I did the test for fun, without much regard for speed or time, so I am not bothered about my own penalties. But I do take a keen interest in contest design and I think it is probably true there are already a higher than 'normal' proportion of penalties here.

Both of the suggestions above are good in my opinion. Indeed, in a recent wpf puzzle gp, cells with direction of travel was successfully used.

How do we calculate the answers to be submitted for numberlink? I tried max segment connecting a number as well as max segment in that direction

Did you understand how the answer keys are calculated in the IB?

No, It looked as if max segment that originates at a number, excluding number block

Puzzle Ramayan is a great concept and I was looking forward to this, but after doing the first part, sadly, I'm not too elated.

Firstly the puzzles were all easy. I understand you want to welcome newcomers but is it necessary that all 4 puzzles are simple?
This way, regulars at LMI now only have sprint tests to look forward to each month.
Perhaps have puzzle #1,#2 being easy/medium and #3, #4 being a little harder?
Because right now, extracting the answer keys takes longer than solving the actual puzzles.
Speaking of which, I recommend using marked squares in Numberlink (like in Japanese Puzzle Land) and Ivan already mentioned improvements for Four Winds and Walls.

Next, as Tom already pointed out, the inclusion of Hitori Blocks was questionable. We have 21 easy/straight-forward puzzles and 1 that potentially ends up being the game decider for the regulars.
I just don't think a test that most people will finish should have a puzzle that is so different in level of difficulty.

Last thing I'd like to nibble on is the Hitori.
Now this will be subjective, but we all know Hitori is a "chore" puzzle with lots of 'noise' (see Thomas Snyder's Hitori where he eliminated all the 'noise' [numbers that are already unique in its row and column] in his old blog). There are some neat steps in Hitori that crop up frequently on Nikoli's handmade Hitoris. But I did not see any here except for the third Hitori where you had big stairs going on. If you're active on nikoli.com you could see that authors put in great effort to minimize the 'global' deduction in their puzzles favouring only 'local' deductions. What I mean is that forgivingly, on a large Hitori, you'd never see the same number spanning the entire row/column. Say: 1213145678 (with 1s tightly grouped together) as opposed to 12314567191 (and solvers have to dart their eyes a lot).

Now, I'm only criticizing the test structure here, the individual puzzles I have nothing against.
To be fair, all the puzzles were clean with some containing beautiful solving paths.
Newcomers will definitely benefit from them and I sincerely hope Puzzle Ramayan will introduce puzzles to a wider audience in India and worldwide.
It is also good to see LMI active again.

Will look forward to the rest of the PRs.


Edited by tamz29 2015-09-06 8:12 PM

I thought the answer keys were fine, and it was good that they were relatively consistent across a lot of puzzles (e.g. 'how long is the longest line' was the key to numberlink and four winds and others). However, it does help that I tend to draw my lines right up into the numbered cell, if you don't do that I can understand why it might be difficult. As an alternative, how about 'For each cell of the given row, give the number it belongs to.' I'm sure I've seen that used before.



I'm a bit embarrassed to say I didn't know either of these tricks (the 8*8 Star Battle or the 6*6 Non-consecutive sudoku). Looks like that might have cost me several places, possibly cost me a win (so far, anyway). I'm not hugely worried about that, what I am worried about is: are there any other such tricks I should know! I'm also curious to know whether any other decent-ish solvers didn't know either?

Entered the code for the Four winds and got an "X". Couldn't find an error and moved on. Another "error" in Four Winds No. 3. I did not only loose the points for this puzzles, but it cost me enough seconds that I couldnt finish inserting the solution code for the Walls puzzle. Anyway, it doesn't matter, because it would have been considered "wrong" for the same reason as puzzles No.1 and 3.
One of the most annoying and misleading solution codes I can remember in any contest.

Also I think the other comments are right about the 8x8 Star Battle and the Nonstandard Hitori Blocks. Walls may be considered "Standard", I think. So, a terrible test for whoever might be the target audience, newcomers or veterans.

I knew about the star battle thing beforehand, but I completely forgot about it while taking the test and wasted like 4 minutes oops. I didn't know about the 6x6 non-consecutive sudoku thing, though, and I don't know any other tricks of this kind.

Fun fact: The star battle trick (well, a generalization) actually appeared on the IMO Shortlist (problem pool for the International Math Olympiad) in 2010: http://www.artofproblemsolving.com/community/c6h418687p2362298

Edited by anderson 2015-09-07 9:27 AM

I enjoyed this, thanks!

I noticed the potential Four Winds answer key ambiguity when entering, decided to risk it (and caught a penalty) instead of looking up the instructions. But careful reading of the booklet would have prevented this, and I think that's generally a good lesson to teach.

My other penalties were just sloppy answer code entry...

I had fun with the Hitoris, though as usual it turns out that half of the interesting logic I found could have been circumvented by noticing obvious things... The Hitori Blocks did seem hard in comparison. I think the variant adds quite a bit of depth, would have been nice to have an (extra?) easier one.

I knew the 8x8 Star Battle thing, but didn't realize the puzzle was 8x8 for a bit. Again I think it's fine to put in things like this, though having this as the first of the set can be a bit unfair for newcomers.

Based on a request from a player, this contest is extended by few hours.

I certainly didn't know about 8x8 Star Battle when one appeared in the WPC 2004 Finals... (and I'm quite sure the organizers back then didn't either).

I agree that the choice to have a single somewhat more difficult puzzle (the Hitori Blocks) was a bit odd, but I am glad that it was included and I really liked how different the techniques for it were. For future rounds I would be interested in seeing "exploratory" puzzles at the end of every section, rather than just two of them. I also think that it would be nice to add some more challenging puzzles in future rounds, though I do like that the lower difficulty level makes the competition more accessible to newcomers.

I am not able to open to do the test.

Hi. The test is still open. Please check the below link and login to start the test.

http://logicmastersindia.com/PR/201509/

You need to download the password protected Puzzle booklet which is present in the top left side of the page.
Login and click Start Classics. You will get a password.
Use the password to open the Puzzle booklet and start solving.

The answer keys need to be entered in the boxes provided

I really enjoyed the test. However despite knowing the trick of 8x8 star battle, I wanted to solve it logically and till now I am not able to solve it. It will be great if someone who solved it logically can post the steps to solve it.

Hi guys. What is the 8x8 Star Battle trick. :)

There are exactly two possible solutions for an 8x8 star battle. One is in the IB. To figure out the other you just need to reflect the configuration.

Yes.Adding to what prasanna said and to make things more clear , this is the scenario regardless of whatever shape the irregular region may take and which means that if you know for sure that one is not the configuration then the other is the solution ! for an 8X8 SB. Incidentally,I got to know about this just before the start of the test.

I am trying to start the test last 25 mins but looks like I am not able to. Any reason

The contest ended around an hour back.

The first episode of the inaugural season is over. Congratulations to Prasanna, Swaroop and Amit for taking the top 3 spots. They were the only three from India to complete the set. Anurag, Rajesh and Zalak took the next three spots.

Amongst the international participants, Congratulations to Endo for the top spot and Ulrich and Hideaki for second and third positions. They were way ahead of other participants.

Hitori Blocks (not surprisingly) and Star Battle 1 (surprisingly) had the least submissions. Four Winds (1 and 3) had less than 80% correct answers. This could have a correlation with the answer key as mentioned by some participants. Numberlink 1 also had 67% correct submissions. This could be due to the puzzle being the first one and most participants would have attempted this puzzle.

Thank you everyone for participating, and hope to see you soon in the next episode (Numbers) by Deb.

New players : If there is any specific Puzzle in which you got stuck, feel free to post here and we will guide you through.

I could not solve Star Battle 1 logically. It will be great if someone could post logically steps to solve it.

I am new. to this puzzles. Can the answers be seen?

Hi. Generally the solutions will be appended to the puzzle booklet in a day or two. Do follow this thread for the update.

Points table for neiighbours

Edited by devarajand 2015-09-11 11:45 AM