You are here

Puzzle GP - round 2 - feedback

Thank you to the Swiss Team ! I really enjoy this round.
I'm not a very good player (too slow) but I try for 3 years.
The puzzles were very possible for me and I continue them after the round.

Thank you for the fun !

Why 310031330228 it's not correct ?
My aswer was not correct why i have a typing error, bat i considering this answer correct !

Why 310031330228 it's not correct ?
My aswer was not correct why i have a typing error, bat i considering this answer correct !

Hi, would you please consider having answer keys not longer than 10 numbers/letters per row? I still believe that solving is the most important part of the competition.


I agree with Matus. You spend 5 minutes to solve a puzzle. Then you spend 1 minute to enter the code. And the code is more important, because you can lose the 6 minutes effort.
About answer codes, I have another concern. Placing unnecessary elements inside the puzzle makes it more difficult to solve. I prefer to see puzzles with answer elements only outside the puzzle grid, as it is used in Sudoku GP.

Of course, ideally, we would have everyone submit their papers and then graders could look at the solution and determine if they are correct, just like at the finals.

Unfortunately, we do not have the budget to hire hundreds of graders who will come to everyone's home and check your papers.

The main reason for having an answer key is so that we can check hundreds of entries every competition.

The main reason for having a long answer key is so that we can be sure the player has actually solved the puzzle. In the case above, 310031330228, it looks like you placed the battleship in column 3, which is not correct. A long answer key is slower than a short answer key, but every competitor has to do it, which ensures fairness.

So, these two factors ensure that our competition is timely and fair.

I think solving is important, but it is not most important; timeliness and fairness is more important than solving.

But I can compromise: anyone who thinks solving is the most important, feel free to mail me your answer sheets and then if they are correct I will call you the unofficial winner. I will be very slow to grade them and it will not be fair, but solving is the most important to you, right?

thank you very much

I'm not sure if this was the intention, but this reply seems designed to shut down down discussion, and even comes across as being glib and/or patronising. I think we're all well aware as to why we might need entry codes.

Discussion about the quality of entry codes is another matter. Surely competition organisers want to improve solving experience as much as is possible within their means - and I absolutely believe that getting answer codes to below 10 characters is a) achievable and b) would improve solving experience.

Taking battleships as an example, an entry code which uses the coordinates of the 1x1 ships goes back at least a decade - this strikes me as a simpler alternative which has been tried and tested (although also not without its own negative points)

you're right. The answer seems patronising, but considering that your initial request came across as rather demanding, IMO it was exactly the right way to respond.

You seem to be under the impression that the rules of the contest are up for "discussion". They aren't. We all knew the rules when we signed up to participate, and I think most players have experienced a correctly solved puzzle not being counted due to a typo or a misunderstanding at some point. It happens, it sucks, life goes on.

I - for instance - did correctly solve one of the Four Winds on paper, but I entered the numbers in the circles in the wrong order (top to bottom, instead of left to right), but you don't see me making a fuss about it. It was a stupid oversight, ... but it was MY mistake, which could have easily been averted had I read the instructions more carefully.

Discussing user-friendliness of solucition codes in itself would have been a fine discussion topic (had you led with that), but you started this topic with a complaint about your answrs now having been counted, clearly indicating that it's not "just about solving" for you, but that you also want official recognition.

However, this contest isn't "just about solving" either. The task at hand is to solve the puzzles AND enter the correct solution codes, like it or not. It's comparable to Long Jump at the Olympics. That discipline requires from the athletes to jump as far as they can AND to make a proper landing. If one falls backwards after landing, then the distance they actually did jump doesn't count.

Ironic and rude reaction is never the proper way to answer... At the beginning, I was not talking about my wrong answers (this time I broke the puzzles, not answer codes), the lenght and senselessness of answer codes makes me worry. I agree that landing is a part of long jump at Olympics, but please note that it's different to land in sand or gravel...

Just for recapitulation: 12 numbers in Battleships (worth 10 points), 14 letters in Tic-Tac, 15 numbers in Hashi... This makes no sense

I was being absolutely serious about offering people to mail me their answer sheets. No irony was intended.

I understand how the length of answer codes can be frustrating, but it is not senseless. There are reasons behind the selection of answer codes, and while these reasons may not be prioritized in ways individual solvers might like, they are not random.

Here are the general priority I follow when creating answer codes:

* A correct code should indicate with reasonable certainty that the solver has solved at least 95% of the puzzle. If a solver can solve 50% of the puzzle and know what the correct code, that is a failure.

* Codes should be of fixed length if possible, to allow the automated real-time checker to inform the user of small, easily caught mistakes. (Some codes for historical reasons are not fixed length.)

* When spanning a row, all columns should be used, to avoid players skipping the wrong columns when they have a correct answer.

* Codes should not mix entry rows and entry columns, to avoid chances of accidental reversal or backwards reading. Rows should be given preferences over columns. Left-to-right should always be used instead of right-to-left.

* Codes should be digits and numbers only, as symbol locations can be very different across keyboards, and punctuation and spaces could be mistaken for answer boundaries. If possible, restricted set of symbols should be given to the automated real-time checker.

* Length of answer codes should be as small as possible, but this takes a lower priority to the previous goals.

Overall, there are now generally four types of answer codes:

(A) Choose certain critical rows, answer full information about those rows.

(B) Choose a cell in each column (usually dotted), answer information about all dotted cells from left-to-right.

(C) Choose certain critical rows, answer partial information about those rows.

(D) Reduce each row to a single character of partial information, answer one character for each row, top-to-bottom.

For new puzzle types (have not appeared in GP before), I try to choose the first type that hits that 95% threshold. I would rather not use (C) because of the length-check condition, but I keep classics like Cave and Slitherlink with those for historical consistency. (D) is sometimes necessary for puzzles where (A) and (C) don't hit the 95% mark and (B) would give away too much information about solving approach, such as Battleships and Star Battle.

If you could go more into detail as to why you feel "12 numbers in Battleships (worth 10 points), 14 letters in Tic-Tac, 15 numbers in Hashi" doesn't make any sense, I would be quite glad to hear it.

Thanks for explanation.
I think for puzzles with binary solution, like snake or nurikabe for example, use of 1 and 0 would be better instead of O and X. O and X are not so close in standard keyboard. 1 and 0 are close if you have numpad.
And actually, (A) full information for rows covers already (C) partial information. For a slitherlink puzzle solution, if I write 1100010011,
it already shows the 212 in partial answer key. I think the full information is more easy to follow.

Thanks for the suggestion. I will consider it, but some of my thoughts:

Not everyone has a numpad (for example, I actually do most of my work on my laptop, which only has a numpad mode). I would worry that 0 and 1 might give an advantage to solvers with numpads. "O" and "X" are in pretty much the same place on most keyboard layouts -- I suppose if you used Dvorak you could get a small advantage.

Full information is indeed more easy to follow, but partial information (like 212 for Slitherlink) is less typing and so I think some solvers like it more. It is impossible to please everyone :)

Oops... and there's another oversight. I thought both the initial complaint and the latest Response were from the same user named "Permalink".

Please ignore my post - isn't there an edit or delete function here? I can't seem to find it.