How can I get better at solving puzzles?

[Deception]

I want to get better at these things, and I'm not sure how. I'd like to think that I'm quite intelligent given I don't have a college degree yet, but I must confess that most of these puzzles trump me forever.

Things I specifically am confused on:

Cryptography: I know some basic cryptography, but there are times when you guys decrypt a series of letters or numbers in a non-intuitive fashion from my point of view. Either there is some thought process you have, or is there perhaps a list of common/uncommon cryptographic sequences?

Well... the right track: If it is clear what the puzzle is asking me to do, then I do pretty well; but a lot of these puzzles are very vague at first glance. Sometimes just a list (scratch that, often times a list), sometimes there is even no explicit question! Are there any 'tricks' you use to help get along?

I really don't know where else to turn because I'm just not sure what to do, and seeing as this site is easily the elite of puzzle-solvers, I figured I'd post here.

[jesternl]

Some can be very hard, my best advice is to read over old puzzles and try and follow the steps.
Most people here will be very willing to explain the solution, as long as it's not a chestnut.

[Zag]

I do get Deception's point. It isn't that he doesn't understand the solution once it is explained, it's that he feels he never would have thought of it on his own. I know how he feels -- I never get the serial killers or the completely wide-open treasure hunt puzzles. I remember my first experience with this feeling was this puzzle, which I saw as a teen:

What do these four words have in common? Sate, Shout, Stew, Thorn

I know that I puzzled over it, trying an assortment of different ideas, before giving up. Once I saw the answer, I hit myself for not thinking of THAT, but then I realized that there must be some technique that the people who can solve those use to find the right association/trick/whatever. In years of doing puzzles, I'm convinced that there is, and that I still suck at it. Sometimes you get them because you've seen a similar association before, or you just stumble on the right thing, or whatever; it just clicks. (Anagrams of East, South, West, North.)

Deception, I would suggest doing a lot of Cryptic Crossword puzzles. They often require the same sort of lateral thinking that is needed for puzzles of this type.

But there are a few tricks -- the first one is to assume that it is solvable. That's especially helpful in this chestnut:

I have a perfect sphere, and I drill a hole straight through the center of it such that the length of the "tunnel" that is left is six inches. What is the volume of the remaining part of the sphere?

It seems that this couldn't possibly be solvable without knowing the size of the sphere or the size of the hole. But it is -- it turns out that as you increase the radius of the original sphere, the part left surrounding the tunnel gets thinner and thinner in order to remain only six inches long, and the volume remains constant. Proving THAT is a challenge, but you don't have to to solve the puzzle. As soon as you simply assume that it must be solvable, you can make some simplifying assumptions and easily come to an answer. (Just assume a 6-inch diameter sphere and a hole of approaching-zero diameter. If those values don't matter, then these are legitimate choices.)

[Quailman]

[Vagrant]

[Zahariel]

Finding the right path to solve a puzzle (particularly puzzle-hunt style puzzles, which seems to be what you're asking about) is one of the hardest parts of solving these puzzles. Conversely, ensuring that a puzzle that has a solve path at all is one of the hardest parts of writing a puzzle hunt. I recently participated in SEAHOP 2011, which unfortunately contained quite a few excellent example of puzzles that had essentially no solve path. (The metameta, in particular, was just criminally bad; no one solved it during the hunt.)

Vagrant is right that there isn't a trick to it really, you just have to do a lot of puzzles and practice. Most metro areas will have a few smaller-scale puzzle hunts each year; these vary in quality but usually you can manage to find a few good puzzles in any hunt. Join an MIT Hunt team as a remote member; the MIT hunt is the biggest one anywhere, and they have no limit on team size, and specifically cater to teams that are mostly remote (as long as you have a few onsite members to do runarounds). I recommend Carnegie Mellon's team, the Manic Sages, as I happen to know they always have extremely good remote-solver support. Eventually you will just start to see patterns in random collections of data. This is a sign that you are going insane (c.f. John Nash), but it's great for solving puzzles.

[Scurra]

Of course suggesting that someone starts with the MIT Hunt is a bit like saying "you want to learn French? Here, try this copy of Proust."

Having said that, it is probably the cream of the crop in terms of puzzle quality and learning potential (and next year will be interesting because Codex have only been waiting and planning for about ten years for the chance to write a Hunt, so I suspect that it will have had excessive testing... )
_________________
_{still Quiz Olympiad champion. Must get a life.
New definitions: COFFEE - someone who is coughed upon}

[Hitchhiker]

If you're not up to the MIT Hunt, try the archives of the much-easier Harvard hunt. Which is not to imply that Harvard is easier than MIT

Scurra: Very much looking forward to MITMH 2012!

[LordKinbote]

I'm also looking forward to the Mystery Hunt, since, well, I'm on Codex.

Hope you guys like Pokemon. That's all I'm going to say.

So if we're talking about puzzles that don't provide instructions (like most but not all Hunt puzzles) there are several ways that you can improve.

1. One can become better at puzzles the same way that one gets to Carnegie Hall: practice, practice, practice! There are many, many outlets for this type of puzzle online...several hunts have already been named, but I recommend P&A Magazine and the Puzzle Boat, Mark Halpin's Labor Day Extravaganzas, CISRA and MUMS competitions, etc. Exposure is the only way to learn new tricks, and there while people are constantly coming up with new ways to confuse puzzlers, more often than not, puzzles are just an old way of solving with a fresh coat of paint.

2. I liked the advice about assuming that the puzzle is solvable, and by taking a leap of faith that the puzzle author intended for the puzzle to eventually fall. I'm going to butcher a quote, but I heard recently that a puzzle is a game between constructor and solver that the constructor hopes to lose. If the constructor is creating a puzzle that he hopes will never be solved, he's not doing his job and he's not providing enough information.

So, assuming that the constructor is doing his job, what information other than the puzzle text gives you clues as to how to solve it? The puzzle? Flavortext? Layout? Presentation? Any one of these might seem innocuous but might be the very key to a solve.

3. One of the best ways to get better at solving is to get better at constructing. I'm definitely a better puzzle solver at the end of this year than I was at the beginning after several months of construction under my belt (working on my fourth, fifth, and sixth MH puzzle right now). Constructing a puzzle helps you understand what kind of constraints are placed upon the construction, and you are much more likely to recognize patterns in other puzzles if you've created similar patterns for your own. I can't tell you how many puzzles I've solved this year by asking the question "Well, if I were the author, how would I have written this?"

4. Luck. Sometimes, you just need to be the right pair of eyes. What may seem confounding to you might seem trivially easy to the next person who sees it, not because they are a better puzzler, but because they have different life experiences. A couple of years ago, there was a Mystery Hunt puzzle that consisted of nothing but sets of blanks to be filled in with words, and no clues as to what words you should fill in. This puzzle confounded us until I got desperate and picked one of the sets with seven words and tried filling in the names of the seven sisters colleges. They actually fit, and spelled ALUMNAE downwards. I am certainly not one of the best solvers on our team, but I *was* one of the few people who did not go to MIT...I went to Vassar, one of the seven sisters. I sincerely doubt any other team broke into the puzzle the same way I did (some of the sets were much easier, like the seven dwarves or the names of playing cards from TWO to ACE) but my life experiences helped me find an opening.

That said, you can get luck on your side by having as broad a knowledge base as possible. A lot of the greatest puzzle solvers are a mile wide and an inch deep, as they say. They know a little about a lot of things; most professional Scrabble players couldn't define a tenth of the words they know, for example.

One thing there is not enough of out there on the interwebs is puzzle decomposition, where a skilled solver picks a puzzle and shows their train of thought as they solve it. We should do something like that here.

[Scurra]

[Zag]

Sometimes just putting it down for a while is enough.

I know that I when I'm solving a Cryptic Crossword, I typically get less than 1/3 of the clue in my first try, and no more than half even when I go through them one by one, considering any known letters, etc. Invariably, if I put it down for a week and pick it back up, there will be two or three that just jump out at me, and I wonder why they were ever hard.

[Hitchhiker]

Agreed. The difference can be so dramatic after a night of sleeping on it, that you might think, "How did I get smarter since yesterday?" Or "this can't be the same puzzle as before -- but the answers are in my handwriting..."

[Scurra]

Then again, Rupert Sheldrake argues that "morphic resonance" explains this. People who haven't looked at the previous day's crossword and then tackle it allegedly solve it faster than those who did it on the day... The principle being that somehow the solution is "out there" for people to sense.
_________________
_{still Quiz Olympiad champion. Must get a life.
New definitions: COFFEE - someone who is coughed upon}

[Zag]

I'm prepared to attribute that to bad testing (like most parapsychology experiments). They probably didn't adequately filter out the people who had solved it the previous day. In any case, the cases I noticed were old puzzles already when I first saw them.

[Suspence]

[Chuck]

How to Solve It by George Polya

[Zag]

[groza528]

So I'd like to expand on the "assume the puzzle is solvable" idea with a feature that also frequently occurs in other various forms of media: The Law of the Conservation of Detail. Little things that seem like they don't mean anything may be the key to the next step. "Boy, this crossword grid sure has a LOT of Xs in it." Focus on the Xs. The writer wouldn't have taken the time to put them in if they weren't important.
The other thing I recommend is to count, count, count. Sometimes something clicks because of a connection to a number. "There are exactly 26 items on this list!" Know what else there are 26 of? Look for a one-to-one correspondence. "There are the same number of blanks down at the bottom as there are sentences in the paragraph above it." Might not be a coincidence. One of the metapuzzles I saw in the last few months was broken because we realized that the answer to the puzzle was almost always exactly one letter longer than a specific other piece of information associated with the puzzle.

[ralphmerridew]

Sometimes you'll notice that most of the answers you've found are in a particular order (alphabetical is very common). Consider the possibility that all the answers are supposed to be in that order and the ones that don't fit are wrong.

[LordKinbote]

[Scurra]

One of the other basic features is "indexing". Given that puzzles generally generate a single solution of some kind, a standard way to do it is to take a letter from each separate answer, often represented by the number of the letter within the word. Sometimes this is given directly by the puzzle, but more often you have to figure out what is going on.

Probably the most common is to use initial letters (i.e. an index number of one!) Reading diagonally may be a close second. These are probably the first things you should try.

But when neither of these work, that's when you start looking for other hints within the puzzle. Do those clues seem a bit oddly phrased, with superfluous words? Does each clue contain something that might be considered a number? Perhaps the introductory text or title contains a clue? (I wrote one once that was called "As Easy As...", and the required indexed letters appeared at positions 3, 1, 4, 1, 5 etc. Of course, the fact that all the clues/answers turned out to be related to Jackson 5 songs was misdirection...)
_________________
_{still Quiz Olympiad champion. Must get a life.
New definitions: COFFEE - someone who is coughed upon}

[Thok]

It might be helpful to try some basic puzzles: you can get a Penny Press or Dell puzzle book with roughly 200 puzzles for 5 dollars (it's a monthly magazine, and I'm sure somebody would recommend Games magazine instead). Alternatively, http://www.janko.at/Raetsel/index.htm has an online collection that you can try. Also, there are multiple free crossword puzzles on the web.

When you do the basic puzzles, you want to try to think up general strategies that will work on multiple versions of that puzzle. Do the easy one first to build up some pattern recognition for common themes before you go to harder problems.

One of the tricks is just building up a lot of cultural knowledge. You should recognize common bits of culture, because a puzzle could be built around baseball, or movies, or common types of cheeses. Or in the case of crossword puzzles, all three of those topics and forty more.

Also, divide and conquer. If the puzzle naturally breaks into parts, look for easy parts and use that to help with harder parts. If it has a beginning and an end, work from the beginning forward and the end backwards.

Finally, just like in any RPG, talk to everybody. Somebody might give you an idea that helps.