Well done, Yes I agree it is very much harder.......
gsf is on the beach for 2 weeks so he may have more thoughts on his return.
I think the latest program from gsf adds ALL the clues in a +2 as opposed to a somehow selected few as it was before - so my search for a minimal stepup takes much longer.
As you say the minimality issue is not addressed with gsfs , it has to be done as a second stage.
I have mentioned to gsf that I think the probability is very high that out there there is a grid with a 38. Also the 37s that we have are far from "full" in terms of number of grid solution per clue.
The grid with the most likely prospect is the PT grid as RW mentioned.
- Code: Select all
123456789457189326689327154231645897745891632896732541318264975574918263962573418
However although I struggled initially to find a 34 in this grid I have now easily with a {-0+2} search option managed to find over 7000 34s and well over 100 35s......which isnt surprising in itself, except that there appears to be no sign of the 34s or 35s tailing off - spreading out throughout the regions of the grid.
A "region" in a grid might well be a maximal pseudopuzzle [with 2 solutions] plus one [solving] clue.
The maximal pseudopuzzle that I have in mind might have 41 clues with the unavoidable set associated with the pseudopuzzle containing 40 clues in this instance.
The resultant region [in effect a non-minimal puzzle] will have 41+1 =42 clues.
These 42 clues [if taken from one of the above 34s or 35s - or other 36s or 37s] we know will have a large minimal puzzle within.
This might be better than a random search [10^9 grids*10^20 regions per grid] !
We might just have to admit defeat on this one !
C
