The New Sudoku Players' Forum

[Patrice]

::: comment :::
I'm using for the first time a mixed rating with my own rating engine and skfr from champagne. That's seems to be promising !
And for the generation I'm using a seed selection which is very near from mike's stocastic generation...

[g.r.emlin]

::: game 0298 status LIGHTNING 2017-12-09+09:51:35-0000 :::

Code: Select all

game 0298 2017-12-09+05:00:00-0000 LIGHTNING duration 4h26m dealer papy999 pages 2334-2335 (open) 4 players
_____________________________________________________________________________________________________________

    1.5/1.2/1.2 papy999        9.2/9.2/9.2 champagne    10.6/10.6/9.9 m_b_metcalf 11.2/11.2/10.0 Patrice
    9.1/9.1/9.1 Patrice        9.3/9.3/9.3 Patrice     11.1/11.1/10.0 Patrice     11.4/11.4/10.0 champagne
_____________________________________________________________________________________________________________

    rare ratings : 11.4

    open ratings : 1.2 1.7 2.0 2.3 2.5 2.6 2.8 3.0 3.2 3.4 3.6 3.8 4.0 4.2 4.4-5.2 5.4 5.6-6.2 6.5-9.0 9.4-10.5 10.7-11.0 11.3 11.6-11.8

       submitter  entries  score  response  notes
     -----------  -------  -----  --------  -----
         papy999        1    -19         0  dealer
     m_b_metcalf        1    126    59.00s  pearl
       champagne        2    430    43m29s  rare-0*2,diamond,pearl,highest-rating=80
         Patrice        4    689    22m44s  rare-0,rare-1,diamond*2,pearl*2,last-entry=54
     -----------  -------  -----            -----
               4        8   1226            rare-0*3,rare-1,diamond*3,pearl*4


[champagne]

ED=11.2/11.2/10.0

Hi Patrice,

not a surprise, I have now 3 of them

[g.r.emlin]

::: rarity :::

Code: Select all

2017-12-09+09:10:00-0000 11.1/11.1/10.0 Patrice rare ED seen once before
2017-12-09+09:22:00-0000 11.4/11.4/10.0 champagne rare ED not seen before
2017-12-09+09:26:00-0000 11.2/11.2/10.0 Patrice rare ED not seen before


[papy999]

::: comment :::
Well done, champagne !
I found this one when I proposed this pattern

Code: Select all

. . 1 | . . . | . 2 .
. 3 . | 4 . . | 5 . .
6 . . | . . . | . . 7
------+-------+------
. 8 . | . 3 . | 9 . .
. . . | 9 5 . | 8 . 4
. . . | . . . | . 6 .
------+-------+------
. 9 . | 5 4 . | . . .
2 . . | . . 1 | . . .
. . 7 | . 8 . | . . .  ED=11.4/11.4/10.0

It's the same !!!
I think there are no higher rating ones !

[champagne]

::: comment :::

Well done, champagne !
I found this one when I proposed this pattern...
It's the same !!!
I think there are no higher rating ones !

And I assume that most of the high ratings found here are already in the data base of potential hardest.
Patrice wrote recently in a pm that one 22 clues pattern has many sudokus in this base and I suspect that this is the pattern.

The 22 clues field has been long searched, most of the potential hardest should be known.

[Patrice]

::: comment :::
I just found this ED=11.4/11.4/10.0
It was number 6178 in my Database, so it came very early...
but skfr rated it as ED=11.2/11.2/9.7 !!! and there was 8 sudoku with the same skfr rating
Serate gives these rating

Code: Select all

..1....2..3.4..5..6.......7.8..3.9.....95.4.8.......6..9.34....7....1.....2.8....  ED=11.1/11.1/10.0
..1....2..3.4..5..6.......7.5..3.8.....98.4.5.......6..9.34....7....1.....2.5....  ED=11.2/11.2/9.9
..1....2..3.4..5..6.......7.5..3.8.....89.4.5.......6..9.34....7....1.....2.5....  ED=11.2/11.2/9.9
..1....2..3.4..5..6.......7.8..3.9.....59.4.8.......6..5.34....7....1.....2.8....  ED=11.2/11.2/9.9
..1....2..3.4..5..6.......7.8..3.9.....95.4.8.......6..5.34....7....1.....2.8....  ED=11.2/11.2/9.9
..1....2..3.4..5..6.......7.8..3.9.....59.4.8.......6..9.34....7....1.....2.8....  ED=11.2/11.2/10.0
..1....2..3.4..5..6.......7.8..5.3.....93.8.4.......6..9.54....7....1.....2.8....  ED=11.2/11.2/10.0
..1....2..3.4..5..6.......7.8..3.9.....95.8.4.......6..9.54....2....1.....7.8....  ED=11.4/11.4/10.0


So using skfr is not so easy, you have to adapt to possible +-0.2 deviance between skfr and serate !

[Patrice]

::: comment :::
champagne
What is the maximum deviance you found between skfr and serate ?

[g.r.emlin]

::: adjustment 2017-12-09+10:55:00-0000 Patrice consecutive-entry -50 :::

[champagne]

What is the maximum deviance you found between skfr and serate ?

I would say +0.1/-0.3, except may be in places where skfr has a bug not investigated (in the high ratings). I use mainly another clone, faster but with a +/- 0.2 deviation in the high ratings

[Patrice]

::: comment :::

Patrice wrote recently in a pm that one 22 clues pattern has many sudokus in this base and I suspect that this is the pattern.

champagne,
I've found now 37 sudokus with this pattern with rating >=11.0
I've verified : it's an isomorph pattern of the 22 clues pattern that I mentionned in my PM...
and there was 37 sudokus (rating >=11.0) in your database with this pattern !!!
EXACTLY THE SAME NUMBER !!!

When this game will be over, I will verify if all the found 11.x are in your "potential highest sudoku database".
So I invite all players to send me by PM their 11.x at the end of the game.

[champagne]

Hi Patrice,

apparently, I have currently 44 puzzles rating 11.x.

[Patrice]

::: comment :::
skfr rating or serate rating ?

[champagne]

I only store serate ratings in the game

[champagne]

Code: Select all

..1....2.
.3.4..5..
6.......7
.8..4.9..
...58.4.3
.......6.
.9.85....
7....1...
..2.3.... ED=10.7/10.7/9.3