As usual, I wish to start by thanking everyone who has contributed to the Singapore Puzzle Hunt 2019, namely Celestine (who wrote many of the hunt puzzles as well as the invaluable setting up the new hunt portal), our many test-solvers, and of course, all the participants, without which the hunt would not be possible. We always appreciate all our fellow puzzle friends, old and new, spending a day of your time to join us in this event :) We hope that you had fun during the hunt, made new friends and enjoyed your onsite hunt experience.

Thanks also to those who had taken time to send us your feedback, be it positive or negative. Some of the points would be covered here. So it's a long post again, and the epilogue is divided into various sections for ease of reference (interspersed with some hunt photos and interesting anecdotes):
1) Bouquets
2) Reflections
3) Key Learning Points
4) Pre-hunt Event Activity
5) Hunt Theme
6) Hunt Structure
7) Online Hunt Portal
8) Ending

1) Bouquets

Beyond the everyone's support and participation, one other thing which we are truly heartened by is the enthusiam and indomidable spirit of our solvers. In spite of the challenges and mental fatigue from the hunt, there were no lack of cheerful faces and everyone was still busily plugging away at the puzzles down to the wire. Some of you would even tell me about how you willingly come back again every year for the "pain" :P And much appreciation as well to those who shared kind words in your feedback. The new teams too, who had no prior experience in puzzle hunts, gamely rose to the occasion and tried their best in an unfamiliar genre of puzzles. The feedback from new solvers was also largely good. So big kudos to one and all who took part! And it's things like this which make all the time and effort spent putting together the hunt worth it!

There were 8 new teams taking part in the hunt this year, featured here in clockwise order from top left are Umibouzu, Pajel, z0 klubbe and Puzzle Solvers. Great to see everyone having a fun time despite the steep learning curve.

9N3S 9NVIW WIS has trolled their friend (who introduced them to SGPH and is solving on Phoenix) with their team name this past two years.
Me: "You realise there's a typo in your friend's name right?"
Wei Chen: (confidently) "No. This is his name now. You can call him this."

2) Reflections

If the number of puzzles solved is used as a measurement of hunt, the results this year is surely less than ideal and disappointing for both teams and ourselves. And as organisers, we need to take responsibility for that. As an example, there were 2-3 longer/harder puzzles which teams took disproportionate time during solving, and that we could have better anticipated.

In most years of SGPH, no team manages to complete the hunt. This is certainly not an outcome by design or one we particularly liked. Like any hunt puzzle, the goal is always to write something that can be solved/completed, because a solver cannot fully enjoy and appreciate an incomplete experience. The biggest challenge with SGPH is the limited time duration available and the experience level of our solvers. As a comparison, all other online puzzle hunts span from a weekend to a week instead, and even then only the very experienced (and less than 10% of) participating teams actually complete. So teams should not feel surprised or discouraged by these results, especially those trying out puzzle hunts for their very first time. The aim of SGPH is to help give local solvers exposure to and experience in puzzle hunts, and to build a local hunt community. So beyond SGPH, it is more about taking the learning points from this experience into future hunts you participate in. And if you found the experience and insights from this year's theme helpful, then that would be a great hunt takeaway for teams and ourselves too.

And could we make the puzzles easier or shorter? A proper hunt puzzle should have an intuitive and non-trivial aha. If the aha is too obvious or straightforward, it is no longer satisfying, and the "puzzle" just degenerates to become following instructions and going through the steps - short and not productive at best, long and tedious at worst. With SGPH's aim, its puzzles should minimally be representative of this basic hunt puzzle criteria, so as to build a good foundation for solvers for other puzzle hunts out there. Hence there is always this balance of writing satisfying puzzles which meet this threshold, while being suitable for the SGPH hunt duration. As an ideal, the puzzles should be short once you get the aha.

Middle Class submits the first correct hunt answer for Jean Valjean puzzle after about 30 minutes, well done!

3) Key Learning Points

To elaborate on the estimate of the length of puzzles, it helps to explain the variables/factors in the context of some hunt observations and related key learning points below which solvers can take away from the hunt.

a) Confident solving - Almost all solvers who approached me during the hunt were able to articulate or come up with the correct next step when I prompted them with just very general questions like "Why do you think the writer included this information?", "What information have you not used yet?" and "What do you think you should do with this information you have?" However, they were mostly uncertain and either gave up on the right approach too quickly and/or had been trying out some other less likely theories instead. Having the confidence to be certain of what is needed to be done at each step when solving a puzzle is something that comes with puzzle-solving experience, since most puzzles are broadly similar in their solving steps. Eg. If you see a list of clues, you solve them and find a commonality in their answers; if you see images, you identify them and find a commonality. Accumulating more puzzle-solving experience thus helps you know the next step based on the information available. A regular solver with more experience and familiarity with hunt puzzles and types, will usually be a better solver than one with aptitude but only solves puzzle hunts infrequently. So, get more hunt practice, keep asking the questions above and use the information available to ring-fence your ideas. Don't give up too early if it is likely you are doing the right thing (eg. identification for Daley Cheng, extraction for Larry Lawrence, decoding for Kaniel Outis), and don't waste time pursuing different approaches in depth if they are not supported by the clues you have (eg. aha for Ellis Redding).

b) Team collaboration - A prevalent solving approach by teams observed at SGPH is that each team member would pick a puzzle he/she likes and work on it silently alone (or at best in pairs), trying to solve it for an hour or more, setting it aside back into the pile for someone else to try when progress is stuck, then picking another puzzle to repeat this process. This is a tried and tested strategy over the five years of SGPH, with the same consistent dismal outcome. Puzzles do not get closed-out and work by one solver is not effectively passed on to other solvers. Hunt puzzles are almost always intended to be, and more easily, solved together with others. The puzzle aha is something which solvers are more likely to miss as individuals, but easy to figure out by discussing observations and theories openly as a team. The aha is the biggest variable in the puzzle length, since solvers cannot proceed unless they get it, and sometimes may never do. So identify easy/short puzzles first and work together as a team to clear them to build momentum. If you want to solve a puzzle you like, convince the rest of the team to work together with you on that.

c) Organized solving - Another common observation was that there were plenty of messy scribbling on the hardcopy puzzles. Solvers sometimes could not make out each other's handwriting, much less explain the puzzle progress to me. In a few cases too, I actually could see the puzzle answer buried in the midst of the handwritten answers, but it just wasn't clear to solvers and a lot of solving time is wasted. Using online collaboration tools such as Google Sheets to tabulate puzzle working and clue answers would help keep information organized and easy to manipulate (eg. to correct, shift, sort/reorder), as well as allow all team members to work together on a common copy. Including in the Sheets template preset fields for observations and theories also forces solvers to ask themselves these pertinent questions, and record them down for each other's reference and discussion.

d) Checking each other's work - Sometimes, minor mistakes by one solver (eg. off-by-one errors, assumptions of fact, wrong indexing) could lead to the puzzle aha or extraction being missed. Both organized solving and team collaboration could greatly help in this. In many cases, when a new solver jumps on a puzzle, they should first start by double checking the previous solver's work. Doing so and correcting errors from the previous solver could help revive previously discarded theories and lead to a breakthrough.

One of most satisfying moments of a hunt is always the moment when we manage to watch a team figure out a puzzle answer together. Jun Han and Kee Wei from Unstoppers had been solving two different puzzles independently but were stuck on their final answer extractions. So I suggested that they try discussing amongst the team and stood behind to watch. By talking out ideas, within around 15 minutes, the team had figured out the answer to both puzzles together!

4) Pre-Hunt Event Puzzle

The pre-hunt event activity is intended to allow solvers to interact with fellow solvers from other teams, whom they would otherwise not have an opportunity to meet. This is a characteristic of the MIT Mystery Hunt, which underscores the difference in the fun elements between an interactive onsite hunt, versus an online hunt which focuses only on puzzle solving. The turn-out was good with plenty of buzz and interactions prior to hunt kick-off this year, making it closer to the hunt atmosphere of the MIT Mystery Hunt. Feedback on the Contraband! activity were almost all positive, so looked like the balance and objectives were both met this time. Although it would be another challenge again to come up with a different activity next year.

During the Contraband! event activity, Melvin from SGoats was asked to identify Cheng Wai from Kaya/d, the team sitting beside them for the past 3 years.
Me: (seeing Kaya/d has only two guys) "So which one of them is Cheng Wai?"
Melvin: "It's 50/50... " (then points to Ying Xiang instead)

5) Hunt Theme

The hunt theme this year was an idea which had actually been conceived years ago, but only now has the time become right to use it. Common with many puzzle ideas, this hunt theme was initially shelved for a more suitable hunt. The theme did not lend itself to a clear Round structure, but the decision to use a single Round structure made this theme viable this year. Besides the source material being a great Stephen King novella and classic movie, the theme of prison escape also brings together this puzzle hunt with the popular escape games genre here. Also, writing good hunts and puzzles has always been a subject I felt strongly about. The puzzle hunt community is a self-governing one, and as the community grows, we need to actively pass down and share the knowledge with newer members on what are some of these undocumented rules and conventions that define hunt puzzles, and which are consequentially helpful during puzzle solving (eg. looking out for information which are contrived or unused). Aspiring hunt puzzle writers could also be reminded through this hunt to note the rules and conventions that make good/bad hunt puzzles, and to get direct feedback before releasing puzzles.

The thematic link for the puzzles in this hunt was thus for each puzzle to highlight an undesirable trait in puzzle construction - be it not following a hunt puzzle rule or convention. The common thematic twist of the puzzles in this hunt though, is that there is actually a redemption arc for each inmate - he has understood and learnt from the issue in his construction, and the trait is instead used as an intentional and necessary part of his puzzle. I felt this thematic link across the puzzles would provide both an interesting construction challenge, as well as a way to bring awareness to show how hunt puzzle rules and conventions are commonly broken, while making these actually part of the puzzle ahas. So observant solvers who picked up on this overarching redemption theme, could then also make use of the highlighted puzzle trait in the flavortext, as a starting point to help figure out the aha.

Klarissa from Kaya/d had finished identifying the actresses in Face Man puzzle but was stuck at the next step.
Me: "The flavortext would have some clues to the commonality between the four actresses."
Klarissa: "I was thinking maybe the four actresses all have the same lover??"
Me: "That would be thematic, but pretty messed up... nope" (and then LOL with the team)

6) Round Structure

One takeaway from SGPH 2018 was that teams could potentially get road-blocked to gated puzzles in subsequent rounds by metas, and timed unlocks might not be easy to fix this given the bunching of teams. Another feedback from SGPH 2018 was that many teams were less concerned about metas, and preferred just having access to more puzzles. These led to the decision this year to have a single Round structure, and just make all the puzzles available to solvers from the start, which would also be a practical structure for a relatively small hunt. The two metas were unmarked, and required teams to figure out which puzzles are associated with which meta. This was a meta structure that was done successfully in past Puzzle Boats, and coincidentally in the recent 2019 Mystery Hunt. The only difference being that given the time constraint of the hunt, the puzzles that fed into a meta would be revealed to solvers once they solved the meta. Unfortunately, none of the teams solved enough puzzles this year to really look at metas, so this aspect of the Round structure was not explored. With the single Round structure this year though, we observed that teams had difficulty choosing the easier/shorter puzzles and ended up with twice as many partially solved puzzles by individual team members which are not closed out. So this wasted a significant amount of solving time for teams and is something we would need to address next year.

A concerned Malcolm from Never Odd or Even jokingly asked me an important question an hour or two into the hunt.
Malcolm: "Are we like the worst team?"
Me: "No... wait, have you solved a puzzle yet?"
Malcolm: "Nope."
Me: "Okay... well, technically you all still better than the no-show."

7) Online Hunt Portal

An online hunt portal was something which I had planned to introduce last year, but thought was relatively less critical given that this is a small-scale on-site hunt. And we had been using a quick hack that I put together using Google Forms and Sheets for submission and tracking of answers pretty okay so far. However, the manual emailing of softcopy puzzles and distribution of hardcopy unlocked metas and puzzles was a hassle and took time away from our interactions with solvers. It also limited the unlock structure by Round meta solves, instead of by other means like puzzle solves or timing. So that prompted more urgency to set up the portal this year. Great thanks to Celestine, who managed to put together the portal very fast and making customizations off the open-source code from the CMU Puzzlehunt. Besides being a more convenient means for solvers to access the puzzles, the hunt site with its Mystery Hunt style graphical interface also helps to enhance the hunt storyline and experience.

Kenny from Flower Power had been working dilligently at the Face Man puzzle since the start of the hunt, so I popped round to check on his progress, but ended up exchanging some facetious banter with fellow team member Junwei :)
Me: "So have you figured out the commonality amongst the four of them?"
Junwei: "They are all women."
Me: "Yes, good observation. Anything else?"
Junwei: "They are all good-looking women."
Me: "Yes, also true. Okay, maybe let's try to move beyond the superficial now..."

8) Ending

You can download this Puzzle Hunt Guide for a summary on puzzle hunts, hunt puzzles, puzzle types and how to solve them.

Check out this series of videos showing you how to solve hunt puzzles and using the SGPH2018 puzzles for illustration:
- Part 1
- Part 2
- Part 3

To round off, check out the do join the SG Puzzlers Facebook group to keep in touch. We share information on other online puzzle hunts throughout the year, and organise teams to take part together in. The best way to get better at hunt puzzle solving is to get more hunt practise. Solving hunts can take up quite a bit of time, but the fun and satisfaction of solving a puzzle together with your team members is great! So do make use of these good opportunities to gain more puzzle hunt exposure and to improve your skills by solving with and learning from other experienced hunters. We look forward to seeing everyone again at next year's SGPH!

Hunt Editor
Ong Kah Kien

Thank you all solvers! Hope you had a fun time and see you again next year!