Monday, May 18, 2020

Crinkly wrinkly shapes with Heptominoes

Seeing as I have an excess of spare time on my hands recently I've been attacking some constructions that I had previously written off as 'too hard' or at the very least on the boundary of what I could do without 'cheating' and using a computer solver. One of these is the following 29x29 rectangle with wrinkly crenelated edges around the outside, and a little 5x5 hole in the centre. And a couple of other little holes too because the harbour heptomino needs something to chow down on.


This wasn't the easiest thing in the world to put together. The first challenge was the outside corners. There are a very limited number of pieces that can occupy those corners, and I only discovered this a little way into the solution - too far in to just backtrack and start again, corners-first. Miraculously I hadn't used up too many of those precious corner pieces when I realised, so I was able to keep on trucking.
The whole perimeter of this was a completer pain in the arse to do. I started out with the left-hand edge, using pieces derived from the C- (or U- if you're weird) pentomino but they ran out fast and by the time I got to the final edge (the right-hand side) I was stuffing whatever pieces I could in, in the desperate hope they'd fit.

And they did. But they didn't leave a nice set of pieces remaining. And they didn't leave a nice shaped internal hole either. Still, I pressed on.

Fig. 2: Approximately what I had left to contend with after completing the perimeter.
From here I filled in the remainder of the space going around clockwise starting from about the 4 o' clock position. In fact, here's a needlessly detailed step-by-step of what I did. I usually refrain from this on the grounds that it's not very interesting, but really, is any of this very interesting?

Firstly, that question mark-shaped piece was just crying out to be put in that question mark-shaped hole, so I did that. Then there were a bunch of really hideous pieces (the other orange ones) that I just felt needed to be used up as early as possible, just so I didn't have to look at them again. Those two sort-of cross shaped bits especially. Winding up with one of those near the end of a solution when you're down to a handful of pieces is the stuff of nightmares.

Then, using up all those long S and L-shaped pieces because they're no fun either. Sadly, in order to do this we had to sacrifice that nice 2x4-with-a-notch-taken-out-of-it piece, but it was for the greater good. And now the edges of the hole in that top-right section are all relatively smooth and straight - ideal for for the very end game when you're solving with relatively squarish blocky pieces.

This is where the going started to get even tougher. Those two red pieces especially, I have a personal vendetta against now. I managed to get them all packed in there - but at what cost? So many potentially useful end-game bits used up in their prime.

Got that top-left corner squared off. And at this point we're left with not the worst hole in the world (that award would have to go to... I don't know, the ozone hole? Or the holes in the end of trombones that allows them to make their noise), but not the easiest to solve either. That two-cell-deep well at the far left was to haunt me for the next hour or so.
And the pieces I was left with to try and plug this hole? Check out this ragtag bunch of misfits:

Admittedly not the worst selection imaginable, but still... especially that one that kind of looks like a number 4 or a heavily-damaged tuning fork (kindly marked in orange for your identifying pleasure), that can suck a fat one. You'd think that piece and the 2-cell well would be a match made in heaven, but try as I might, it just was not to be...

At this stage in the game, there really is no more technique to speak of, no prioritisation of pieces or synergies between two pieces that are painful to deal with on their own but make a well-behaved 14-omino when coupled together. It's just a case of try something then if it fails tear it out and try something else. And that makes solution times vary wildly. Sometimes just through blind luck you'll stumble upon a valid solution after five minutes; other days it's hour after grueling hour trying various configurations in vain.  I guess the factor that determines how long this takes will be how many solutions to this particular endgame there are lurking in the search space as you wander randomly through it.

I got all but one piece in several times as I was doing this one. That's always the worst, that feeling of so near yet so far... And there's another weird thing that happens occasionally when finding a correct solution. It's when you get that 'eureka' moment, that "Yes! I've done it!" spark of excitement a fraction of a second before you've actually understood how the last two or three pieces are going to fit in there. I'd write it off as premature celebration if not for the fact it only ever seems to happen when I've got an actual solution on my hands - maybe I'm subconsciously spotting the solution and the rest of my mind takes that split-second to catch up? 'Tis a mystery.

Here's the eventual configuration I found. A quick consultation of some solving software tells me that there are 9 solutions to this, and that 5 of them do fit that '4'-shaped piece into the two-cell well, so goodness knows how I managed to miss 'em all.

In fact, here's all nine, since I'm being so generous with the pictures today. Mine is number 7.



And while we're on the subject of wicked tough constructions, here's another recent one that might be a contender for 'hardest construction so far', or at the very least most time-consuming:


As I've probably mentioned at some point before, octiamonds are just not a pleasant bunch of shapes to work with.
Sorry, no step-by-step account for this one. Partly because I solved this about two weeks ago so I can't remember the solution process clearly (other than that it was all over the place), and partly because I'd rather not re-live the experience. All I know is I solved the bottom edge first, and the rest followed, four hours later...

No comments:

Post a Comment