Earlier, we discussed the Ace: getting an all-clear on the 10th piece. Why the 10th piece? At 10 pieces, we've played 10*4=40 squares worth of tetrominos, which is enough to fill exactly 4 (4*10=40) rows. Looking at the math, given no damage received, we could also all-clear on the 15th piece and 6th row (15*4 = 6*10 = 60) or 20th piece and 8th row (20*4 = 8*10 = 80).
We also know that each deck in Tetris consists of 7 pieces. If we all-clear on 15, we'll have played two whole decks of pieces, and, given the piece in hand, have our pick of the first two pieces of the third deck. If we all clear on 20, well, three decks = 3*7 = 21. We'll have played three whole decks, minus one piece in hand.
The process of finding an all-clear, whether improvised or planned, is much like solving a jigsaw. We start with one corner filled in, figure out how to fill in the other corner, then fill in the center. Let's think of some strong setups we know which already give us one of the corners.
1. The Jail15
Above is the Zesty Jail setup. We can see that the top left corner is already filled in: the J piece forms a corner at height 6. This is a candidate for a 15. How shall we fill in the right?
Above we see various fills which end up in the same place: two pieces remaining to play, and us already having hit them with a nice T-spin triple for 6 damage. I find it useful to classify these both by what pieces are at the bottom of the build, and by what piece we don't play and end up with in hand. So above we have the O-no-S (x2), O-no-Z (x2), Z-no-I (x2), and S-no-I, all ending up with the same hole. This is a pretty small hole to fill in, and I only see three solutions:
We can solve it with T,O (T<O) T,J (no constraint) and I,L (I<L). Fortunately, three of the above fills don't use the I, so we can just use our I in hand from the second deck. Note: the I-spin is tricky; read up if you're not sure how to do it. Given a no-I fill, we get the all-clear whenever we either have an L in the first two pieces of the third deck, or have TO, TJ, or JT as the first two pieces. In total, we get it 5/14 of the time. But note that only 1/21 of the time do we get a second T-spin, and then only a single. So we're probably doing 16 damage total with this opening, and a small part of the time, 19. 16 won't kill a Long Combo. Do we have something that might?
2. The Izzy20
Above is the Izzy-dubdub setup and second T-spin. We've already done two T-spin doubles for nine total damage and have another one to come. We notice the top right is a filled-in corner at height 8 for possible 20 (4 + the four rows we've already cleared.) Can we fill in the rest?
Well, much like the O fill of the Jail15, we could just put a 4x3 block on the left:
We can also be a bit more ragged here, leaving a staircase:
Filling in the rest, we get:
Again classifying these by which piece of the last deck we don't play, these are the Izzy20 No-S (x2), No-Z (x2), No-L, and No-S3 variants. My favorite of these is the No-L, as this gets us a third T-spin double if we can play the J and O before the T. Note that the No-S1&2 and No-Z variants place the T vertically by the O, requiring both that the O is placed after the T, and the T-spin is only a single.
Let's look at constraints based on these pictures. Note that use of soft-drops and spins means we can evade some stacking constraints, but once we have our No-piece in hand, the ordering constraint becomes a hard-constraint: we can no longer use swapping to evade it.
No-S 1: Hard: T < O, L < Z < J. Somewhat evadable: T|O < I (evadable by O or T&O last)
No-S 2: Hard: T < O, J < Z. Somewhat evadable: T|O < I (evadable by O or T&O last), Z<L (evadable by Z-spin)
No-Z 1: Hard: T < O, S < L. Somewhat evadable: T|O < I (evadable by O or T&O last), J < L|Z (evadable by J<O|I|T and a soft drop)
No-Z 2: Hard: T < O, L < S. Somewhat evadable: T|O < I (evadable by O or T&O last)
No-L: Hard: J < Z. Somewhat evadable: T < S|I (can be evaded by T played last, but we lose the T-spin), O < I (evaded via O > Z|I|S) S|Z < I (evadable via S/Z spins)
No-S 3: Hard: T < I|L, I < O, Z < J. Somewhat evadable: I < Z (evadable via I-spin in one mirror image), I < L (evadable via I < Z|J)
Whew! That's a lot of constraints. Let's figure out some rules of thumb:
- For the No-L, once we've played J, O, and T (any order) we're good to go. Even just J and T works almost all the time, given consistency with S/Z spins.
- For the No-S1&2/No-Z, we're good once we've played T and O, T needing to come before O. Meanwhile we play our 4x3 rectangle in whatever order and with whatever pieces are convenient.
- We can get screwed by needing to play our I early. The otherwise mostly-useless No-S3 can give us a last chance if we have an early T, I.
Back when I first wrote about the Izzy I alluded the the inverted-J variant and its dubdub:
This has many fewer constraints to set up than the above flat Izzy-dubdub. We still have Ju<Z<S (upright J is before the Z which is before the S) but otherwise get to place our second-deck pieces independently. Can we all-clear from here? Indeed, if we just tweak the above No-L fill and instead use No-J:
Third deck constraints: Hard: L < Z. Somewhat evadable: T < S|I (can be evaded by T played last, but we lose the T-spin), O < I (evaded via O > Z|I|S) S|Z < I (evadable via S/Z spins)
We have fewer all-clear possibilities here, but the one we do have is the best one: three T-spin doubles for 14 damage plus an all-clear for 10. 24 lines of damage in the first 20 pieces ain't bad. Of course, we could've killed a Long Combo opening with just 19. Is there a way to get 19 lines off the Izzy and kill at 15?
3. The Izzy15.
Suppose instead of using an inverted, upright, or facedown J after the Izzy, we just leave it faceup and drop it on the right:
We can then fill in a 3x4 rectangle on the right in various ways, to get our right corner at total height 6 (4 high + 2 already cleared lines from our T-spin double):
Filling in the left with our remaining pieces other than the T, we can get to the same hole in various ways:
I only see three solutions to this hole:
Since we have a T in hand, we just need an O or L in the first two pieces of the next deck, and we can do one of these solutions. Note that the LJ solution is pretty useless, since whenever we can do it, we can also do TL for a T-spin plus the all-clear.
It turns out we get O or L as one of the first two pieces of our third deck 11 times out of 21 (1 - 5/7*4/6 = 1 - 20/42 = 22/42 = 11/21), or a bit over half the time. The L gives us a T-spin single, while the O gives us a second T-spin double. Getting the O happens 2/7th of the time, or almost 30%. With the O, we've done 4+5+10 = 19 damage in our first 15 pieces. Not bad.
Also note one of these setups looks a lot like the second-T-spin-at-11 version of the Izzy, with just more stuff on top. If we get the needed pieces early in our second deck--Z,S,O,T--we can just go for our quick T-spin at 11, and while dropping the T, look at the pieces-to-come and see whether we have the O or L available. If we see one of these coming up, we continue on to the all-clear. If we have the L but not the O, we can take the T-spin single and plan ahead.
These aren't the only possible all-clears. Given some practice with this, related ones can easily be improvised based upon your third deck pieces.
For example, the other day I did something like this:
Here I had an I in the beginning of my third deck, so I moved the J and O in the right-JO setup and instead cleared with the J and I in the top right.
Coming up with a related one on the fly, this one uses a Z in the third deck and shifts the second-deck O to clear the central hole:
4. Following up
The good news with a 20 is that you pretty much have a full deck to work with, plus whatever piece in hand. You have your choice of opening play here, but I suggest you usually just do a variant of the Izzy again, and get your followup T-spins before they start clearing your garbage.
With a 15, you've used up one or two pieces of your second deck already. If you used up the T, don't fear to just build a Tetris if you don't feel like a T-spin. It's not like you have a T to play until the next deck anyway. Oh, and don't try to Left-O, OJ, or OL if you used up the O.
5. Use in competitive play
These all-clear setups are prevented by receiving damage before they can be pulled off. So against openings which do not do damage early, such as the Long Combo, they really shine. The 15 and 20 all-clears turn the soft-counter of quick T-spins into a very hard counter a good proportion of the time. The 15 and the 20 also can occur when initial attacks cancel out--as in a quick T-spin mirror match if both players hit at about the same time. But note that most of our attacks in building these are T-spin doubles--so a T-spin triple will still put damage on our board and prevent the all-clear.
Most of these do not need to be explicitly played for. We can follow our usual game plan of attacking as fast as we can with T-spins and Tetrises, and, if the opportunity comes, all-clear as well. But of course if we see our opponent doing something vulnerable to these, such as setting up a Long Combo, we can take a little more time and explicitly use one of the many setups as a counter.
In a metagame sense, these lower the viability of openings which are slower to put damage on the board, such as the Long Combo and the Left-O, while raising the value of openings which do a T-spin triple within the first couple decks, such as the Jail and especially the OJ/OL-trip.
Conclusion
Learning to all-clear at 15 and 20 can be a powerful counter and a strong addition to one's repertoire. Additionally, practicing the various setups can teach one to improvise all-clears in midgame to a greater extent. However, given the penchant of strong players for hitting as hard and fast as possible in early game, do not expect to be able to bust these out in top-level Tetris vs. Tetris too much of the time.
No comments:
Post a Comment