LMI October Puzzle Test 1 - Nikoli Selection

Information and discussion on Puzzle Competitions
Post Reply
debmohanty
Posts: 129
Joined: Fri 09 Jul, 2010 12:36 pm

LMI October Puzzle Test 1 - Nikoli Selection

Post by debmohanty » Sun 03 Oct, 2010 10:42 am

Logic Masters India announces October Puzzle Test 1 — Nikoli Selection

Date : 9th and 10th October

Puzzle Types : 15 Selected puzzle types from nikoli.com — 18 main puzzles + 3 Marathon puzzles

Author : Tom "detuned" Collyer

Length : 100 minutes

IB and Submission Link : http://logicmastersindia.com/M201010P1

PuzzleScot
Site Admin
Posts: 2740
Joined: Fri 18 Jun, 2010 10:45 pm
Location: Edinburgh, Scotland

Re: LMI October Puzzle Test 1 - Nikoli Selection

Post by PuzzleScot » Sun 03 Oct, 2010 4:47 pm

Well Done Tom! I look forward to having a go at this next weekend! Not sure about the 'marathon' puzzles though :?

detuned
Posts: 1869
Joined: Mon 21 Jun, 2010 2:25 pm
Location: London, UK
Contact:

Re: LMI October Puzzle Test 1 - Nikoli Selection

Post by detuned » Wed 06 Oct, 2010 3:26 pm

Yeah, this ought to be a good one ;)

The marathon ones are there because I foresee I great difference in solving speeds for this one, and having people sat twiddling there thumbs (albeit with huge amounts of bonus points) seems a little pointless. It also acts as a nice bonus for your average solver to go through at leisure after the test is done. Feedback on the large puzzles so far has been really encouraging :)

detuned
Posts: 1869
Joined: Mon 21 Jun, 2010 2:25 pm
Location: London, UK
Contact:

Re: LMI October Puzzle Test 1 - Nikoli Selection

Post by detuned » Fri 08 Oct, 2010 9:52 pm

Just a reminder on this - I'll be very interested to hear what everyone thinks about this!

debmohanty
Posts: 129
Joined: Fri 09 Jul, 2010 12:36 pm

Re: LMI October Puzzle Test 1 - Nikoli Selection

Post by debmohanty » Sat 09 Oct, 2010 1:04 am

The test has started.

Maxelkat
Posts: 45
Joined: Thu 19 Aug, 2010 5:55 am
Location: Coventry UK

Re: LMI October Puzzle Test 1 - Nikoli Selection

Post by Maxelkat » Sat 09 Oct, 2010 2:17 pm

I had said that I wasn't going to try any of these competitions against the clock, as I'm just not at that level yet. So with that in mind I printed it off and sat watching TV with my daughter, but after I solved a few I thought I might as well submit my answers. I solved 9 of the main 18, including one of the 100-pointers, and was very pleased with my overall score of 455.
Great test Tom, really great fun!

Elliott

detuned
Posts: 1869
Joined: Mon 21 Jun, 2010 2:25 pm
Location: London, UK
Contact:

Re: LMI October Puzzle Test 1 - Nikoli Selection

Post by detuned » Sat 09 Oct, 2010 3:28 pm

Naturally I'm a little disappointed with the way things have turned out, but I'm currently trying to fix the marathon heyawake. The nurikabe error was entirely stupid, and I guess I just can't make numberlink!

Glad everything else was fun though :)

detuned
Posts: 1869
Joined: Mon 21 Jun, 2010 2:25 pm
Location: London, UK
Contact:

Re: LMI October Puzzle Test 1 - Nikoli Selection

Post by detuned » Sat 09 Oct, 2010 4:59 pm

So I'm in a very conservative mood - I'm not going to post this at LMI until it's been confirmed that this is indeed error free - but here is the heyawake and the answer. I believe this is good - although I thought that about the first one too :roll:


Puzzle
Solution

drsteve
Posts: 727
Joined: Sun 27 Jun, 2010 7:23 am

Re: LMI October Puzzle Test 1 - Nikoli Selection

Post by drsteve » Sun 10 Oct, 2010 4:54 pm

Great test, Tom. Lots of good practise there.

685 for me - spent faaarrr too long not getting one particular puzzle out (no spoilers here until it's over) and didn't attempt three of them - no prize for guessing one of them :D

Thanks again

nickdeller
Posts: 600
Joined: Tue 29 Jun, 2010 11:41 am

Re: LMI October Puzzle Test 1 - Nikoli Selection

Post by nickdeller » Sun 10 Oct, 2010 7:42 pm

An interesting 465 here. I'm particularly upset to discover that there's absolutely nothing wrong with one particular puzzle which I fouled up during the test and then fouled up again afterwards in two entertainingly different ways! Third time's the charm, apparently... :oops: Nice test generally, thanks Tom and Deb!

GarethMoore
Posts: 1209
Joined: Thu 24 Jun, 2010 9:27 pm
Contact:

Re: LMI October Puzzle Test 1 - Nikoli Selection

Post by GarethMoore » Sun 10 Oct, 2010 8:37 pm

890, which would have been 920 if I had counted the number of bridges on the Hashi puzzle correctly (my puzzle's correct, but my answer key is out by one).

Successfully finished all of the main puzzles still in the test, and somewhat half-heartedly tried one of the marathons because I figured I wouldn't finish it in the 15 or so minutes left.

PuzzleScot
Site Admin
Posts: 2740
Joined: Fri 18 Jun, 2010 10:45 pm
Location: Edinburgh, Scotland

Re: LMI October Puzzle Test 1 - Nikoli Selection

Post by PuzzleScot » Sun 10 Oct, 2010 10:55 pm

Dismal 370 here. About 200 in the first 30 minutes. Might have helped a little to read the instruction file before starting!
Just one counting error, one small puzzle I wasted over 20 minutes on (still without an answer), and 20 minutes on an aborted mega-masyu.
Off to bed to finish them off. Nice puzzles Tom, thanks!

debmohanty
Posts: 129
Joined: Fri 09 Jul, 2010 12:36 pm

Re: LMI October Puzzle Test 1 - Nikoli Selection

Post by debmohanty » Mon 11 Oct, 2010 5:32 am

Good score Gareth (even with loss of 30 + 24 bonus points)
Thank you everyone for participating (around 12 players from this forum, if I can count correctly)

david mcneill
Posts: 412
Joined: Mon 21 Jun, 2010 2:02 pm

Re: LMI October Puzzle Test 1 - Nikoli Selection

Post by david mcneill » Mon 11 Oct, 2010 8:04 am

Sat up late last night to have a go at Tom's puzzles. Very late as it turned out. I'm not a Nikoli subscriber and my 555 is probably a fair indication of where I'm at with these puzzle types. I also solved the bigger Yajilin (a thing of beauty) within the time but forgot to submit my answer!!! Incidentally, I found the first Yajilin much harder despite its smaller size. Some of the puzzle types I have solved before on the free part of the Nikoli website, but I didn't have a good way of marking them on paper. Akari was a good example. I wasted huge chunks of my time on the Slitherlink puzzles - only solving the first one. I thought the Heyawake was very hard, but the LITS seemed OK. Both of these were solved after the 100 minutes was up.

I then decided to look at the marathon puzzles and I thought these were brilliant. I did the Kakuro and most of the Masyu before deciding that 3am was late enough. Looking forward to the Heyawake which Tom has posted.

Congratulations to Tom on putting together such a fine collection of puzzles. Sorry that the early solvers met a few gremlins, but that shouldn't take away from Tom establishing himself as a fine puzzle setter. The number of Japanese competitors was fantastic.

I see Thomas Snyder has some light entertainment lined up for next weekend. Looks like another late night.

David McNeill.

PuzzleScot
Site Admin
Posts: 2740
Joined: Fri 18 Jun, 2010 10:45 pm
Location: Edinburgh, Scotland

Re: LMI October Puzzle Test 1 - Nikoli Selection

Post by PuzzleScot » Mon 11 Oct, 2010 8:52 am

Nice to see some new names under the UK banner!

I wasted loads of time on the 1st Yajilin. Got the larger one out fine though. Miscounted on the first SlitherLink.
Didn't try the one with the gates, but solved in 30 seconds (after time) once I realised the rules!
Some very tricky ones, and overall a nice selection of puzzles.

Tom, well done on your first contest! I hope you aren't too disheartened by the errors - pass the buck on to your test solvers!

drsteve
Posts: 727
Joined: Sun 27 Jun, 2010 7:23 am

Re: LMI October Puzzle Test 1 - Nikoli Selection

Post by drsteve » Mon 11 Oct, 2010 9:41 am

Wasted loads of time on the Kakuro. Maybe 15 minutes without finding the mistake. The big Slitherlink was my favourite - a great example of the advantage of shading inside and outside cells.

Well done, Tom

Maxelkat
Posts: 45
Joined: Thu 19 Aug, 2010 5:55 am
Location: Coventry UK

Re: LMI October Puzzle Test 1 - Nikoli Selection

Post by Maxelkat » Mon 11 Oct, 2010 10:06 am

My favourite was the Nurikabe. But I agree the big Yajilin was great too.

detuned
Posts: 1869
Joined: Mon 21 Jun, 2010 2:25 pm
Location: London, UK
Contact:

Re: LMI October Puzzle Test 1 - Nikoli Selection

Post by detuned » Mon 11 Oct, 2010 11:59 am

PuzzleScot wrote:Nice to see some new names under the UK banner!

Tom, well done on your first contest! I hope you aren't too disheartened by the errors - pass the buck on to your test solvers!
Disappointed, but not disheartened :)

The main thing from my point of view actually was messing up the heyawake. Modulo the stupid thing I overlooked, I thought it was the best of the 3 biggies. I have sort of fiddled it as posted above, but it's not quite the same and has definitely lost some je ne sais pas. The thing with the LITS was down to the rush (I'd spotted and fixed that typo early on, must have had too many versions of the PB file flying around), the smaller nurikabe was just silly, and well, numberlink. Thomas Snyder made a very interesting comment on my recent blog entry on numberlink which was very enlightening - and I'm far more confident of making uniquely solvable numberlink because of it!

The whole thing was done in a bit of a whirlwind, which is in some contrast to the sudoku thing coming up - so have no fear there! Incidentally the larger nurikabe/slitherlink/yajilin were almost afterthoughts, but I'm glad they've gone down well. I'd agree that the smaller yajilin was harder actually - the grading was left in the hands of the testers whilst I was off spending a long weekend in Barcelona - by the time I was back I thought the effort required to nail the gradings properly wasn't worth the effort given that the initial offerings won't overly inaccurate.

Bad luck on the Hashi btw Gareth - the answer keys were one aspect I'd have liked to have had more time to think through properly - some were quite fiddly really and conducive to errors. I'm looking forward to getting back to solving next weekend, and having silly errors only affect me :)

drsteve
Posts: 727
Joined: Sun 27 Jun, 2010 7:23 am

Re: LMI October Puzzle Test 1 - Nikoli Selection

Post by drsteve » Mon 11 Oct, 2010 12:22 pm

Just done the Nurikabe. Very nice indeed - lovely logic on the big numbers. Good job.

GarethMoore
Posts: 1209
Joined: Thu 24 Jun, 2010 9:27 pm
Contact:

Re: LMI October Puzzle Test 1 - Nikoli Selection

Post by GarethMoore » Mon 11 Oct, 2010 1:51 pm

I haven't tried the marathons properly, but in the main body there were some great puzzles with solving logic that felt very Nikoli-esque too.The puzzles I enjoyed most were the Fillomino (just right I felt), both the Slitherlinks (again just the right level), both the Yajilins (the first definitely harder, and both with some nice loop end counting bits) and perhaps most of all the LITS since I thought it was the most original in design - fantastic use of the 'all connected' rule. Unless I'm wrong I think the Heyawake required exploring one of two possible guesses, which I didn't like (perhaps I just couldn't see the logic however!). The Kakuro was also nice although I made a mistake to start with which meant I spent far too long on it. Overall a very enjoyable 1 hour 40 minutes.

detuned
Posts: 1869
Joined: Mon 21 Jun, 2010 2:25 pm
Location: London, UK
Contact:

Re: LMI October Puzzle Test 1 - Nikoli Selection

Post by detuned » Mon 11 Oct, 2010 5:49 pm

Gareth - there was definitely logic to the heyawake. The main technique came twice with the 2x5 room in the top and bottom left corners. (And here I'm assuming with the bottom left room you get to the bit where the top row of the room is confirmed to be empty).

At first it seems there are lots of ways to fit the 4 squares in, but on closer inspection it becomes quite constrained. Both times you can see that at no point can you have 3 shaded squares in any 2x3 subroom (only three of these possible, with two being direct contradictions, so not an outrageous amount to check). This means that the room essentially breaks down into 3 bits - a 2x2 with 2 squares in, a 1x2 with 0 squares, and another 2x2 with 2 squares.

Other than that, there were perhaps one or two more slightly more involved "if I do that I get a contradiction" moments, but nothing approaching what I'd call trial and error.

GarethMoore
Posts: 1209
Joined: Thu 24 Jun, 2010 9:27 pm
Contact:

Re: LMI October Puzzle Test 1 - Nikoli Selection

Post by GarethMoore » Tue 12 Oct, 2010 6:39 pm

Well I just re-solved it and I found it perfectly fine (in fact it was excellent) so I must have missed something during the competition. ;)

Just out of interest, though, re what you said I don't see how you prove early on that the entire top row of the bottom-most 4 is empty. I can prove early on that all of the top row of that shape except for the second square are empty - that last one only comes via experimentation with the two remaining options for the 4 at (3,3).

Here's where I get to where I think you have already proved that square empty: (note bottom row obscured - scroll the image to see!)
IMG_0536.JPG
Stage 1
IMG_0536.JPG (30.52 KiB) Viewed 17839 times
And then my next move is to try the two options for the 4 at (3,3) and end up with this: (bottom row still hidden)
IMG_0537.JPG
Next step
IMG_0537.JPG (29.5 KiB) Viewed 17839 times
That's what I meant by a guess, but in fact I don't think it's fair to call that a guess since you can see immediately the bottom option is no good - I think I hadn't marked in as many other squares at that point during the competition so I had to follow the guesses a lot further before working out which one was good.

Anyway, at this point you can then clearly see that the top-left 4 is two 2x2s as you say, and it all follows fairly smoothly and without any guesses.

detuned
Posts: 1869
Joined: Mon 21 Jun, 2010 2:25 pm
Location: London, UK
Contact:

Re: LMI October Puzzle Test 1 - Nikoli Selection

Post by detuned » Tue 12 Oct, 2010 11:53 pm

Gareth - the first shaded squares I intend to be shaded are in the top left box - I'm fairly impressed you've got the second four-room down solved without to be honest.

Here's my model solution:

The top left is immediately forced, by the logic above:
a.png
a.png (3.86 KiB) Viewed 17829 times
Which ever way the second 2x2 subroom goes, the following is forced:
b.png
b.png (3.93 KiB) Viewed 17829 times
After a bit of thought, the second 2x2 subroom must go this way to make sure 4 shaded squares go in the 4-room directly below it.
c.png
c.png (3.89 KiB) Viewed 17829 times

...continued
Last edited by detuned on Tue 12 Oct, 2010 11:58 pm, edited 1 time in total.

detuned
Posts: 1869
Joined: Mon 21 Jun, 2010 2:25 pm
Location: London, UK
Contact:

Re: LMI October Puzzle Test 1 - Nikoli Selection

Post by detuned » Tue 12 Oct, 2010 11:56 pm

I'm sure getting to this point now is fairly standard:
e.png
e.png (3.89 KiB) Viewed 17829 times

second application of the trick:
f.png
f.png (3.91 KiB) Viewed 17829 times
and then finally (there's a little thing to be careful with in the bottom right)
g.png
g.png (3.93 KiB) Viewed 17829 times

GarethMoore
Posts: 1209
Joined: Thu 24 Jun, 2010 9:27 pm
Contact:

Re: LMI October Puzzle Test 1 - Nikoli Selection

Post by GarethMoore » Wed 13 Oct, 2010 12:04 am

Ah, I completely missed your first move! That seems really obvious now you point it out. ;) Thank you!

I did something pretty similar to get to where I was, though - I realised whatever the arrangement of the top-left 4 and the 4 at (3,3), the two touching cells of the 4 below had to be empty. That in turn forces the whole of that room.

Your logic is easier and better. :)

Post Reply