The New Sudoku Players' Forum

[blue]

my results for the 134 puzzles in te3:

Code: Select all

4tp:  18
5tp: 115
6tp:   1 : (26)


Same here.

[P.O.]

thanks for checking blue.

[denis_berthier]

.
My implementation of the rules defined at the start of this thread were correct, but there was something missing for level T3: the elimination of a template[2] didn't lead to all the eliminations of templates[3] extending it. This is now corrected. I don't have time now, but I'll publish the additions tomorrow.

As a result:
- any puzzle I previously found at template-depth T0, T1, T2 and T3 is unchanged;
- any puzzle I previously found at template-depth T4 is now at T3 (and computed much faster).
.

[eleven]

My guess is, that it is provable, that 3-templates (strong version) can solve any sudoku.
Probably Mathimagics would be interested trying it, but unfortunately i cannot talk to the deceased ones.

[denis_berthier]

.
Good luck with this (I mean prove your guess, not talk to the dead).

Something possibly easier: every puzzle in T1 is in T&E(1).

Note that:
- all the known T&E(3) puzzles are in T3;
- all the tested T&E(2) puzzles are in T3. The list of tested puzzles is not very long ( a few thousands) but I'd bet there are none not in T3 among the known ones.

The results for the first 254,760 cbg puzzles, restricted to the 165,272 of them in T&E(1), show that T&E(1) spreads over all T1, T2, T3, with the largest part in T3:
T1 : 43551 =. 26.35%
T2 : 22141 = 13.40%
T3 : 99580 = 60.25%

With such results, my current view is, the Tn hierarchy is useless for puzzles in T&E(2 or 3). Whether is brings anything useful to the analysis of T&E(1) puzzles is as yet undecided.
.

[denis_berthier]

.
I think I found the first puzzle (really) in T4; it is in eleven's pre-tridagon T&E(2) tamagotchi collection:

Code: Select all

12.....8...7..9..66...2.1.....93...4..4.7.9.......4.6...6..7..3.......5.8..5.....  11.3 11.3 10.6

The T4 classification resists:
- 100 random relabelings
- 50 random full isomorphisms.

Blue, would you like to try it with your implementation?
.
[Edit: added blue colour.]

[eleven]

Hehe, a stable conjecture by me is, that at least 50% of my sudoku conjectures are wrong

[coloin]

.
I think I found the first puzzle (really) in T4; it is in eleven's pre-tridagon T&E(2) tamagotchi collection:

closely related is this puzzle .. sensors indicate its got similar template properties

Code: Select all

12.....8...7..9..66...2.1.....93...4..4.7.9.......4.6...6..7..3.......5.8..5..... ED=11.3/11.3/10.6### 98509 FNBP C23.m/M2.2.72160  ##rating:   3908 ,    1431 ,
12.....8...7..9..6....2.1.....93...4..4.7.9.......4.6...6..7..378.....5....5..... ED=11.3/11.3/2.6 ### 99367 FNBP C23.m/M3.361.1472 ##rating:   4012 ,    2242 ,


[blue]

.
I think I found the first puzzle (really) in T4; it is in eleven's pre-tridagon T&E(2) tamagotchi collection:

Code: Select all

12.....8...7..9..66...2.1.....93...4..4.7.9.......4.6...6..7..3.......5.8..5.....  11.3 11.3 10.6

The T4 classification resists:
- 100 random relabelings
- 50 random full isomorphisms.

Blue, would you like to try it with your implementation?
.

Yes, T4
Nice.

closely related is this puzzle .. sensors indicate its got similar template properties

Code: Select all

12.....8...7..9..66...2.1.....93...4..4.7.9.......4.6...6..7..3.......5.8..5..... ED=11.3/11.3/10.6### 98509 FNBP C23.m/M2.2.72160  ##rating:   3908 ,    1431 ,
12.....8...7..9..6....2.1.....93...4..4.7.9.......4.6...6..7..378.....5....5..... ED=11.3/11.3/2.6 ### 99367 FNBP C23.m/M3.361.1472 ##rating:   4012 ,    2242 ,


I have the first in T4 and the 2nd in T3.

[denis_berthier]

closely related is this puzzle .. sensors indicate its got similar template properties

Code: Select all

12.....8...7..9..66...2.1.....93...4..4.7.9.......4.6...6..7..3.......5.8..5..... ED=11.3/11.3/10.6### 98509 FNBP C23.m/M2.2.72160  ##rating:   3908 ,    1431 ,
12.....8...7..9..6....2.1.....93...4..4.7.9.......4.6...6..7..378.....5....5..... ED=11.3/11.3/2.6 ### 99367 FNBP C23.m/M3.361.1472 ##rating:   4012 ,    2242 ,


I have the first in T4 and the 2nd in T3.

Same.

[denis_berthier]

.
I think I found the first puzzle (really) in T4; it is in eleven's pre-tridagon T&E(2) tamagotchi collection:

closely related is this puzzle .. sensors indicate its got similar template properties ...

Could you say more about those "sensors"? Do you have any clues that suggest two puzzles are close wrt templates?
.

[coloin]

Could you say more about those "sensors"? Do you have any clues that suggest two puzzles are close wrt templates?

its a quote by "spock" on early Startrek ...
but its based on this puzzle which has similar very high -q2 and suexratt ... and higher than eleven's original T4...
Maybe P.O.'s analysis will show it as "t6"

[P.O.]

Maybe P.O.'s analysis will show it as "t6"

i classify this puzzle in 6-template, not T6, the Tn classification refers to Denis’s template implementation.

[blue]

Maybe P.O.'s analysis will show it as "t6"

i classify this puzzle in 6-template, not T6, the Tn classification refers to Denis’s template implementation.

Oh no.
I get "5-template" for the last 3 puzzles.

BTW: I'm sure that Colin used lower case "t6" to mean "6-template", not "T6".

[P.O.]

3 last puzzles? what puzzles are you talking about?
if Coloin means 6-template why doesn't he say it?