croco-puzzle

[gareth]

can't exactly do this with Nikoli's app, since the notes don't get on the grid lines which is 10x better for spotting things in the important rows/columns, but I've accepted "hold spacebar and hit number" as a pretty friendly interface where it pops up.

!!! I never knew you could do that on Nikoli.com! I've always found it incredibly slow and annoying clicking in each corner to place pencil marks! I never tried holding down space and pressing numbers, and since there are no instructions I probably never would have done.

So thank you. Very useful!!!

[drsteve]

OK, new question

I've just hit 600 points (hurrah!) and was comparing myself to the puzzler below me.

Our scores are:

Played Correct Score
Me 57 56 602
hansauger 87 59 600

So despite my higher score and significantly better percentage correct, he's a 20kyu and my red line is less than half the way to the 100 mark, which I understand is the benchmark.

So my question is basically, huh? How is the red score worked out? As a mathematician, I figure I should be able to understand it, but if it is, as I understand it, an average of the last 200 somethings, surely my red line score should be roughly 3?

[Calavera]

So my question is basically, huh? How is the red score worked out? As a mathematician, I figure I should be able to understand it, but if it is, as I understand it, an average of the last 200 somethings, surely my red line score should be roughly 3?

Just add up the rating of your last 200 in which you participated (602 for now, take 0 for the days before you started) and divide the result by 200. I'd be more than surprised if you reached "3"...

[drsteve]

So my question is basically, huh? How is the red score worked out? As a mathematician, I figure I should be able to understand it, but if it is, as I understand it, an average of the last 200 somethings, surely my red line score should be roughly 3?

Just add up the rating of your last 200 in which you participated (602 for now, take 0 for the days before you started) and divide the result by 200. I'd be more than surprised if you reached "3"...

By the rating, do you mean the points score? Because averaging about +12 per puzzle (which I am) will give me an average score of 12. Is the rating something different?

[motris]

By the rating, do you mean the points score? Because averaging about +12 per puzzle (which I am) will give me an average score of 12. Is the rating something different?

The red rating which determines kyu/Dan is simply a 200 day (for days when you compete) average of your overall score. Right now that score is at 600, so tomorrow the average will improve by 600/200 = 3 points, as you are replacing a value of 0 from 200 days ago. So you have a lot of room to grow as you still have 170 or so days of no value to fill.

Since the other player has played for more days, his/her average ranking over the last 200 days is higher even if the rate of ranking improvement is slower. You'll pass that player eventually if you keep at your current performance.

I believe the figure Berni often cites is it takes ~300 days to reach your true rating, and ~500 days for your red score to reach the same place. Of course you may get better/worse at solving puzzles in the interim, but this is the expected lag to climb the leaderboard.

[motris]

can't exactly do this with Nikoli's app, since the notes don't get on the grid lines which is 10x better for spotting things in the important rows/columns, but I've accepted "hold spacebar and hit number" as a pretty friendly interface where it pops up.

!!! I never knew you could do that on Nikoli.com! I've always found it incredibly slow and annoying clicking in each corner to place pencil marks! I never tried holding down space and pressing numbers, and since there are no instructions I probably never would have done.

So thank you. Very useful!!!

It's sort of in the instructions but not incredibly transparent. Something like: "When the check mark is in the "Memo" button you can put temporary numbers in the four corners of a cell with "Space-key" + "Number-key"."

What I like is the easy functionality - I simply use the thumb to transition to note mode - and that it overwrites everything that was in that box, but if you type accidentally and erase the notes all come back. Short of what I do on paper, I think it's my favorite note entry method. And it is identical on Kakuro and Ripple Effect so if you ever want to make a number note you know how to.

[drsteve]

By the rating, do you mean the points score? Because averaging about +12 per puzzle (which I am) will give me an average score of 12. Is the rating something different?

The red rating which determines kyu/Dan is simply a 200 day (for days when you compete) average of your overall score. Right now that score is at 600, so tomorrow the average will improve by 600/200 = 3 points, as you are replacing a value of 0 from 200 days ago. So you have a lot of room to grow as you still have 170 or so days of no value to fill.

Since the other player has played for more days, his/her average ranking over the last 200 days is higher even if the rate of ranking improvement is slower. You'll pass that player eventually if you keep at your current performance.

I believe the figure Berni often cites is it takes ~300 days to reach your true rating, and ~500 days for your red score to reach the same place. Of course you may get better/worse at solving puzzles in the interim, but this is the expected lag to climb the leaderboard.

Oh - so it's an average each of my ratings at the end of last 200 days. Got you. Cheers for that.

[gareth]

It's sort of in the instructions but not incredibly transparent.

There are instructions? I didn't know that either! Very useful. I think they must have added them at some point since I never found any before. (or maybe I just didn't find them previously). I also just learnt that if you drag to place black squares in Heyawake it automatically chequer-boards them. I did not know that! Ditto (much less usefully) Hitori.

[jbclements]

Heads-up: Firefox 4 RC prevents all keyboard input in croco-puzzle applet

Starting two days ago, I started a puzzle and discovered that the
keyboard inputs were completely ignored by the java applet. I clicked to
give the applet focus (a common problem), but this didn't help. I
finished the puzzle with mouse-only.

I've now switched back and forth from Firefox 3.6.15 to the
Firefox 4 Release Candidate, and I can confirm that under Firefox 4, the
*second* puzzle you solve--and all subsequent puzzles done without
restarting firefox--ignores keyboard input.

Let me be very specific about the sequence of events required:

1) Kill all running Firefox applications
2) Start Firefox 4
3) Open a new tab
4) Go to http://croco-puzzle.de/Ue-Raetsel/ue-raetsel1.php (Nicht angemeldet)
5) Click on Initialisieren (puzzle starts immediately)
6) Note that keyboard input works correctly
7) Click on U-Raetsel 2
8) Click on Initialisieren (again, puzzle starts immediately)
9) Now: no keyboard input works at all! Aah!
10) Downgrade to Firefox 3.6.15 to get applet working again

I should point out that I have *no idea* whether the problem is a bug in Firefox, a bug in the croco-puzzle applet, or simply one of those unspecified-behavior-changing-oops issues, but it may be easier to change croco-puzzle than to change Firefox...

Let me know if there's another place I should be submitting this report; this appears to be the de facto English-language croco-puzzle forum.

Thanks!

[Calavera]

Posting here should suffice as berni is also following this thread.

The problem you described seems to be similar to a problem that was part of the second last (?... the one before the last) firefox update. It was corrected shortly after this by the last firefox update. So it seems to be a bug that is/was part of the firefox developement right now and theoretically they already found a solution for it.

[ronald]

I uhh also have exactly the same problem running Google Chrome for Mac, as discussed before.

[dickoon]

I uhh also have exactly the same problem running Google Chrome for Mac, as discussed before.

FWIW, same here, though it's fine using Chrome on a PC running Vista.

[ronald]

also no problem running Safari for Mac.

[jbclements]

@calavera, "they fixed it before so they can/should fix it again"; unfortunately, I'm not nearly as confident as you. In particular, it strikes me as entirely possible here that the existing croco-puzzle applet depends on unspecified behavior that's been changed to other unspecified behavior or actually specified in some way that the existing java implementation is not. So... I think that a fix is more likely to come from Croco-puzzle than from Firefox. Especially since Mozilla is now in the *late* stages of release for Firefox 4. I think it's very likely that Firefox 4 will ship with (what we see as the) bad behavior.

john "crossing fingers but without much optimism" clements

[Calavera]

Well, as the bug appeared the first time (other browsers aside) with Firefox 3.6.14 and no longer appeared with the next version, published very shortly afterwards, I'm still hopeful. Why change something and then change it back directly unless there's a reason!? So I hope the reason was that the bug was Firefox-related and recognized and Firefox 4 is still based on a previous build containing this bug.
Also, programming based on unspecified behavior doesn't really sound like berni...

Let's hope for the best,

Calavera

[PuzzleScot]

Worst luck possible on great solves.
UR 1) Faster than Calavera! ... but missed one star - grrr!
UR 2) 2'35 - hooray! Hang on, scoresheet not loading. Damned Internet connection lost, as well as my time!!
Pity the applet doesn't remember my time and submit it next time there IS a connection...

I may need some time off puzzles - they're meant to be relaxing, but they're stressing me, and not for the first time.

[Calavera]

I may need some time off puzzles - they're meant to be relaxing, but they're stressing me, and not for the first time.

Not looking at the rating list might help. I haven't looked there since last november or so and it still feels better. Although the last few days there were some ... suboptimal puzzles. But as I don't know how much they cost me I hardly care .

[ronald]

If I beat Calavera on one puzzle then it's a good day for me
If I beat Calavera on two puzzles then it's a good day for me AND a bad day for Nils

I have had several suboptimal days in the last week or two. My tagesrating has fallen precipitously. And so DavidMcN and detuned closed some of the gap congratulations today to Tom on reaching 16 kyu.
But - must look on the bright side: my plan is to catch sf2l in April

[berni]

Damned Internet connection lost, as well as my time!!
Pity the applet doesn't remember my time and submit it next time there IS a connection...

The applet only looses the time, when you end it (which means you leave the website or close the browser). If sending the time does not work it will show a warning (with red letters) and instructions to press "Testen" once more after the connection is there again to submit it again. I've got plans to change the applet, making it submitting automatically after a certain time period. I've got also an other idea, how to avoid this dataloss, but I'm not sure if it works - still need to think about it.

Concerning Firefox 4: Someone told me, that the same problem appears with other applets too. So I don't think, that there is anything I can do.

[PuzzleScot]

If sending the time does not work...

Yes, time was in red text, but I didn't read/understand the text. I'd already clicked on highscoretabellen to see how well I'd done, then the page didn't load and that was it. It must be a rare event for this to happen to anyone, and as Calavera points out, it should be for fun, not competitive, so if the applet stays the way it is, there won't be any complaints.

To put it in perspective, I would be delighted if we could host a daily puzzle or 2 on our website, regardless of scores/speed/rating, so what you have done is really quite extraordinary and very much appreciated by us all!

[GaS]

Someone could please help me in yesterday's ABCD-Radstel? It's the first time I tried this type of puzzle (as I remember) and I can place no letter at all...

Thanks very much,
Gabriele

[Calavera]

Sure...

First thing to look at are always the greater numbers. In this case horizontally "3" and vertically "2". When there is an odd number of columns/rows this often places the first letters automatically.
In this case when looking at the horizontal "3"s we have to use the help of column information. Most promising is the A: There are only 4 columns in which an A appears and as same letters must not touch each other orthogonally (diagonally was allowed this time) it should be clear where to place three As in row 3. This also leaves only one field for the remaining A.
The "3" of the letter B didn't help as at first, but now that we placed an A in the first row we can also locate one B to the far right. In the row with the 3 As a D can be placed now for the same reasons.
Now as a last hint something a little bit more advanced: After the letters we placed in the first row there are now 2 Bs left and 4 available fields. In the second row there are 5 available fields for 2 Bs (the last one isn't available because of the neighbor rule). Now if the second row wouldn't contain a B in the second last column the 4 Bs of these two rows had to be placed in a zigzag, placing one B in each of the first 4 columns. That would make the placement of the 3 Bs in the last row impossible. So the second last column of the second row must contain a B.

That's probably not the easiest path after my start, but I wanted to give an example about what thoughts are possible.

[GaS]

Now it's all clear, thanks very much!
I was not aware that adjacent cells should not contain identical letters. I can't read german and so I just saw images in "anleitung" section of the puzzle without more information and so I thought, erroneously, I understood the instructions, my fault.

Best regards,
GaS

[sf2l]

If I beat Calavera on one puzzle then it's a good day for me
If I beat Calavera on two puzzles then it's a good day for me AND a bad day for Nils

I have had several suboptimal days in the last week or two. My tagesrating has fallen precipitously. And so DavidMcN and detuned closed some of the gap congratulations today to Tom on reaching 16 kyu.
But - must look on the bright side: my plan is to catch sf2l in April

won't be too difficult, as I have just hit the wall nose-on

[detuned]

I think my gap closing is entirely dependent on the puzzle distribution. If the puzzle is very easy I think I have a decent raw speed, but otherwise there are some puzzle types that I'm clearly still missing a few tricks on. Thankfully Masyu isn't one of them!

On a more sobering note - Thomas is already 17th overall, and still climbing steadily (although this probably shouldn't be such a surprise as I'm sure he mentioned on his blog the other week his tagesrating was round about 2700!) It'll be interesting to see how close things are when everything settles down. If I can stay this close to David I'll definitely be chuffed!

btw Nils, these walk-throughs you're posting are much appreciated, it's always fascinating to see how others go about their puzzle solving!