The Stack

OK, let’s try to analyze this a little. The Next Tetris practice level 19, “Zipper”, has the initial layout pictured here. There are seven rows of garbage squares. Only the bottommost row — the grey one — counts for completing the level, but the gap in that row can only be accessed by fist getting rid of the row above it, which in turn needs the row above that deleted, and so on. The level is supposed to be completed in nine moves, which means adding 36 squares to the board. There are 18 empty spaces, which leaves 18 squares to lump up on top. The board is 10 squares wide, so at most one additional row above the initial garbage can be filled, leaving 8 squares left at the end. One of the big questions here is whether it’s worthwhile to try to complete such an extra row.

Here are the nine pieces, in the order they appear:

(Recall that color is significant. Blocks of the same color fuse together and fall as a unit, while a piece composed of multiple colors will break apart upon first coming to rest. This has some non-obvious consequences. For example, if you drop the first piece here the opposite way up into one of the gaps on this board, it’ll stay the shape it is until you delete the first row, at which point the right side will fall one row farther than the left side, and the red squares on both sides will stick together, forming a unit two squares wide.)

Now, all of the gaps in the initial layout are one square wide. Of the 36 squares that compose our nine pieces, 11 have a horizontal neighbor of the same color no matter how they’re rotated, and thus cannot possibly fit into a one-square-wide gap. Let’s call these “blocker squares”. Another 11 squares are directly attached to blocker squares of the same color. If these are placed in a gap, the adjacent blocker square will have to be above it, preventing access to any further gaps in that column, unless you can complete that extra row and delete it. The remaining 14 squares are detached units or straght columns that can be used to freely fill in whatever needs them, provided they don’t glob onto something else of the same color and get stuck.

The obvious thing to do with the straight piece (or “I” piece) is to put it in the 4-deep straight hole on the right. But obviousness is no guarantee of correctness — the I could also very well fill the vertically-arranged four gaps in the very middle. This works well with the level’s name — the I is like the zipper that pulls downward through the middle and tears the whole structure apart. Also, the I is a color that isn’t used in any of the other pieces, which makes it more useful in the middle, where it helps to keep things from getting stuck to each other. But if you put it there, you’ll have to come up with some other way to fill in the straight gap, and nothing else fits quite as easily. Filling an extra row is almost necessary in such a scheme.

Two other pieces are of special note. The 2-by-2 block (or “O”) is composed entirely of blocker squares, and is of no use whatever for filling gaps in the garbage. It’s also composed of a color almost unused otherwise, so it isn’t much use as a complication either: just throw it to the side and it won’t interfere. This makes me think that it serves in the puzzle to help complete the extra row. At this point, there’s so much circumstantial evidence in favor of the extra row hypothesis that I’ll be really surprised if it’s wrong.

The other is the final piece, the green “S”. Because it’s the last piece, it has to either fill in the final gap, or trigger a cascade that fills it in. On purely aesthetic grounds, I suspect that a final cascade won’t involve only part of the initial layout. Either the garbage above the goal row will all be gone by the time this piece is placed, or it’ll all still be there. In the latter case, there’s some difficulty in placing the S. This scenario requires all but one of the gappy columns to be capped with a tower of pieces waiting to fall into place. The only original gap that could possibly accommodate the S in this situation is the straight pit on the right, which would have to have three pieces in it already and have nothing on its left to keep the S out; this could clearly only happen if an extra row had been deleted. Alternately, the S itself could complete the extra row.

So, I’m coming to the conclusion that there’s probably an extra row to be built and deleted, and the I could plausibly go in the middle. I’ll have to try things along these lines the next time I play.

Well, I’m still stuck on that last practice level. It’s a toughie.

The puzzle title is “Zipper”. It features filled and unfilled squares in a checkerboard pattern, which means that whenever you use a piece that’s more than one column wide (after taking splitting by color into account), you’ll wind up with an overhanging bit that prevents you from filling in the space beneath it. Recall that pieces (or piece fragments) of the same color stick together, potentially fusing into a piece that’s more than one column wide after you’ve placed them. This makes this puzzle unique among the practice levels in that the challenge has a lot to do with not just making the most of the pieces you’ve got, but also finding ways of getting rid of the ones you don’t want.

At one point, I was about ready to give up on this one. I hit Gamefaqs, hoping to find a nudge in the right direction. And indeed there is a guide to the practice levels there! It even features nice gentle hints, describing the general approach you have to take for each puzzle and letting you work out the details yourself. It’s always a relief to find well-written hints. Far too often the hint-writers just give everything away. Unfortunately, the person who wrote these hints couldn’t figure the last puzzle either, so it’s not a lot of use to me.

What an terrible title! Not only is it far from distinctive — a completely different Tetris variant called The New Tetris was released around the same time — it’s also quite arrogant. Or at least it is if you interpret it as “The thing that will be as significant to game culture as Tetris was”, as in “Every game designer hopes that their creation will be the next Tetris“. If you interpret it as “One more Tetris among many”, it’s a bit less so. Under that view, it even implies that the gameplay is so generic that they couldn’t come up with a descriptive title. Which isn’t really true: The Next Tetris delivered some genuine innovation within the Tetris format, subsequently reused in other titles. Instead of making every block completely static once it’s finished falling, staying with its row even when isolated and unsupported, TNT makes blocks only stick together if they’re the same color and fall otherwise. Delete a row, and bits of the row above have the potential to fill in gaps in the row below, potentially making Bejewelled-like cascades. To promote this kind of behavior, the designers decided to make two-color pieces that have the potential to split apart immediately on placement, if placed right.

Doesn’t this all make it easier? Well, yes. Yes it does. Even ignoring the potential for cascades, the frangibility of the pieces means that you don’t need to spend so much time waiting for the straight piece to come up. The designers compensate for the increased ease with horrible, horrible time limits.

But that’s not what I want to talk about. The main game here is, for me, a sideline. The real reason I picked this game up is that I learned Scott Kim was involved in its creation. Although he’s probably still best-known for his invertable calligraphy, Kim is also an accomplished puzzlesmith with about a dozen games under his belt. Among them is Obsidian, which provided me with one of the most beautifully transcendent moments of realization I’ve experienced in a lifetime of gaming. ¹Figuring out the override for the non-regulation flight, in case you’ve played it and were wondering. This memory was still fresh in my mind when TNT came out.

Kim’s specific contribution to TNT is a series of puzzle levels — yes, real puzzles, with designed solutions which you can discover by thinking about them, rather than the random configurations and speed play of normal Tetris and most other so-called “puzzle” games. In each, you’re given a fixed initial board and a fixed sequence of pieces, and you have a figure out how to delete the initial “garbage” in the minimum number of moves. (If you exceed par, the game lets you keep playing, but you know you’ve failed.) Some of them seem to be just a matter of trying different configurations until one works, but the better ones have some gimmick, some underlying principle that makes the whole thing easy when you think of it. For example, in the screenshot included here, the trick is to make the blocks into an inverted copy of the empty space, so that everything just cascades into place when the final piece unlocks the whole structure.

The puzzle levels are not a large part of the game, or one that it draws a great deal of attention to. I suspect that there are people who have played TNT a great deal who don’t even know they exist, just as I ignore multiplayer mode on most games. They’re listed under the main menu as “Practice Mode”, which is a lie: practicing the main game by playing these levels would be like practicing swimming by taking a long shower. It involves the same elements, but applied in a different way. Normal Tetris doesn’t allow you to replay the same situation over and over until you get it right, and that makes a huge difference. So does the certain knowledge that there is, in fact, a solution. You can’t blame the random number generator for your problems. No, you’d have to blame Scott Kim.

There are 19 practice mode levels. In my last session, I solved all but one, although not in order. If it takes me more than one more session to figure out the last, I will be sad.

So, the second year of this blog ends with another unplanned month-long outage. It’s been a pretty dismal year for the blog, with only 14 games knocked off the Stack, if I count correctly. I haven’t even finished the Orange Box yet. This is in large part because of the demands of my new job. (The first month-long outage basically coincided with my the first month of employ.) Don’t get me wrong: it’s a great job, miles better than the one I left to take it. But there have been long hours and tight deadlines, on top of a killer commute. It’s nearly an hour and a half each way by bus, which, unless I switch to a portable system, doesn’t leave a lot of time for gaming. Or, to be more accurate, it leaves a certain amount of time for gaming, but not nearly enough time to both game and write about it. I’ve really got to find quarters closer to the office, but not having a lot of time also means not having a lot of time to look for a new apartment.

And so the Oath has backfired: in order to avoid the obligation of blog, I’ve been playing games that aren’t on the Stack. But I’m not giving up. Now that the most recent tight deadline has passed, I’m going to try to ease myself back into this by writing up some non-stack games.

As for what’s remaining on the Stack, I think it’s about time I made my secret files public. There are two ways to view it. First, at some point in 2008, I discovered Backloggery through a link to this blog from a comment thread. Backloggery is a site devoted to people doing exactly the same thing as me, except with less commentary. I had always assumed that when I wanted to put my list online I was going to have to find or create my own HTML interface to it. Seeing that someone else had done the work already, I entered my entire list, and have maintained it ever since.

I found this solution unsatisfying, though, because it didn’t categorize things the way I wanted them. Backloggery sorts by platform, but not by genre. Their list of game statuses includes several degrees of finishedness (“Beaten”, “Completed”, “Mastered”), but only one unfinished status; I had been tracking only one degree of completion, but had several kinds of non-completion (“untried”, “played partway”, “was unable to complete due to unresolved technical problems”).

Then Gunther Schmidl started his own game backlog blog and showed me what I should have done in the first place: just upload the spreadsheet to Google Documents and make it world-readable. So I’ve done that too. My Backloggery page is here and the Google spreadsheet is here.

You may notice that the Google document has 301 rows, while the Backloggery reports only 299 games unfinished. I always spend a moment confused when I look at them. Well, the spreadsheet has an extra row because of the column headers, while Backloggery is missing Pokémon from the “Unfinished” list: by their standards I’ve beaten it and it would be dishonest of me to list it otherwise. I should try to contact some of the other backloggers with Pokémon on their lists to try to arrange trades. It’s probably my only hope of finding any. (Craigslist was a bust.)

At any rate, that means we currently stand at exactly 300 games listed, which is a satisfyingly round number to start the new year on. Not that this number is really all that meaningful: I’ve got 8 points to spend (that’s $80 worth of new games by the terms of the Oath, which can go quite a long way these days), and there are a number of games whose stack status is iffy. Does Team Fortress 2 count? I did buy it, but only because it came with the Orange Box. I suppose I’ll write it up when I get around to trying it, but it’s not in the list right now. What about Peggle Extreme, also from the OB? I don’t think so: it’s really just a demo, not a full game. Or The Next Tetris — a puzzling thing to be on the Stack, perhaps, as it’s not the sort of game that’s finishable, but it has a finishable component, which is what I’m counting for Stack purposes. Except I can’t for the life of me remember if I ever finished it or not. So it’s on the list just in case.

I’m sure that there are other things on the list that will provoke questions, or at least raise eyebrows. That’s why I was so reluctant to publish the list. Anyway, expect another post tomorrow (I’ve already started writing it), and happy new year.