on the too-hard side, and then if you find you're running long, hint the top teams to victory. Yuck! Frankly, I really disliked the hunts I played in where the hosts had to hint us along. Hints destroy the "aha" moments, which is what the players come for!

Buzz Lime Pi tried hard to design a correctly-timed hunt. (Our two betas indicated we were running long. In fact, at 10pm on Saturday, Jay was still firmly asserting that we wouldn't have a finish until well after sunup.) But we deliberately erred on the side of "too short for Peter" (and just right for the many dozen non-blogging players :v) to avoid the side of "too long with a hint-storm disaster on Sunday". While I think that future hosts should contemplate improvements (like the team-size handicapping idea), I'm satisfied with the way PHA turned out.

Since this seems likely to become the canonical thread for PHA, big congrats to BLP for running a very, very solid hunt. Great, great work!

I think the real problem is this: "The disparity between the top teams and the next tier continues to grow." In particular, the top teams are getting meaningfully _faster_ - instead of having two or three talented people who have done lots of hunt-like-events before with other very experienced hunters and thus have the right intuitions to break straightforward puzzles, those teams now have nine or ten people with those skills. Pair them up to beat puzzles that are in the same family as ones we've seen before, and that's what's going to happen. This is also really difficult to pick up in a beta test or everybody-takes-home testing.

(Similarly, teams that had nobody who could do this now have have six people who are just better - thus 12 finishers this year.)

I don't believe this hunt was objectively simpler than PH3 (i.e. the first hunt where Google mattered) or PH4, for example. It was just similar in form and function, and at this point people have seen ~400 puzzles.

That's why I don't believe that size changes, for example, will really change that - it's an artificial slowdown, sure, but the disparity won't go away, and over time will continue to increase.

One thing that makes the MS Puzzle Hunt unusual in this space is that it does enable teams 15-30 and 30+ to have some fun solving (though since I've never lived through that, I don't really know from which I speak). I don't know what smaller teams would do to that, but I don't see how it would be good. I'd take the known accessibility of PH over the questionable benefit of structural change.

While I understand the desire to be competitive and play with people you know, I suggest that this isn't a problem for the event organizers to solve: it is a problem for the participants to solve for themselves. Are you on a team that isn't competitive enough for your taste? Form a new competitive team or kick a few people off your team to bring in fresh blood. Is your team too good for events to be enjoyable any more? Split your team into two teams and bring in some fresh blood. Not only will you be self-handicapping, but you'll also be teaching other people to be better puzzlers - breeding your own competition, as it were. I don't think it is fair to ask Hunt organizers to solve a problem that is within the team's own control to solve.

On a side note, I thought the final meta was absolutely dreadful. Note to future event hosters: please for the love of god NEVER build a meta-puzzle that requires darn near all of the answers to all of the puzzles. No team solved this puzzle without 92% or more of the base puzzles solved, even if all the penultimate metas had already been solved. If the team hasn't solved those base puzzles, it is because they think those puzzles SUCK, so why do you make them go through them if they already have all the other metas solved? In my opinion, you should be able to solve the each meta with 75-80% of the base puzzles solved and solving the final meta only requires solving 80% of the metas leading up to it (no more base puzzles required). TLA did this same thing on PH7 and it sucked then too.

Overall, I liked the event despite the fly in the ointment discussed above. The puzzles were mostly solid with a few brilliant ones, but more stinkers (4-5) than in the past. My biggest diappointment was the return of the "puzzles in which you can do absolutely nothing until you have an aha stroke of genius". Common Threads was a great example of this. Most of these puzzles were sorely lacking on internal, built in hints, but fortunately enough of them were approachable that it wasn't devestating.

The best part of this hunt, IMHO, was the introduction of the automated answer system. No more worries about waiting on the HQ folks to answer if you were right or not. I think it still needs some tuning, but it was an idea whose time had come.

> It's hard to argue with the notion that the right place to target is closer to the center (or right of center) of the bell curve.

As Jon said, we made a deliberately decision to aim the focus away from the top teams; not to the center of the bell curve, but at the second-tier teams. The Hunts I've participated in definitely seemed to be designed for the amusement of the top teams. Perhaps the most clear signal of that was that the handful of finishers were treated to an exciting spectacle that nobody else saw. One of the innovations I'm most proud of in PHA was to turn that into something everyone got to enjoy. I hope future hosts make it a tradition.

> the hunt was essentially unthemed, as aside from a beginning-, mid-, and closing-game skit, there was no Atlantean flavor anywhere in the body of the hunt.

I understand where Peter is coming from; the puzzle theming was somewhat incoherent, and I take responsibility for that as the team member in charge of the theme. But I think he overstates it. As another member of our team said, "there was no theming whatsoever in any puzzle... even ones that required us to be completely submerged under real water."

I'm sure we'll hear more about the ideas discussed in this thread. The handicapping idea is very intriguing. But of course the ultimate contribution to the conversation is to design a hunt that rocks, and Peter will get his chance soon since Death of Dr. Zero seems likely to pass on their invitation to host.

> A is 10 in hexadecimal. Oh, those crazy computer geeks!

Apparently C is for Cracking Good Toast... I'm looking forward to it!

> No team solved this puzzle without 92% or more of the base puzzles solved

Sorry; our beta testing didn't bear this out. I (one guy in a room by himself with an Excel spreadsheet) test-solved the ubermeta with only 37 of 42 answers (88%). We figured the likes of CGT could do even better.

I think I hear violins playing. When your biggest complaint is that you finished the event too quickly, you're in a pretty cushy position.

The trend that more people finish the hunt can only be considered a good thing. Why waste time and resources designing cool activities that only a few players get to enjoy? Pushing more people through a difficult hunt by releasing waves of hints is a pretty crappy solution. Forced hints tend to suck in general - either they tell you nothing that you haven't already figured out, or tell you more than you need to know to get going again. When the hosting team has enough staff to do so, personalized hints on demand make for a much more enjoyable event.

Moving the final skit up to the main stage at the wrap-up was a brilliant move. I've sat through previous events where the final stages were revealed, and it truly sucks to miss out on some amazing spectacle. Along these lines, I wish there were some way to gauge when something neat was waiting for players that get sent out to do something. I'm a little bummed I didn't venture over to RedWest to see the neat PVC device, although playing with it at the ending was almost as good. Maybe I just had poor luck this time - my team finished the communications device while I was over in Cafe 9 peering for scraps of paper taped to the undersides of tables...

In an event with no artificial throttle on solving speed, it's really tough to have a high solve rate for the final meta, but still remain meaty enough to satisfy the fastest teams. As Rich said above, this is a problem that can be at least partially solved by the players themselves. Funny handicapping rules would be overkill and take some of the fun out of the event.

For the ubermeta, which answers you didn't have possibly mattered more than how many you were missing. There was a decent mix of short and long answers, and enough high indexes in the grid to guess where the missing answers would sort. This turned out to be fairly easy with 39 answers (93%), although we got lucky and were missing two adjacent answers in the high Rs or low Ss. We could read all of the instructions and start making out words in the second grid with 4 of the 5 meta pieces.

There's a difference between "water stuff" and "Atlantis". A story was set up about how Atlanteans were trying to flood the world and we had to stop them. But the puzzles themselves did nothing to reinforce this theme. Many of them had aquatic titles, and yes, one required people to swim. If the theme had just been "water", as in Shinteki Aquarius, that would be have been perfectly fine. But that's not the theme we had-- it was about an Atlantean conspiracy. And aside from the three skits, that didn't come through in the hunt at all for me. Contrast that with PH6, where the change in the timeline was reflected in the puzzles themselves, and where a time machine was shattered and the pieces were scattered through time and embedded in the puzzles. Or PH8, where each meta completed a step of the heist.

Did the level of PHA's theming hurt my enjoyment? Not at all. Please understand that I had a great time, and in fact thought a lot of the puzzles were quite good. But it's possible to enjoy something and also critique it. A hunt doesn't have to be strongly themed-- if PHA had set itself up with no story and just water-themed puzzle titles, that would have been fine. But it set different expectations, and then didn't deliver.

That's my beef with the length as well. I had a lot of fun during the hunt, but I expected that fun to last longer. Had I gone into it knowing it would be 15 hours for our team, I would have been in the proper mindset to enjoy its completion rather than being disappointed that there wasn't more. My juices are just starting to flow around midnight. Pushing through those wee hours is part of the hunt experience that I enjoy. I expected to be there overnight, and that expectation wasn't met.

The underlying theme here is managing expectations. You can run a James Bond hunt with funky gadgets, spying on other teams, crazy stunts, and discovering that three of the puzzles are double agents designed just to slow you down... or you can just run one where the puzzle titles are all character names and movie titles. Either way is fine, but if it seems like it'll be the first way and then it turns out to be the second, some people will be disappointed.

I disagree with Rich on the Common Thread puzzle, which I think was perfectly reasonable. Yes, if you don't have the AHA you're sunk. But it's completely getable, and I think AHA puzzles are fair game. In fact, hunts are really the only fora for those kinds of puzzles.

I still believe it's possible to follow the old school philosophy of making the hunt challenging for the leaders but easier for those trailing behind, without resorting to shoveling out hints-- which, I agree, has its own problems.

I think Scott raises a good point, that part of the reason the top teams are getting better and better is because all hunts have a certain sameness to them. Oh, look-- semaphore. A bunch of 2x3 grids? Braille! Transinsertions! Transdeletions! 5 bits of binary data? That's the 1-26 range, baby! Hey, that grid's 26x26. True innovation in puzzle design is rare at this point, and the top teams simply recognize the commonalities more quickly now thanks to their collective experience. And while there's undeniable pleasure in solving a familiar form that's well-executed (In the Axeman's Basket, for example), the real delights are the rare puzzles where even grizzled veterans must blaze new trails. Problem is, true innovation is hard. Word search variants are easy.

I also question the implied virtue of an event where many teams finish. How important is that, really? It's fun to finish, sure, but is having an additional 10 teams finish by the end worth the cost of having 4 or 5 teams finish halfway through and having a much shorter experience? Would the overall happiness level of all players combined been higher had only 4 or 5 teams finished somewhere between noon and 6 PM on Sunday, but everyone was engaged and happily puzzling for 25-30 hours?

As a player, I don't want to see the event simplified so that more teams finish. I want to see teams' skill improve so that more of them finish. I want people to run faster so they break more records, rather than have the track shortened to achieve the same goal. As for top teams breaking themselves up, I suppose we could play with one hand tied behind our backs, too. What's the point?

In the end, the hunt is about having fun. And if the top teams have less fun because the event doesn't challenge them enough, that too is a problem worth examining.

BTW, I went back and looked at PH2 last night, since it's the only one online (which I still find depressing, but what can you do).

No disrespect to the very talented folks who created puzzles for that hunt (some who did so for this one), but this is a great example of the remarkable growth in both hunt creators and hunters. Read those puzzles: anagrams, ah-ha's, grinds galore. Look at the stats and remember the timeline: the teams that did finish didn't finish all that early (If early at all? I don't really remember.), and there were only 26(25?) non-metas. Puzzles that seemed (to me, anyway) very clever or challenging then (like the golf puzzle) have begun to seem standard.

(BTW, BLP deserves _huge_ credit for deliberately and explicitly removing the dependence on anagramming, which is a crutch for every puzzle designer. That was great, both as a solver and as an admirer of good puzzle design.)

Oh, and Rich, we solved the final meta with 38/42 solved, and I'm pretty sure we could have done it with somewhat fewer (i.e. we had 38 solved by the time we put together the final meta). The final meta was like much of this hunt: a well-executed variation on a standard theme. In general, it's always about the puzzles you haven't solved, and sometimes it works for you, sometimes it doesn't.

Regardless of personal opinions on puzzle design, I'd bet that if the final meta had not been straightforward and had required some mental leaps to figure out what was going on, the last few teams would not have finished - and that's not necessary.

This month's pandamagazine.com issue has the same theme as PHA!

== = = = = = == = = = == = = = = = = == = = = =
Issue #7 Is Available!
There's a storm of evil brewing, and it''s up to Captain Clear Skies and his co-horts to end this wicked weather.
== = = = = = == = = = == = = = = = = == = = = =

I for one loved finishing early, as pushing through the night is not my primary motivation for doing the hunt. The *only* all-nighters I've ever pulled were in previous puzzle hunts, and other members of CGT can attest that I get cranky at 4 in the morning. My wife also appreciated that I got home at a quasi-reasonable hour.

However, that said, my primary argument in the past for making the hunt easier is that more people would see the cool stuff. It's frustrating to put lots of effort into something (like our closing skit in PH6) that only four or five teams get to see. Similarly, if we build a cool location-based puzzle, we'd like teams to see it.

BLP solved these problems (mostly) elegantly by having the final skit at the wrap-up, by dragging most of the location-based stuff to the wrap-up, and showing good videos of some of the other stuff. Those innovations, to me, were the best things about this hunt. The fact that the puzzles were almost entirely top-notch was also much appreciated.

So as long as we find ways to make sure people get to see the "good stuff", I don't know that it matters much how many teams finish. As long as every team sees the majority of the puzzles and sees the theming elements and the cool location stuff, I don't know that finishing is that important.

BTW, BLP deserves _huge_ credit for deliberately and explicitly removing the dependence on anagramming, which is a crutch for every puzzle designer. That was great, both as a solver and as an admirer of good puzzle design.

Everyday Heroes did that in PuzzleHunt 9 also.

> I suppose we could play with one hand tied behind our backs, too. What's the point?

The point is that you'd get exactly what you're complaining that you don't get now. You'd get 24ish hours worth of puzzle solving and be challenged to finish the hunt. If you can convince all the other top tier teams to do the same you'll have a lot more teams in competition.

The top tier teams are mostly hardcore solvers. The second and third tier teams are not as hardcore, or at least aren't teams as full of the hardcore.

I'm not saying you should HAVE to divide your team. You can do whatever you want to, but dividing your team seems like the easiest solution to your complaints. :)

It seems to me like what is happening now is akin to having a local volleyball league with a few Pac 10 teams playing, a few good local teams, some people who have liked playing volleyball in the past. :)

I didn't mind finishing early. The only thing I dislike about long puzzle hunts is the no-win question of when and how long to sleep: either you miss a bunch of fun puzzles in the present or you have less fun due to fatigue in the future. As it stands, long hunts inherently reward the people who are most willing to run on little to no sleep, and I dispute the notion that wearing one's body down and getting snippy with one's teammates is part of being a "hardcore" puzzle nerd. I would be thrilled if some GC found a good way to introduce, say, designated 4-hour sleep periods during which no new puzzles would be released, but I've never thought of a way to implement this in which the most hardcore teams wouldn't just keep solving existing puzzles during that time. Anyway, I was happy to finish at 1 AM (or whenever it was) and know that I would be able to sleep in a real bed for 8 hours without missing a bunch of amazing puzzles or the end of the hunt. In fact, I wish that we'd finished 10 minutes earlier still, so we could have beaten you guys. ;)

Couple comments on the final meta:

> I (one guy in a room by himself with an Excel
> spreadsheet) test-solved the ubermeta with only
> 37 of 42 answers (88%). We figured the likes of
> CGT could do even better.

Yes, I agree that it was possible to solve with more than 4 incomplete - Everyday Heroes came very, very close with 6 incomplete (BOTTOM OF SEVENTEEN GARAGE IN ... DIAMOND). If we'd solved one more, we probably would have gotten it, especially if we'd been allowed to go to 17 and actually look around to see the crates. My point was more to the fact that no one actually *did*, including top teams that were all very capable. There's a big difference between a test solve solo in a no-pressure situation and one in which you're short of sleep, brain-fried, and under the gun due to other competitors and/or the looming deadline of the end of the Hunt, so I'm not surprised you got different result in beta testing.

> Oh, and Rich, we solved the final meta with
> 38/42 solved, and I'm pretty sure we could have
> done it with somewhat fewer (i.e. we had 38
> solved by the time we put together the final
>meta).

Yep, I agree that it could be done (barely). I'm just saying it was just complex enough that no one could, practically. 38 compeleted puzzles is the fewest completed puzzles that anyone who finished had.

> The final meta was like much of this hunt: a
> well-executed variation on a standard theme.

Here is where I have to disagree with you Scott. I think it was a poorly executed puzzle. Most of the "good" Puzzlehunts, in my experience, have relied on completed sub-metas, not on completed base-level puzzles. This results in far less potential for random screwing because the puzzles your team can't solve (despite having all sub-metas solved) prevents you from spelling "CRATE".

Oh, and Peter, I personally believe you're just wrong - puzzles that start with an unclued AHA have no place in any sort of event that purports to be a competition and not merely an intellectual curiosity. I am fully aligned with the Staggering Geniuses on this one - you won't find any of these type of puzzles without instructions or built in clues in either PH8 or PH9. I think this was a step backward for PH as an event that I hope Scrubbers recognize and return to the "righteous path" for PH11. I'm pretty adamant about this, but I'm happy to agree to disagree. :)

Oh, and lest anyone is confused: I really liked the event overall, most of the puzzles were great/brilliant, and it was the best managed event I've seen to date. I just like to dwell on the things that I found frustrating/annoying. :)

>> BTW, BLP deserves _huge_ credit for deliberately and explicitly removing the dependence on anagramming, which is a crutch for every puzzle designer. That was great, both as a solver and as an admirer of good puzzle design.

> Everyday Heroes did that in PuzzleHunt 9 also.

Actually, so did Staggering Geniuses in PH8.

Rich: I'd agree that unclued AHA puzzles are to be avoided. Common Thread, however, was not unclued. The flavortext had clues to the AHA. The puzzle also told you that all the pictures had something in common. Once the puzzle gives you that, it becomes an exercise in lateral thinking and brainstorming, and I think that's completely fair. If the puzzle had just been the grid of pictures with no instructions or flavortext, my opinion would be different.

"righteous path"? my zealot radar is beeping like crazy.

rich, your team didn't know it was the bottom of the garage until after the deadline had passed. it's unfortunate that you were too late, but that isn't the fault of the puzzle. further, you weren't told not to go. when you said "top of seventeen garage" you were told that not all of your information was correct.

I'm not sure where you draw the line at "aha" puzzles. To me, finding a common link in a grid of images doesn't seem much different than recognizing Roman numerals hidden in text, interpreting bits of color as resistor values, etc.

I see at least a couple of puzzles from PH8 that required an initial bit of insight to get going - Treasure Island (homophone and synonym literary titles), and Craps (dice cryptogram). The initial leaps required in some of the PHA puzzles don't seem significantly different.

Actually, Dana, we did figure out that it was the bottom of the garage before the deadline passed, but the fact that we didn't solve it before the deadline is a completely spurious argument. I'm not complaining that we didn't finish (which I attribute to team-internal issues), I'm complaining about the structure of the final meta. It is objectively true that no one was able to complete the hunt with more than 4 incomplete base puzzles, even if they had solved all the sub-metas. That, to me, is poor design of the final meta. Also, the inference of the statement "call us before you go anywhere" is that we should not go somewhere without permission, but it sounds like we read that wrong.

And I specifically put "righteous path" in quotes because it is something that I feel strongly about, but wanted it to be clear that I was using hyperbole, but apparently didn't come across. Sorry. :)

rich, i appreciate your clarifications. as long as we all keep in mind that we are discussing opinions i think we'll do fine.

I really liked the PH9 system in which the meta-meta didn't require anything besides the meta answers to be completed. However, there are several good reasons to use a PH7/PH8-style meta-meta that requires a few additional puzzles be solved beyond the requisite metas. The most compelling such argument, IMHO, is the following.

Imagine a team that's just solved the final meta and is now working on the meta-meta. It's unlikely the meta-meta will be amenable to parallel solving by all 12 people on the team. So, everyone not working on the meta-meta must sit tensely by, feeling powerless to control the fate of their team at a particularly exciting and crucial moment.

On the other hand, if you design the meta-meta to need, or at least be substantially aided by, solving a couple more puzzles, then all those other people have something productive to do in those waning moments of the hunt. Everyone can feel like he or she is contributing to the final outcome for the team up until the very end.

If you have 100% of the sub-metas, solving the final meta should be relatively straightforward (i.e. not much down time), and if you don't have 100% of the sub-metas, other people can be working on those or the puzzles that lead into those. Most of the work to solve the final meta should be done (if not completely solved) by the time all the sub-metas are complete. In other words, I don't buy your argument, Jay. :)

rich, that depends on how the meta meta works. i have definitely been in events where at the end there is one person crunching on the meta meta and there's not really anything productive for anyone else to do. i remember being that one person.

Another aspect of having answers feed into the meta-meta is that puzzles which feed into solved metas don't become "worthless" - there is still value to solving them.

(For PH8, this was a particularly important design point, since the mapping of puzzles to metas was done solely based on the answers. You couldn't know for sure which puzzle was going to feed to which meta until after you'd solved it.)

Here's my perspective from a team that would probably fall in the middle tier of puzzle hunters. this was the 2nd puzzle hunt my team "where's lunch" has taken part in, and this year, we placed 24th which was much better than last year (and our first "first solve" ever!) ... although we may have been the team to win the "most emails sent to PH team" award. =)

We really enjoyed puzzlehunt (both 9 and A), and we DO come to win (even though we know its not very likely). This year our team consisted of msft people (1 PM, 2 Testers, and some Devs), a 19 year old CS major from UW, an actuary, an elementary school librarian, and my brother, who worked remotely from DC (ex software dev in law school). So as you can see, a pretty interesting mix of people.

I would say there are a couple of problems with our team. Mostly, we aren't "hardcore" enough. Only about half the team was ready to go at the start of puzzlehunt, the rest of the team arrived around 1pm. Next, during the night, only 3-4 of us stayed up for most of the puzzling, and by noon, only 9-10 of us were left working on puzzles ... so i would say our effective team size was probably 7-8 people.

another problem with our team is that we aren't very "cultured". We are terrible when it comes to crossword puzzles (and the like) and puzzles involving music/movies/etc from before 1995 (avg age of team is about 23 years old!).

But despite all of this, we have a really fun time at puzzlehunt, and we always look forward to the next one. Do we wish we could take part in all of the puzzles? of course! but we also realize that unless the puzzles are difficult, theres no fun in it. This years puzzlehunt was probably the right level of difficulty for a team like ours, but that said, i agree that puzzlehunt should be a bit more difficult, why? I really think our team has tons of room to grow (we are actually trying to do puzzles during the "off season"), and we want to continually be challenged to do better. The onus should be on the team to get better, and not PH organizers to "soften" the puzzles to accommodate the teams.

anyways enough rambling, my whole point is that I think it is OK if PH caters to the top 10 (or so) teams, because it gives them more enjoyment, and it gives us lesser teams something to aim for.