Asian Sudoku Championship 2025
Sudoku Champs 2024
LMI Screen Test #4 — November Sudoku Test — 15th-18th November109 posts • Page 4 of 5 • 1 2 3 4 5
@ 2013-11-18 5:20 PM (#13515 - in reply to #13503) (#13515) Top

vopani



Posts: 739
50010010020
Country : India

vopani posted @ 2013-11-18 5:20 PM

Fred76 - 2013-11-18 2:46 AM

Hey my friend-despite-the-fact-you took-part-in-WPC,

Interesting reasoning, although it is wrong.

Yes, I am also an 'exception' B-) B-)


"...If '3' is in R6C3, it also has to be in R5C4..." you can conclude that 3 in R6C3 is not the answer of the JOC sudoku, but you can't conclude that 3 cannot be placed in R6C3 (it can be one of the multiple solution).

I was in such a hurry about submitting my solution after having found the 3 in R5C4 that I didn't thought that it could not be possible, as it has to be in R6C3 as well (what I've seen however).

Fred

I know, I feel so stupid. I had a hearty laugh after I realized what I did.
@ 2013-11-18 5:51 PM (#13516 - in reply to #13371) (#13516) Top

detuned



Posts: 152
1002020
Country : United Kingdom

detuned posted @ 2013-11-18 5:51 PM

Can't remember if I made this point before - I think I did for a previous screen test - but having an "undo" button seems to take away the whole point of doing things in your head. I certainly made liberal (ab)use of it...
@ 2013-11-18 5:59 PM (#13517 - in reply to #13371) (#13517) Top

detuned



Posts: 152
1002020
Country : United Kingdom

detuned posted @ 2013-11-18 5:59 PM

I'd also like to know what is "incorrect" with my E5 - JOC5 - Killer solution.

I entered R1C4 as a 6. The reasoning goes like this: 13 in the bottom right is either 256 or 346 - either way we certainly can't have R6C4 as 6. By the law of leftovers this means R1C3 cannot be a 6. The 6 in R1 must therefore go in C4 because it can't be part of a 7 or 8 cage...

edit: n/m I see e5 seems to be missing on the results page, i assume it's a bad puzzle...

Edited by detuned 2013-11-18 6:26 PM
@ 2013-11-18 6:27 PM (#13518 - in reply to #13371) (#13518) Top

An LMI player



An LMI player posted @ 2013-11-18 6:27 PM

 What was your opinion of the distribution of easy/hard puzzles? Just right
 What did you think about the puzzle quality of the test? Average
 How did you feel about the length / time limit for this test? Just right
 What is your opinion about the solving interface? Average


@ 2013-11-18 8:21 PM (#13519 - in reply to #13371) (#13519) Top

An LMI player



An LMI player posted @ 2013-11-18 8:21 PM

 How balanced do you think the puzzle types of this test were? Fairly balanced
 What was your opinion of the distribution of easy/hard puzzles? Just right
 What did you think about the puzzle quality of the test? Very nice
 How did you feel about the length / time limit for this test? Just right
 What is your opinion about rules, points and scoring for this test? Very fair
 What is your opinion about the solving interface? Very nice


@ 2013-11-18 8:58 PM (#13520 - in reply to #13371) (#13520) Top

tamz29



Posts: 225
10010020
Country : Thailand

tamz29 posted @ 2013-11-18 8:58 PM

A nice set of puzzles. I underestimated myself and decided to skip two of the JOCs and end up finishing 20 minutes early. :(
@ 2013-11-18 9:17 PM (#13521 - in reply to #13516) (#13521) Top

rob



Posts: 170
100202020
Country : Germany

rob posted @ 2013-11-18 9:17 PM

detuned - 2013-11-18 5:51 PM

Can't remember if I made this point before - I think I did for a previous screen test - but having an "undo" button seems to take away the whole point of doing things in your head. I certainly made liberal (ab)use of it...


I agree, felt a lot faster on the harder puzzles once I started trying out things. Maybe one or two levels of undo to deal with actual typos?
@ 2013-11-19 12:11 AM (#13523 - in reply to #13521) (#13523) Top

rvarun



Posts: 268
100100202020
Country : India

rvarun posted @ 2013-11-19 12:11 AM

Hi. I completed the test. But my name is not in the scoresheet.
@ 2013-11-19 12:23 AM (#13524 - in reply to #13371) (#13524) Top

rvarun



Posts: 268
100100202020
Country : India

rvarun posted @ 2013-11-19 12:23 AM

Sorry guys. The ST 4 Page had this link after finishing the test. This was for ST 3.

You have completed the test. Check score page at http://logicmastersindia.com/ST/ST3/score.asp



Edited by rvarun 2013-11-19 12:24 AM
@ 2013-11-19 10:07 PM (#13537 - in reply to #13521) (#13537) Top

prasanna16391



Posts: 1801
1000500100100100
Country : India

prasanna16391 posted @ 2013-11-19 10:07 PM

rob - 2013-11-18 9:17 PM

detuned - 2013-11-18 5:51 PM

Can't remember if I made this point before - I think I did for a previous screen test - but having an "undo" button seems to take away the whole point of doing things in your head. I certainly made liberal (ab)use of it...


I agree, felt a lot faster on the harder puzzles once I started trying out things. Maybe one or two levels of undo to deal with actual typos?


We actually discussed a lot about this, and it is the reason why we ensured that if you change color, it'll change throughout, meaning color changes are only for player preference rather than guesswork. With undo, we're not sure what is a great way of doing things. My point is someone solving could honestly make a mistake and it'd help to undo to retrace. So the only use isn't guesswork, and I'm always in favor of keeping things that help normal solving even if they could be used towards shortcuts, so we decided to keep it. Maybe restricting to one or two levels is a good idea, but I personally am still non-committal to the same. Would love to hear more views on it.

I'll read back now to other points and queries. I only addressed the more pressing ones while I was away conducting a puzzle event here simultaneously over the weekend.
@ 2013-11-19 10:09 PM (#13538 - in reply to #13510) (#13538) Top

prasanna16391



Posts: 1801
1000500100100100
Country : India

prasanna16391 posted @ 2013-11-19 10:09 PM

rajeshk - 2013-11-18 12:16 PM

In the solving interface, it would have to better if we could have been given different colors for backtracking the puzzles.


joshuazucker - 2013-11-18 4:02 AM

The colors interface would have been nice for bifurcating at times if I could have figured out how to make it work for me instead of changing the color of everything when I clicked on a color.


As mentioned in my post above, we deliberately decided to keep it this way, and keep the undo button as the sole method of backtracking.
@ 2013-11-19 10:18 PM (#13539 - in reply to #13498) (#13539) Top

prasanna16391



Posts: 1801
1000500100100100
Country : India

prasanna16391 posted @ 2013-11-19 10:18 PM

Semax - 2013-11-17 10:21 PM

And another little thing that I don't understand: Why is there an automatic submission after 5 minutes? There are so many situations where this is bad. Can you tell me one situation where this is good?


It is probably the same reason why normal tests have a time limit instead of just saying, solve and submit whenever you like. Here the expected difficulty per grid is really low compared to normal tests, so a per grid time limit adds to the competitive aspect of things. In other words, I'd just say its easier and so we can experiment with a different quick-fire setting for it so we do. It is the general idea of a screen test (the same was used in WPC 2011 too) that we've tried to extend to Sudoku. The bigger challenge is in ensuring that at least most of these are easy enough for everyone to submit some in spite of time limits. Looking at the general feedback I feel at least that has been achieved this time.
@ 2013-11-19 10:25 PM (#13540 - in reply to #13488) (#13540) Top

prasanna16391



Posts: 1801
1000500100100100
Country : India

prasanna16391 posted @ 2013-11-19 10:25 PM

Nilz - 2013-11-17 3:14 AM

Given that a lot of people won't expect to complete every puzzle, 'tactics' and puzzle choice becomes very important (given that the strategy 'never skip' clearly cannot be optimal, except for the best), so I wonder if it would be better to have a more accurate gauge of the level of difficulty of the puzzles, if possible?


We were actually worried about this. Ideally the solution is to have a system where a solver can get back to skipped puzzles only IF all of the order is finished before time. Unfortunately we did not have the time this time to get that system in place, so made do with the note in the instructions. A more accurate gauge of difficulty is interesting though, just in case the system itself proves difficult to implement. Is there any suggestion on how this can be done? It'd probably become too formulaic if we say that a player with X ratings is expected to be on the border of finishing, but I see that as the most realistic way of providing an idea about difficulty.
@ 2013-11-20 12:53 AM (#13543 - in reply to #13371) (#13543) Top

uvo



Posts: 21
20
Country : Germany

uvo posted @ 2013-11-20 12:53 AM

Is there a way to make the scoring more transparent? I have no idea why the ranking is the way it is, for example places 19 and 20 look especially strange.
@ 2013-11-20 1:49 AM (#13545 - in reply to #13543) (#13545) Top

prasanna16391



Posts: 1801
1000500100100100
Country : India

prasanna16391 posted @ 2013-11-20 1:49 AM

uvo - 2013-11-20 12:53 AM

Is there a way to make the scoring more transparent? I have no idea why the ranking is the way it is, for example places 19 and 20 look especially strange.


A few things -

B2+E5 have been canceled and points have been added considering the amount of time wasted on those two puzzles. Timothy wasted more time on them.

E4 is not considered for bonus criteria, because some had a bad experience on it with rules being misplaced. JOCs are considered separately and only their own points are removed. So, Timothy has only 6 cells wrong across the other grids (E2 and E6) and is eligible for the bonus.

We'll add a few notes explaining the scoring add-ons on the test page soon. As of now this is the gist. There is the drop down list with the choice of Points instead of Seconds where you can check about the points being given for each grid.
@ 2013-11-20 5:21 AM (#13546 - in reply to #13545) (#13546) Top

prasanna16391



Posts: 1801
1000500100100100
Country : India

prasanna16391 posted @ 2013-11-20 5:21 AM

prasanna16391 - 2013-11-20 1:49 AM

We'll add a few notes explaining the scoring add-ons on the test page soon. As of now this is the gist. There is the drop down list with the choice of Points instead of Seconds where you can check about the points being given for each grid.


These are now added to the test page here.
@ 2013-11-21 4:19 PM (#13552 - in reply to #13546) (#13552) Top

Administrator



20001000500202020
Country : India

Administrator posted @ 2013-11-21 4:19 PM

Score page is now final. The bonuses were activated manually based on eligibility mentioned in the Guide and on the submission page.

http://logicmastersindia.com/ST/ST4/score.asp
@ 2013-11-25 12:10 PM (#13576 - in reply to #13371) (#13576) Top

prasanna16391



Posts: 1801
1000500100100100
Country : India

prasanna16391 posted @ 2013-11-25 12:10 PM

Bit late to posting a summary because of traveling and then falling ill. Anyway, just a few quick notes.

Firstly, congratulations to jaku111 (2982.4 points), TiiT (2882.6 points) and deu (2865.1 points), the podium finishers.
This screen test definitely had some things that were bad, pretty obviously my own doing. But the biggest positive I'll take from it is that the general format/timing/etc. seems to have been a success. I wanted many solvers to be able to attempt most of the grids, and this is what happened.

Also, about Just One Cell idea, I think its probably fine to do but it takes more checking, and probably not using variants with too many constraints.

The interface is superb as always, but as usual had some feedback directed towards it. Firstly, the color button implementation, I think, was fine. It'll stay. I'm still not sure about undo button being limited to 2/3 clicks. One change we must work to implement, especially with this easier format, is to enable a solver to visit back to a grid if the solver goes through all 30 grids with time remaining.

Thank you all for participating, and also for the feedback
LMI Screen Test #4 — November Sudoku Test — 15th-18th November109 posts • Page 4 of 5 • 1 2 3 4 5
Jump to forum :
Search this forum
Printer friendly version