The New Sudoku Players' Forum

[m_b_metcalf]

Code: Select all

 . . . . . . . . .
 . . 9 . 2 . 3 . .
 . 4 . 5 . 6 . 7 .
 . . 6 . 8 . 7 . .
 . 7 . 1 6 3 . 5 .
 . . 3 . 4 . 2 . .
 . 6 . 4 . 8 . 1 .
 . . 8 . 1 . 4 . .
 . . . . . . . . . SE 8.7

Good morning,

Mike Metcalf

[gsf]

Code: Select all

. . . | . . . | . . .
. . 5 | . 2 . | 3 . .
. 2 . | 6 . 1 | . 8 .
------+-------+------
. . 2 | . 5 . | 7 . .
. 5 . | 2 6 7 | . 3 .
. . 8 | . 9 . | 4 . .
------+-------+------
. 3 . | 4 . 5 | . 6 .
. . 4 | . 8 . | 9 . .
. . . | . . . | . . .  ER=6.5


[m_b_metcalf]

Code: Select all

 . . . . . . . . .
 . . 6 . 7 . 9 . .
 . 1 . 2 . 3 . 4 .
 . . 1 . 9 . 8 . .
 . 9 . 8 3 2 . 5 .
 . . 4 . 6 . 7 . .
 . 7 . 1 . 9 . 3 .
 . . 3 . 8 . 1 . .
 . . . . . . . . . SE 5.8


Regards,

Mike Metcalf

[TTHsieh]

Code: Select all

 . . . | . . . | . . .
 . . 1 | . 2 . | 3 . .
 . 2 . | 1 . 4 | . 5 .
 - - - + - - - + - - -
 . . 2 | . 6 . | 5 . .
 . 4 . | 7 5 8 | . 3 .
 . . 9 | . 4 . | 7 . .
 - - - + - - - + - - -
 . 8 . | 2 . 5 | . 1 .
 . . 7 | . 1 . | 9 . .
 . . . | . . . | . . .   ER=4.7



TTHsieh

[TTHsieh]

start: The referee selects a dealer and posts a message that the game will start in 8 hours.

start: The referee selects a dealer and posts a message including the game pattern that the game will start in 8 hours.

TTHsieh

[gsf]

start: The referee selects a dealer and posts a message that the game will start in 8 hours.

start: The referee selects a dealer and posts a message including the game pattern that the game will start in 8 hours.

TTHsieh

this would change the game, probably for the worse w.r.t human participants
it would allow everyone to precompute submissions
this might be the good part about it because it would put the dealer at the same (dis)advantage as the other players

are there any dealers who have precomputed before others have seen the deal?
the fairness streak in me thinks that dealers should not use precomputed data

for my part when I search for new patterns to deal there is some precomputation to
get a feel for how hard the pattern would be to generate to, but that data is thrown
out before game time and not used during play

and wouldn't this reduce the games to lightning rounds?
I also think it would skew the pearl strategy, which I'm still trying to flesh out
(e.g., how long to hold on to a non-pearl submission in hopes of finding a pearl for it
before someone else grabs the rating, pearl or not)

[JPF]

Congratulations TTHsieh for the score.

The game is becoming a race for professional players.

Even with TTHsieh's proposal it's going to be a rush to post a maximum of puzzles after the 8 hours-delay.

The ability to post a puzzle as quickly as possible was not the initial purpose of the game and is of no interest IMO.

I think we need to revisit the rules (again).

JPF

[gsf]

The ability to post a puzzle as quickly as possible was not the initial purpose of the game and is of no interest IMO.

I think we need to revisit the rules (again).

agreed

how about minimum time between posts for each player, say 10 min

also, strong pearls seem fairly scarce
game 0011 only had 9 { 7.2 7.3 7.4 8.3 8.4 8.5 8.8 8.9 9.0 }

how about treating the first strong pearl for a rating as a trump
i.e., it counts even if a non-strong pearl has already been posted for that rating

ratings above 9.2 also seem scarce
maybe a weighted bonus for ratings > some value?

increasing the score for "harder" puzzles might make it interesting (and level the scoring field)
for niche players that might be able to make a few good submissions per game

[JPF]

Code: Select all

 . . . | . . . | . . .
 . . 1 | . 2 . | 3 . .
 . 2 . | 4 . 3 | . 5 .
-------+-------+-------
 . . 2 | . 6 . | 7 . .
 . 6 . | 2 7 8 | . 3 .
 . . 8 | . 4 . | 5 . .
-------+-------+-------
 . 7 . | 5 . 6 | . 4 .
 . . 5 | . 3 . | 1 . .
 . . . | . . . | . . .      ER=4.699999999999999



JPF

[m_b_metcalf]

how about minimum time between posts for each player, say 10 min?

Sounds a good way to slow down the game in its initial phase and to even out GHz ratings.

also, strong pearls seem fairly scarce
game 0011 only had 9 { 7.2 7.3 7.4 8.3 8.4 8.5 8.8 8.9 9.0 }

how about treating the first strong pearl for a rating as a trump
i.e., it counts even if a non-strong pearl has already been posted for that rating?

A good idea, but how would it work in practice? It's possible that someone posts a pearl without being aware of it and a subsequent deliberate pearl posting would inadvertently become a duplicate.

ratings above 9.2 also seem scarce
maybe a weighted bonus for ratings > some value?

submission ordinal + 100 x (rating - 9.0) ?

Are we getting close to version 2? And should we switch to the latest version of SE at that time (for consistency within a thread)?

Regards,

Mike Metcalf

[JPF]

Some thoughts :

No dealer (g.r.emlin is the dealer)

The winner could be the player with the last puzzle [with an obligation for each player to post a minimum of puzzles/day to stay in the game].
A pearl would be equivalent to 2 or 4 puzzles ?

We should take some time between 2 games to understand why some ratings are missing.

JPF

[tarek]

How about:


2 consecutive submissions by the same player: minimum 4 hours
2 sunbmissions with 1 intervening submission by another palyer: minimum 10 minutes
2 sunbmissions with more than 1 intervening submissions by other palyers: no minimum time

tarek

[gsf]

Code: Select all

. . . | . . . | . . .
. . 8 | . 2 . | 3 . .
. 4 . | 1 . 5 | . 6 .
------+-------+------
. . 3 | . 6 . | 1 . .
. 7 . | 2 5 8 | . 3 .
. . 4 | . 3 . | 5 . .
------+-------+------
. 5 . | 6 . 3 | . 1 .
. . 7 | . 4 . | 8 . .
. . . | . . . | . . .  ER=5.2


[TTHsieh]

Code: Select all

 . . . | . . . | . . .
 . . 1 | . 2 . | 3 . .
 . 2 . | 3 . 4 | . 5 .
 - - - + - - - + - - -
 . . 6 | . 3 . | 7 . .
 . 3 . | 8 7 2 | . 4 .
 . . 2 | . 9 . | 8 . .
 - - - + - - - + - - -
 . 6 . | 2 . 9 | . 8 .
 . . 7 | . 6 . | 9 . .
 . . . | . . . | . . .   ER=9.2



TTHsieh

[gsf]

there is no incentive for a player to hold on to puzzles
this leads to lightning rounds that have a bias towards quantity over quality
I think pearls might provide the incentive

puzzles would be submitted with both an ER and pearl index (PI=.. or PEARL=..)
the referee would keep track of the highest pearl index for each rating
the pearl indexes would be listed in the status so players can keep track

add one more rule:
any player may trump another player's submission with a higher pearl index
(players may not trump their own submissions)
where trump means any previous scores for that rating are nullified

a player that submits early and/or fast risks being trumped
a player that hoards risks being stuck with unplayable submissions if the game goes idle
this may also skew strategy to quality over quantity
and would definitely place a premium on strong pearls