I just want to say thanks again to Deb for facilitating the test and Grant for test solving, and best of luck to everyone who will be taking the test. I hope you all enjoy the puzzles a bunch!

Edited by davmillar 2010-11-20 12:53 AM

Enjoyed the test. Don't know why, but I had trouble entering the solution code for puzzle 7. Please check my answer manually.

David McNeill.

David, It is already marked as correct for you.
Was not expecting that one will enter 13 commas along with 20 digits :-)

Nice test, thanks! That went really quickly. I wasn't expecting so much math, but that's just me not paying attention to the IB.

I screwed up my answer to #2, entering the wrong end of the consecutive chain. Is that a minor enough screw-up that it could be overlooked?

That went rather fast, but every moment was enjoyable. I particularly liked the mirror puzzles with 0/2/5. Thanks for these puzzles.

Unlike Renban, you can see everyone's total score after you complete the test.
Note that you can refresh the score page only once in 30 minutes.

I actually found I could check the score before my clock ran out. This had two bad problems: first I had to wait 30 minutes to actually see my score due to the refresh penalty, but more importantly it seems this is not info someone should have anytime the test is running. So maybe a lock until clock has run out is best?

Edited by motris 2010-11-20 11:04 AM

Disabled the score page completely until this problem is fixed.

Puzzles were good and enjoyable. I liked ESB, Slitherlink and Mirror 0-2-5 the best.

I feel 50 minutes is too short for a test. By the time you get into the groove of solving, before you know its over. And if someone has a problem like a jammed printer (like me), then there's absolutely no time to coverup.
Also, if players finish the test within 50 minutes (or maybe 40 after seeing Thomas's 'rather fast'), then there should have been more puzzles. The concept of FLIP is fantastic and I dont see why there couldn't have been more puzzles.

Did anyone have trouble with the '1's and '7's? or '0's and '9's? or '5's and '6's? I had to strain a bit to solve (especially the sudoku) some of them. The font was a little disturbing, or maybe it was just my sleepy eye :|

@David
The concept of FLIP was brilliant and I would rate it as one of the best concepts of LMI tests after Mastermind Twins.

Fixed. Thanks for reporting.
The score page was always designed to display score only after the clock has run out.
But with some changes yesterday (especially the 30 minutes penalty which we added to reduce load on server / database), a 'bug' crept in.

We store all the times (time for submission, time of checking score etc) in the database. A quick check revealed that no one else had visited the score page before completing.
And you visited 'long time' after you submitted your last answer :-)

I also think 50 minutes is more of a warm up than a test, especially in an online scenario. I think this was one puzzle test where you did not need a printer and the puzzles could be solved equally well online (paint).

The idea of a puzzle test on a 'FLIP' theme is a good one, but it could have perhaps been a 100+ minute test.

GUYS 50 MINS IS TOO LESS..
You have to be Thomas to play in such a short time.....
I wasted 10 mins (20% of the total time) as the printer gave way.....
but nevertheless puzzles are nice... enjoyed doing them.............

I think you all are being a bit harsh to say that this test is too short and needs more puzzles to be satisfying. I don't think this was a "warm up" at all. And if all authors are required to make 2 hour tests as a prerequisite in the future, you'll likely get fewer tests since writing good puzzles is tough!.

The 12 puzzles in FLIP are a very good set and have a lot of novelty which is refreshing. I guess maybe what you are saying is you would have wanted more because they were good, and with this I certainly agree.

This discussion brings up another point which is that we do set time limits on tests that don't necessarily need to be hard limits. 1h or 75 minutes could also be used on this test, and more solvers get an opportunity to finish, without compromising the quality of the puzzles. I wonder if I should experiment with a whole new scoring format on my next LMI test (many, many months away) where perhaps there isn't a set clock, but instead the value of the puzzles declines with time. So solvers that finish the whole test in 50 minutes will get more points than someone that spreads it out over the weekend and takes 6 hours, but everyone can enjoy the test at their own speed, within the two day window, and if things like printer errors or internet disconnections or phone calls come up, you can still continue the test after the interruption ends (but with a small sacrifice to your score).

@Motris

I couldn't agree more..... :)

Rishi

Ours Brun had organised a championship in May last year in which the scoring format reflects what you suggested. I quite liked the idea and it will be interesting to see if someone takes that up.
More about the contest here: http://rohanrao.blogspot.com/2009/05/sudoku-tournament-by-ours-brun...

It would be interesting to do a test with no time limit, because I'm curious how long it would take me to complete this (or any other test, for that matter). It would have to be a shorter test like the FLIP, because I'll never get more than two hours at a time to solve one, (much less the 10 hours it would take me to do something like the USPC). Maybe the points could be based entirely on how quickly someone finishes the test, with demerits based on how many puzzles remain unsolved (or wrong)? Of course then the solver would have to figure out if it's worth more to check for errors or to keep the points.What about having 10 different PDFs, with 10 different passwords, and a week to solve all 10 puzzles?

Because of the disparity in printing out paper and such, fragmenting a test into so many pieces may not be such a good idea. But this would be an ideal option for a test that could be solved on an applet. Take a sudoku test that can be solved online; it would be an interesting format to set up 10 "start" buttons for 10 puzzles, so the individual time for each could then be recorded. Maybe scoring is based on finishing puzzles (so time doesn't matter) and top 20 solvers on each puzzle by time get small amount of bonus. So it is a race for the best solvers, and a slightly more relaxed competition that can be done over a weekend without needing 2 straight hours for everyone else.

Just to let anyone know that a Sudoku contest based on this very principle is currently running on argio-logic.net at this address [url]http://www.argio-logic.net/?p=headshrinker[/url]
the contest is based on "tough" diagrams (er, at least for me...) and the time starts ticking when a player opens an applet (take care not to open two or more applets at the same time...)

if you want to have a try.....

Stefano

Edited by forcolin 2010-11-21 12:06 AM

Great puzzles David.... It was real fun attempting these.... looking more of such puzzles to keep me interested :D

0-2-5 and the mirrors were great fun....

Thanks David - short and sweet - and definitely well executed. And not forgetting Deb on the ESB too - an entertaining puzzle!

david and deb

I typed the results for the puzzle 4 BOTH BOTTOM LINES instead of as marked (last second rush) - I wonder if I can be awarded the points for that puzzle


by the way all diagrams (at least all those I solved) were very exciting - and I like this sort of "blitz" contests. easier to find an hour in a (usually) busy weekend rather than 2.5 hours

Stefano

Edited by forcolin 2010-11-21 3:51 AM

Yes, awarded. It is taking little more time to check the claims this time because of unstable internet connections, but all the claims will be checked before the final results.

That brings me to the very interesting discussion motris has started.

1) Given a set - e.g. Decathlon - how much time should the test run so that everyone can enjoy the puzzles?
It is likely that many players don't get to solve the puzzles once the test is over. And also there is some incentive for solving the puzzle after the 'official' time.
It will be very interesting to have the points reduce with time. I think Gotroch had proposed something similar earlier.

2) The other issue is players not finding enough time at a stretch.
This is a problem for most especially given that there are so many internet competitions these days, and all of them offer something different.
Splitting into 10 sets is rather difficult to manage, but what about splitting into 2 sets each of one hour?
Those who find it difficult to get 2 hours at a stretch, may take the tests in 2 parts.
Players who needs some warm up can take the full test at a stretch.

Hi Stefano,

yes that is exactly the kind of contest that is being proposed with no time limit.... the best part is there are 3 Indians in the top 4 positions with a canadian participant at 3rd, Cant remember when that happened last time :D

The position might change as the competition is still on and somebody could always solve it faster.

Rishi

I'm not sure what you mean by "exactly that kind". I thought Headshrinkers is about (very-)hard Sudokus, and that is why the contest has no time-limit.
The purpose is completely different here.

When I say exactly that kind of contest I mean a contest with no time limit... I am not referring to the difficulty level of the puzzles....

The discussion here is about a competition which does not have a time limit. However this type of contest is possible only in an online format ... unless a separate pdf is created for individual puzzles and time for each puzzle calculated when the pwd for individual puzzle is taken from the site.

Rishi