croco-puzzle

Links to puzzle sites all around the web
dickoon
Posts: 615
Joined: Mon 21 Jun, 2010 12:52 pm
Location: Stockton-on-Tees, UK
Contact:

Re: croco-puzzle

Post by dickoon » Thu 17 Feb, 2011 9:57 pm

Calavera wrote:you are assigned a "virtual" score between 0 (not solved/worst time) and 3000 (best time) with 1500 being at the median of all participants.
OK, there's something here I'm not quite clear on and would appreciate an explanation, please. When the median time is calculated, for the purpose of setting the 1500 level of performance, are Fehlversuch five-minute penalties included?

Second question: on the techdata page, what do the figures mean, please? I associate 6 kyu with 1500, or the median, so I'm not clear why the time listed for 6 kyu tends to be faster than the median time. Is it the case that the time listed for for 6 kyu on that page reflects an "average 6 kyu performance", i.e. half-way between the lower boundary from a 1500 performance and the higher boundary from a 1600 performance, thus the 6 kyu time reflects a 1550 performance? If this is the case, is it also true that the time listed for 7 kyu represents a 1450-level performance, and so on up and down the scale, with the time listed for 20 kyu representing a 150-level performance and the time listed for 9 dan representing a 2950-level performance?

I'll add a little more to what Calavera said about kyu and dan. When you click on someone's name on the rating list, you'll see a graph for them with three lines: a black line, which shows your rating updated every day, a green line which reflects a rolling average of your 50 last daily ratings and a red line which reflects a rolling average of your 200 last daily ratings. If you have fewer than 50 or 200 ratings, the rolling average fills up the 50 (or the 200) with lots of zeroes, which will drag it down. When the red line reaches a multiple of 100, you earn the appropriate kyu or dan title; 100 earns you 20th kyu, 200 earns you 19th kyu and so forth; 2000 earns you 1st kyu, 2100 earns 1st dan and so on up to 2900 which earns 9th dan. It took me 70 days to earn 20th kyu, and it took Thomas Snyder (mars) something like 30 days, so I suspect it will take you somewhere in between. :)

Titles are awarded for life, even if your rolling average sinks well below that threshold later on; this is potentially quite a strong assertion and I cannot back it up, but I have a suspicion that the overall standard of croco-puzzle solvers has improved over the years, so someone who earns 1st dan in 2012 arguably represents greater strength than someone who earnt 1st dan in 2010.

Chris
(who did, indeed, correctly fill in all the numbers and then pressed "Testen" before filling in all the black squares)

Calavera
Posts: 73
Joined: Wed 03 Nov, 2010 1:42 am

Re: croco-puzzle

Post by Calavera » Thu 17 Feb, 2011 11:41 pm

dickoon wrote:OK, there's something here I'm not quite clear on and would appreciate an explanation, please. When the median time is calculated, for the purpose of setting the 1500 level of performance, are Fehlversuch five-minute penalties included?

Second question: on the techdata page, what do the figures mean, please? I associate 6 kyu with 1500, or the median, so I'm not clear why the time listed for 6 kyu tends to be faster than the median time. Is it the case that the time listed for for 6 kyu on that page reflects an "average 6 kyu performance", i.e. half-way between the lower boundary from a 1500 performance and the higher boundary from a 1600 performance, thus the 6 kyu time reflects a 1550 performance? If this is the case, is it also true that the time listed for 7 kyu represents a 1450-level performance, and so on up and down the scale, with the time listed for 20 kyu representing a 150-level performance and the time listed for 9 dan representing a 2950-level performance?
First answer: The five-minute penalty is only for the daily topscorer list. When calculating the median and the daily rating a flexible time penalty is taken. This time penalty is exactly the median of all solvers who didn't need Fehlversuche. After this penalty is added to the individual times the median for the rating and the ratings are calculated.

Second answer: Yes, your average-theory is correct.

ronaldx
Posts: 407
Joined: Mon 21 Jun, 2010 2:41 pm

Re: croco-puzzle

Post by ronaldx » Fri 18 Feb, 2011 1:36 am

Dickoon: yes, I agree something odd is going on.

The Japanese Sums of 17 Feb is a good example to test because
1) nobody above the median got a fehlversuch
2) the fehlversuch penalty worked out to be more than 5 mins
meaning the fehlversuch penalty has no effect on the median

Here the "Technische Daten" median is (correctly) listed as 6:12, which we know scores 1500.
but the 6kyu score is listed as 6:00
According to brief calculations, you're right and the listed 6 kyu score is equivalent to 1550 (/in between 1500 and 1600 would be approx the same)
Similarly the 9 dan score is equivalent to 2951.7 (as close as you can get to 2950 when measuring to the nearest second)

Interesting, I had also assumed the listed 6kyu score was equivalent to 1500.
Thanks.

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

Re: croco-puzzle

Post by drsteve » Fri 18 Feb, 2011 9:42 am

Lesson Number 3: You don't have to mark empty squares as such on Heyawake, unlike Japanese Sums.

Is there a list on the site somewhere of individual rules, or should I just experiment with the prize puzzles?

Calavera
Posts: 73
Joined: Wed 03 Nov, 2010 1:42 am

Re: croco-puzzle

Post by Calavera » Fri 18 Feb, 2011 11:27 am

drsteve wrote:Lesson Number 3: You don't have to mark empty squares as such on Heyawake, unlike Japanese Sums.

Is there a list on the site somewhere of individual rules, or should I just experiment with the prize puzzles?
Once more I can only recommend the puzzle-wiki I linked on the previous page. If you look at the Japanese Sums there the first sentence starts with "Blacken some of the fields...". So blackening the fields is part of the job. Without looking I'd bet Heyawake rules don't include a sentence like "mark all other fields as empty".

By the way: A few weeks ago there were another Japanese Sums and some Germans ran into the same mistake of not blackening all fields necessary as it was a small diagram. So it's not so much a question of language as you might think and more a question of interpreting (or reading) the rules (which is certainly easier for people speaking German).

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

Re: croco-puzzle

Post by drsteve » Fri 18 Feb, 2011 2:52 pm

Very helpful wiki. It's cleared up another couple of problems I had. Thanks

willwc
Posts: 4
Joined: Mon 08 Nov, 2010 12:06 am

Re: croco-puzzle

Post by willwc » Fri 18 Feb, 2011 6:03 pm

drsteve wrote:Lesson Number 3: You don't have to mark empty squares as such on Heyawake, unlike Japanese Sums.

Is there a list on the site somewhere of individual rules, or should I just experiment with the prize puzzles?
in addition to the Wiki, if you click on Anleitung [Instructions] when you open the applet, the last sentence of the Aufgabe [Task] page describes the requirements for what is considered a complete solution. For example, the Japanese Sum instructions state (translated by Google): "For a correct solution must in addition to all numbers, all black boxes are located."

dickoon
Posts: 615
Joined: Mon 21 Jun, 2010 12:52 pm
Location: Stockton-on-Tees, UK
Contact:

Re: croco-puzzle

Post by dickoon » Fri 18 Feb, 2011 6:15 pm

Calavera wrote:First answer: The five-minute penalty is only for the daily topscorer list. When calculating the median and the daily rating a flexible time penalty is taken. This time penalty is exactly the median of all solvers who didn't need Fehlversuche. After this penalty is added to the individual times the median for the rating and the ratings are calculated.
Thank you! That clearly explains the "Strafzeit pro Fehlversuch" figure on the techdata page whose meaning I had not properly understood.

dickoon
Posts: 615
Joined: Mon 21 Jun, 2010 12:52 pm
Location: Stockton-on-Tees, UK
Contact:

Re: croco-puzzle

Post by dickoon » Sat 19 Feb, 2011 11:26 pm

It's more than a little cheeky to post this, but it was briefly true.

I keep such good company, albeit not for long.
Attachments
croco-podium.png
I'm #3! I'm #3!
croco-podium.png (88.85 KiB) Viewed 20940 times

ronaldx
Posts: 407
Joined: Mon 21 Jun, 2010 2:41 pm

Re: croco-puzzle

Post by ronaldx » Sun 20 Feb, 2011 12:09 am

hehe you may have posted the wrong link but I saw that originally dickoon, great time! (oh no I see you mean the sidebar, hurray!)
Also that's a lot of browser tabs!

I had a nightmare on the Kakuro today... solved it bar 2 cells, then hit the wrong key, browser changed to another page, had to go back and solve it again... mistyped one cell... fehlversuch... solved.
great stuff!

berni
Posts: 163
Joined: Mon 01 Nov, 2010 11:38 am
Contact:

Re: croco-puzzle

Post by berni » Mon 21 Feb, 2011 3:39 pm

drsteve wrote:Might be nice to be able to gauge how long it will take me to climb the greasy pole.
Just a small clearification: After about 300 days (assuming you tried both puzzles every day) your rating reaches "your level". After about 500 days the dan/kyu-Rating reaches "your level".

PS: Somehow it seems that I don't get notifications on new posts in this forum... A few weeks ago it worked...

PuzzleScot
Site Admin
Posts: 2755
Joined: Fri 18 Jun, 2010 10:45 pm
Location: Edinburgh, Scotland

Re: croco-puzzle

Post by PuzzleScot » Mon 21 Feb, 2011 7:44 pm

Somehow it seems that I don't get notifications on new posts in this forum
Anyone else got issues? If so, it's something to look into...

Calavera
Posts: 73
Joined: Wed 03 Nov, 2010 1:42 am

Re: croco-puzzle

Post by Calavera » Mon 21 Feb, 2011 9:44 pm

I don't remember ever checking the relevant box so I can't say anything about that.

But I have a wild guess about how the problem might occur: If you post again in the subscribed thread without checking the notification option the forum might unsubscribe you from the thread. Just a theory though...

berni
Posts: 163
Joined: Mon 01 Nov, 2010 11:38 am
Contact:

Re: croco-puzzle

Post by berni » Tue 22 Feb, 2011 8:21 am

This time it worked. :wink:

detuned
Posts: 1953
Joined: Mon 21 Jun, 2010 2:25 pm
Location: London, UK
Contact:

Re: croco-puzzle

Post by detuned » Wed 23 Feb, 2011 11:23 pm

ouch...not had some fehlvers for a while now (although a couple of slow times when spotting and correcting an error). Now two typos come along and bring me two in a day! :oops:

This is almost as bad as my retrospective observation that I neglected to fill in two digits of Steve's irregular sudoku on Saturday! :roll:

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

Re: croco-puzzle

Post by drsteve » Fri 25 Feb, 2011 6:45 am

In terms of shortcut keys, is there a quick way for Arukone other than hitting the letters on the keyboard?

berni
Posts: 163
Joined: Mon 01 Nov, 2010 11:38 am
Contact:

Re: croco-puzzle

Post by berni » Fri 25 Feb, 2011 11:06 am

drsteve wrote:In terms of shortcut keys, is there a quick way for Arukone other than hitting the letters on the keyboard?
You can use mouse-dragging (click a letter and move the mouse while keeping the mouse pressed). This also works in a couple of other applets (Masyu, Sternenhimmel, Snake, Thermometer).

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

Re: croco-puzzle

Post by drsteve » Fri 25 Feb, 2011 8:32 pm

Thanks. It should speed things up next time.

PuzzleScot
Site Admin
Posts: 2755
Joined: Fri 18 Jun, 2010 10:45 pm
Location: Edinburgh, Scotland

Re: croco-puzzle

Post by PuzzleScot » Sat 26 Feb, 2011 11:01 am

It's astonishing that today some people solved both puzzles in under one minute combined! Roland did the pair in 33s! Wow.
Must be a record for combined times, if not for individual puzzles too.

I hope with the inclusion of B teams, Roland makes it this year.

berni
Posts: 163
Joined: Mon 01 Nov, 2010 11:38 am
Contact:

Re: croco-puzzle

Post by berni » Sat 26 Feb, 2011 11:34 am

PuzzleScot wrote:It's astonishing that today some people solved both puzzles in under one minute combined! Roland did the pair in 33s! Wow.
Must be a record for combined times, if not for individual puzzles too.
This is indeed a new record. Up to now it was 59 seconds by mars at 31st of January. For individual puzzles the record was 4 seconds (21st of august 2010 by pwahs); at that time there was no second puzzle.

Calavera
Posts: 73
Joined: Wed 03 Nov, 2010 1:42 am

Re: croco-puzzle

Post by Calavera » Sat 26 Feb, 2011 12:33 pm

PuzzleScot wrote:I hope with the inclusion of B teams, Roland makes it this year.
In the last years the reason Roland didn't participate in the WPCs wasn't that there's only one team. He simply didn't take part in the qualifier and thus also not in the Logic Masters whose first four places are sent to the WPC. As soon as he changes this he should have quite a chance to be in the top four.

Still I'd like to see B teams at this years WPC for -eh- personal reasons ;).

PuzzleScot
Site Admin
Posts: 2755
Joined: Fri 18 Jun, 2010 10:45 pm
Location: Edinburgh, Scotland

Re: croco-puzzle

Post by PuzzleScot » Sat 26 Feb, 2011 2:02 pm

Talking of potential German participants - Is Marian Kraus still puzzling? I got on very well with him at two WPCs...

Calavera
Posts: 73
Joined: Wed 03 Nov, 2010 1:42 am

Re: croco-puzzle

Post by Calavera » Sat 26 Feb, 2011 2:21 pm

PuzzleScot wrote:Talking of potential German participants - Is Marian Kraus still puzzling? I got on very well with him at two WPCs...
I'm not entirely sure but I think that he's SenorDido on crocopuzzle. He's definitely still active as he's reached 7th place in last years Logic Masters.

motris
Posts: 77
Joined: Tue 21 Dec, 2010 5:35 pm
Contact:

Re: croco-puzzle

Post by motris » Sat 26 Feb, 2011 3:45 pm

berni wrote: This is indeed a new record. Up to now it was 59 seconds by mars at 31st of January. For individual puzzles the record was 4 seconds (21st of august 2010 by pwahs); at that time there was no second puzzle.
If only I'd known I had that record from my great Jan 31st Rundweg solve, I would be more bothered to lose it on a day like today. By my rough estimate 20 solvers broke the record today. So this is just a super outlier day with two very easys, much like that 4s solve must have been. I've not seen any puzzle in the 200 I've played yet where even knowing the solution I could replicate a 4s time, except when I jump over to say the 2007 Advent puzzles and go to the easiest category.

I know the puzzle type is determined randomly by an assigned formula depending on first/second group, but are the grid dimensions randomly chosen in a biased way where certain sizes are much more likely? For example, a 10x10 Tapa is 40% likely while an 11x5 is 4% likely while a 5x5 is .4% likely? Certainly some puzzle types (Doppelstern) can't go below 9x9 without a serious problem or 8x8 at all, but what is the smallest grid that a surprise puzzle can be? I'm guessing you limit it to 4x4 on the Latin square types, but could something like a 3x3 Hochhaeuser ever appear?

PuzzleScot
Site Admin
Posts: 2755
Joined: Fri 18 Jun, 2010 10:45 pm
Location: Edinburgh, Scotland

Re: croco-puzzle

Post by PuzzleScot » Sat 26 Feb, 2011 6:03 pm

Berni/Calavera, Is it possible to replay old daily puzzles?

Post Reply