I've managed to cut down the time for a Jelly check to about one minute and for each Sword check to about half a second. So I could check a couple of hardest puzzles within reasonable time.
I checked
coloin's M3 list of 2007-11-20 and the numbers 1, 6, 8, 9 and 10 didn't even require a sword check.
But
col-201107-M3-3 scores 1 jelly & 5 swords and
col-201107-M3-7 even 1 jelly & 11 swords (SE rates it 10.6 BTW).
I also checked
JPF's list of last April and I found 20 puzzles requiring a Jelly check. Funnily of the top three SE 11.4 puzzles only the Easter Monster requires a jelly check. And further down the list there's
coloin's St. Patrick 2 with 1 Jelly & 4 Swords and SE 10.6 next to
tarek's ULTRA0213 which doesn't even require a sword check and rates SE 10.7.
As I said, my results don't exactly match those of
AW. He built some basic solving techniques like Quads and Jellyfish in his solver, while I purely rely on pair implications. But the puzzles that require a Jelly check for my solver can't be solved with a combination of SSTS and Implication Tabling up to Sword check either. And I don't think that my Sword check is less effective than
AW's because he rated
dml 1/07 with 1 Jelly & 17 Swords while my solver only needs 1 Jelly & 12 Swords. (I'd really like to see how many Swords
AW would award the Golden Nugget).
[edit] Houston, we have a problem ...
I just stumbled over
coloin's old waiting list with the following SE 11.4 puzzle:
- Code: Select all
coloin-04/13-1414
1.......2.9.4...5...6...7...5.3.4.......6........58.4...2...6...3...9.8.7.......1
After 7 Jellys & 11 Swords my solver resigned!
AW, I'd really like to have your results on this one. Currently I'm only checking for Jellys with True True assumptions, but at least for the Sword checks I never saw the other possibilities yield new implications when True True failed ...
[edit2] Another one from that list, that won't surrender after 5 Jellys & 8 Swords (SE 11.3):
- Code: Select all
coloin-04/13-1426
1.......2.9.4...5...6...7...5.3.4......96.........8.4...2...6...3...9.8.7.......1