2010 US Puzzle Championships

[nickdeller]

...the message here to myself is to spend at least 30 seconds on each puzzle at the start and see what's what!

I learned that lesson on my very first attempt in 2000, and have been forcibly retaught it every year since!

[drsteve]

Well, with the benefit of hindsight, I wish

a) I hadn't had a stab at Count Me In

b) I'd noticed that I hadn't tried Sukoro and then done it

c) I'd remembered the odd-even number rule for Route (although I still couldn't do it quickly)

d) I'd been luckier with my guessing when I go stuck halfway through Hex Masyu

e) I hadn't missed completely that the 2 in the top middle box was obvious in the Sudoku

I think 200 would have been achievable on a good day, but no more than that.

Pretty sure that with all the time in the world I'd have been stuck on X-Agony, Pentapath, Digitile, Mirror Mirror and Hex Equation. The jury's out on Sukazu until I have another go at it with the benefit of knowing that 1444455555 is a possibility for some rows.

It's a shame in some ways that the first time we try and publicise the championship, it's the hardest one in years...

[drsteve]

...the message here to myself is to spend at least 30 seconds on each puzzle at the start and see what's what!

I learned that lesson on my very first attempt in 2000, and have been forcibly retaught it every year since!

My lesson that it's taken me five years to learn - check you've typed your answers in correctly - this was my first year with no mistyping!

[GarethMoore]

Well, with the benefit of hindsight, I wish
a) I hadn't had a stab at Count Me In

Haha, I completely agree! That was a -5 of points and 5-10 minutes of time I could have done one of the others in.

b) I'd noticed that I hadn't tried Sukoro and then done it

Again, me too! I know it sounds silly but two (well three if you count Sudoku) puzzles with almost identical names confused me into forgetting one of them. Or that's my excuse anyway. But it's doable in 10 or so minutes, perhaps 15 with extracting the answer key which I think takes about 15 minutes of the entire test time with checking.

In fact I am sure I spent about 20 minutes in total on extracting and checking answer keys - I know I started typing them in about 8:15pm and only finished a minute before 8:30, and I had already calculated some of them so in total I'm sure I spent about 10% of the total time on key extraction and then re-checking by counting inverses and so on. I spotted I'd missed the "2" region in the Corral by counting the centre region too, which would have lost me a lot of silly points if I hadn't spent time on that.

c) I'd remembered the odd-even number rule for Route (although I still couldn't do it quickly)

Well forgetting it entirely is less embarrassing than remembering it at the bottom-left (when placing the 1 etc) and forgetting it in the middle (when placing the 69 to 73, where it would have been most helpful!) which is what I did.

d) I'd been luckier with my guessing when I go stuck halfway through Hex Masyu

I should try this puzzle again. I didn't find it difficult but I did find it confusing and I kept thinking there was a very high chance I'd made a mistake somwhere! Luckily the very low number of clues convinced me I must be right in my general approach so I kept going.

e) I hadn't missed completely that the 2 in the top middle box was obvious in the Sudoku

I couldn't believe it when I solved this in 5 minutes on Sunday. Grrrrr! Another silly thing - I saw the 20 points and convinced myself it must be much harder than it really was, so I went about it with brute force (badly) rather than my usual method (which can be slow but at least is less error prone!).

I think 200 would have been achievable on a good day, but no more than that.

Well if I hadn't tried the counting one, and had got Sudoku, Criss-cross sums and Sukoro (which I could have done in the time I spent on the ones I abandoned) I could have had... a maximum of 200 points too! Snap! But that would have required a little bit of luck with my puzzle selection, and if we're allowing luck then you could fluke the solution to Digitile for another 10 points.

I think if you didn't have to type in the answer keys you could do one more puzzle too in the time allowed.

Pretty sure that with all the time in the world I'd have been stuck on X-Agony, Pentapath, Digitile, Mirror Mirror and Hex Equation.

Hex Equation I spent too long on, but it's not really that hard I don't think - once you start staring at it you realise where the two sides of the equation must go and then there's only I think three or four possible solutions for one side, and maybe the same of the other. But for whatever reason those 16-odd reasonable paths are hard to evaluate so it took me at least 10 or 15 minutes to do. Still, like last year I successfully did a puzzle motris didn't. (last year it was wolf and sheep in fences). Of course he got more than double me overall, but hey.

The jury's out on Sukazu until I have another go at it with the benefit of knowing that 1444455555 is a possibility for some rows.

If you do let us know how you break in, once you've done the left and lower edges. I figured 55555 was needed during the competition but couldn't see how to work out where it went and where it didn't. Maybe it's just a case of counting rows, columns and boxes and cross-referencing, but I couldn't be bothered to try that! Is there a better way?

It's a shame in some ways that the first time we try and publicise the championship, it's the hardest one in years...

It was my favourite one I've done, including the older ones I didn't do "live". Usually there are a few puzzles I find nightmarish just to think about and can't imagine ever doing for fun, but this year I would be happy to do any of them just for fun (perhaps excluding X-Agony, although I haven't tried it!). And the lack of standard types (mostly) surely is an advantage to those who are new, relative to the norm? Maybe.

[drsteve]

The jury's out on Sukazu until I have another go at it with the benefit of knowing that 1444455555 is a possibility for some rows.

If you do let us know how you break in, once you've done the left and lower edges. I figured 55555 was needed during the competition but couldn't see how to work out where it went and where it didn't. Maybe it's just a case of counting rows, columns and boxes and cross-referencing, but I couldn't be bothered to try that! Is there a better way?

OK, I presume that you've done enough to establish that rows 3, 7, 9,10 and columns 1, 2, 4 and 8 must be 122333444 and that there must be a 5 somewhere, as there aren't enough spaces for the 1's in the top 4 rows if everything is 1223334444.

So, either there are 5 columns and 5 rows that have five 5s in them, or 6 columns and rows.

If it's the first, then it must be column 3 & row 4 or column 5 & row 8 that doesn't have a 5, but this would put six 5s in the top right area. Hence for each of the six columns and rows, there are five 5s in them. You can fill in columns 3 and 5 and rows 4 and 8 straight away.

Let me know if you need any more after that.

It's a shame in some ways that the first time we try and publicise the championship, it's the hardest one in years...

It was my favourite one I've done, including the older ones I didn't do "live". Usually there are a few puzzles I find nightmarish just to think about and can't imagine ever doing for fun, but this year I would be happy to do any of them just for fun (perhaps excluding X-Agony, although I haven't tried it!). And the lack of standard types (mostly) surely is an advantage to those who are new, relative to the norm? Maybe.

But it might be nice to put a couple of friendlier, more familiar puzzles to attract the layman. The iterations of wordsearch and criss cross this time were a lot of fun, but replaced puzzles that were easier to get into. I think there were too many puzzles this time that the common or garden puzzler couldn't start.

As a maths teacher, I'm trying to encourage a bit of puzzling here and there, with one eye on publicising the competitions, especially the USPC. Next generation and all that... but if this was the first attempt, I can see a lot of people scoring very low scores and thinking, stuff this for a game of soldiers. If you look at the sub-100 scores that experienced puzzlers on this board are claiming, then inexperienced puzzlers may be in single figures.

I certainly don't agree with Tom Snyder that there should be a bucketload of standard puzzles in the WPC and USPC - I think it levels the playing field more if you have to develop strategies on the hoof, but would be nice if there were, say, 50 points of puzzles that are of a low level of difficulty - there ususally have been in the past - basicish Criss Cross, Word Search, Spot the Diff and other more difficult puzzles that were at least attemptable by the inexperience puzzler, such as the harder Criss Cross varients. This year seemed to have more WPC level puzzles which I hope didn't put off the amateurs.

[GarethMoore]

I will try the Sukazu later - thank you very much!

I think now you say it you're completely right about the lack of recognisable puzzles - one of the comments on my forum today was from someone who is a very keen puzzler, but she took one look at the USPC test and decided that she could never do it. And I don't suppose she'll ever look at it or any qualifying test again now! Even the wordsearch was superficially scary (of course it wasn't when you read it, but certainly at a glance it didn't even look like it was necessarily a wordsearch).

[ronaldx]

"I certainly don't agree with Tom Snyder that there should be a bucketload of standard puzzles in the WPC and USPC"

I don't agree with Tom here either! It's a Puzzle competition - which should be about solving puzzles, not following the best and quickest rote algorithm.

[ronaldx]

I'd like to clarify that now that I'm less angry (and I read what Tom Snyder wrote, he was saying he thought there should be more classics, not all classics)
Maybe it is about following the best and quickest algorithm indeed, I take that back, but it shouldn't be about following a rote algorithm the quickest!

[drsteve]

I certainly don't agree with Tom Snyder that there should be a bucketload of standard puzzles in the WPC and USPC - I think it levels the playing field more if you have to develop strategies on the hoof, but would be nice if there were, say, 50 points of puzzles that are of a low level of difficulty - there ususally have been in the past - basicish Criss Cross, Word Search, Spot the Diff and other more difficult puzzles that were at least attemptable by the inexperience puzzler, such as the harder Criss Cross varients. This year seemed to have more WPC level puzzles which I hope didn't put off the amateurs.

Tom's clarified his opinions now on his blog - it looks like I got something near the wrong end of the stick. In fact, it seems we agree with the amount of approachable puzzles in the USPC this year being too low. I'd encourage anyone who's interested to read his blog - http://motris.livejournal.com/ - his recent WSC sounds very interesting, despite not being a sudoku specialist. Hopefully these puzzles will be published somewhere sometime. I still think he must be a cyborg though...

[puzzlemad]

Presume that you've all had your email with your score on. I hadn't bothered checking my answers after the comp and now realise that I'd made a couple of careless errors in the answer submission, misreading my notation, so my pathetic score is now even worse. Ah well at least it's stopped raining. I'm going to get on with the pointing on the house!

[detuned]

Having completed all the puzzles, even more self-kicking. However I'm pleased to say that I appear to have gotten my "full" 110 points, which is a little consolation bearing in mind my general "rabbit in headlights" performance.

X-agony was an absolute brute. I'd be very interested in seeing a pleasing logical solution to this. I was fine with di-agony from last year which I thought solved beautifully. Everything else was fine without the time pressure. Although the four square puzzle was a bit of a grind. Perhaps I need to get a bit better with "snake" style puzzles as this seems to be the main work in.

I particularly liked the constant sum, can't believe I couldn't finish the hex masyu (which actually solved rather nicely,albeit I don't think there's much potential for further variety in this variant), can't believe I did the normal masyu 3 times before realising my "mistake" was nothing of the sort and can't believe I didn't get groundhog day out. The pentapath was also very easy for the points once you worked out the common rectangle-shape the path had to lie in. D'oh.

I wish I'd seen the clever trick with count me in. Seems utterly obvious now. I guess I should have thought further about how unique vertices can be specified in pentagons when the instructions were released. E.g. in previous examples you could use right-angles in triangles. And I wished I hadn't assumed that sukazu was forced to be 1-4 as that came out quite nicely too.

I'd also like to add that the sudoku in my humble opinion was a really good puzzle, up there with takeout with the highlights from my 2 and half hours. I suppose there's always next year!

[drsteve]

Although the four square puzzle was a bit of a grind. Perhaps I need to get a bit better with "snake" style puzzles as this seems to be the main work in.

I found this a bit problematic starting during the test as it took me ages to find a snake that had 14 corners in it - once I had that though, and I'm pretty sure there's not much flexibility in it, the rest fell quite quickly, despite a nagging voice in my head telling me that 2222222 was bound to be wrong.

[PuzzleScot]

Although the four square puzzle was a bit of a grind. Perhaps I need to get a bit better with "snake" style puzzles as this seems to be the main work in.

I did this in 15-20 mins in the comp. There weren't too many ways of getting that length snake in with 14 turns! Some basic placements on the skyscrapers left a small range of values in the battleships. I needed to do pretty much all of all 4 puzzles to be sure of the answer key.

[drsteve]

Although the four square puzzle was a bit of a grind. Perhaps I need to get a bit better with "snake" style puzzles as this seems to be the main work in.

I did this in 15-20 mins in the comp. There weren't too many ways of getting that length snake in with 14 turns! Some basic placements on the skyscrapers left a small range of values in the battleships. I needed to do pretty much all of all 4 puzzles to be sure of the answer key.

I do find this puzzle a little annoying due to the fact that the aim is to find the numbers that make the puzzle solvable, rather than to solve the puzzle itself. The more confident puzzler can get the numbers whilst still leaving two grids at least mostly undone. Of course the less confident puzzler, i.e. most people, still solve it

By the way, am I right in thinking that the rows for the battleships/skyscrapers part MUST be 122333 in some combination? Not sure if that helps in solving it, but maybe...

[drsteve]

Hurrah! Finally did that swearwordy Sudoku!

16 minutes, with the advantage of knowing where to at least start from last time.

So, as there are Sudoku experts out there - is there a good resource for advanced Sudoku strategy out there? I know part of my trouble is just missing things (like, occasionally, having 8 numbers filled in for a row) but are there any deep solving secrets known only to those who have passed some sacred ritual? Or should I just practise lots?

[ronaldx]

In this Sudoku, as I was doing it under the timed conditions, I noticed that each filled in number "pointed" you to the next one, i.e. there was a followable route around the puzzle. It was kind of clear that each new number restricted your options in one particular place.
Although I wasn't always very efficient at identifying what it was pointing at and ultimately resorted to guess and check.

Does that make sense to anyone? I might take another look.

[PuzzleScot]

By the way, am I right in thinking that the rows for the battleships/skyscrapers part MUST be 122333 in some combination?

No reason for not have a 4 in there is there? You must of course have exactly one 1.

[ronaldx]

So (in theory) it could be 122333 or 122234?

[drsteve]

So (in theory) it could be 122333 or 122234?

But it can't be 122225 because that stops the Battleship from having a solution

[detuned]

Hurrah! Finally did that swearwordy Sudoku!

16 minutes, with the advantage of knowing where to at least start from last time.

So, as there are Sudoku experts out there - is there a good resource for advanced Sudoku strategy out there? I know part of my trouble is just missing things (like, occasionally, having 8 numbers filled in for a row) but are there any deep solving secrets known only to those who have passed some sacred ritual? Or should I just practise lots?

The best place for that sort of thing is scanraid (I refuse to acknowledge the silly name change as it's clearly not a wiki). Although my opinion is that its a bit dry and technical - it's one thing having a strategy explained to you, and quite another to then execute in a non-safe environment . A lot of speed-solving is down to experience - I found that I developed a good "eye" for sudoku and can quickly home into the bits of a puzzle that need to be homed into to progress. The other thing that I see a lot of is bad notation. If you go even slightly over the top with it, it can slow you down incredibly. I find that it clouds over my "eye".

99% of sudoku you'll see are covered by strategies 1-7. Sometimes "super-fiendish" labelled puzzles will use 8-11, but that's pretty rare, esepcially so in competition. If anything else is needed, to be honest you are better off guessing - or bifurcating as it's known in the lingo. This USPC puzzle safely fell into the 99%, although there were a fair amount of pointing pairs going on, which sometimes aren't obvious to see. What made it a great puzzle is for a large majority of it there was only ever one place you could fill in a digit - as ronaldx says as you filled in one digit you had to see what implications that had for other digits.

[SimonAnthony]

Have now been able to try Saturday's test myself. Found it tough but then so, it seems did everybody. Re Tom's X-Agony experience, bizarrely the example took me longer to solve than the actual puzzle!! I made a decent start using the 100 and 63s and then it just fell out - only needed a small tweak on the right hand side from my initial 'guess' My favourite puzzle of the test was the Suzaku, which seemed to bring together lots of other puzzle types into a new (for me, anyway) and interesting puzzle. I found the Sukoro very tough for some reasonand probably left easy points on the table by not even attempting Mirror Mirror, Pentapath or Rectangle Maze - won't know until I get a chance to try them on a 3 hour plane trip tonight. Biggest fail was completely bombing on the Hex Maysu where for some reason I stalled horribly after doing a bit of the top. Another one for the plane...

Simon

[drsteve]

Have now been able to try Saturday's test myself. Found it tough but then so, it seems did everybody. Re Tom's X-Agony experience, bizarrely the example took me longer to solve than the actual puzzle!! I made a decent start using the 100 and 63s and then it just fell out - only needed a small tweak on the right hand side from my initial 'guess' My favourite puzzle of the test was the Suzaku, which seemed to bring together lots of other puzzle types into a new (for me, anyway) and interesting puzzle. I found the Sukoro very tough for some reasonand probably left easy points on the table by not even attempting Mirror Mirror, Pentapath or Rectangle Maze - won't know until I get a chance to try them on a 3 hour plane trip tonight. Biggest fail was completely bombing on the Hex Maysu where for some reason I stalled horribly after doing a bit of the top. Another one for the plane...

Simon

What was your score, Simon?

[ronaldx]

I spotted I'd missed the "2" region in the Corral by counting the centre region too, which would have lost me a lot of silly points if I hadn't spent time on that.

I missed that 2 region as well! Although... I also split the 5 region into 3 and 1 - so luckily my answer key mistake matched my puzzling mistake...!

[SimonAnthony]

BAH HUMBUG. I thought I should check my answers if I was going to announce a score and I've made a mistake on the Sukuro somehow so that's -25 , which I think makes me 180. In case it's interesting I did:

Battleships
Sudoku
Islands
Groundhog Day
Maysu
Takeout
Corral
Sukoro (except I've got it wrong)
Criss cross Sums
Sukazu
X Agony

I wasted time on Hex Maysu, Four Square (which I needed probably ten more minutes to get as I was on the right path); Constant Sum (where it emerges I didn't understand the instructions on what should have been an easy 20 points); Route (where I got in a right mess).

[drsteve]

Congrats, Simon. That looks like the highest claimed. So far...