Given the increasing difficulty of finding a good match, probably due to the age of the game, the start of the school year, various people topping out the ladder ratings, Nintendo Online's switch to paid, and the time consuming travails of my own life, I've decided to hang up Puyo Puyo Tetris for now. Thanks to all the players who inspired me with amazing gameplay in our matches, especially Jonathon, Wumbo, T-block, Karaage (から~げ), and iljain.
As a final parting note, here's various writings of mine that never made it to this blog, not being entirely finished, edited, or diagrammed. If any particular one looks useful to you, drop me a line either in a comment on this post or at "robert <dot> tetris <at> gmail <dot> com" and let me know, and I may finish, edit, and diagram it just for you.
-----
currentmetaopenings-izzy-fd-j
Openings in the Current Meta: Izzy with the Facedown-J
One thing that surprised me coming back from my Switch-repair-induced hiatus was the newfound popularity of certain openings and some particular followups, especially among some of the higher-ranked Japanese players. While this one isn't my favorite, I can't deny its effectiveness.
We open with an Izzy with the J facedown, rather than placing it upright or faceup on the right. From there, we can easily follow up with a T-spin on the left followed either by a quick Trips or a powerful repeated T-spin pattern. Let's take a deep dive.
First deck
Second deck
Above we see two different variants, differing only by whether we have an L or a J in hand. We should note that the left variant, with the J in hand, gives a pretty high chance of an all-clear: we just need an I in the first two pieces, which happens 2 times out of 7. With the L in hand on the variant on the right, we need both the I and J at the beginning of our third deck, which happens only one time out of 27.
(example splits above paragraph. put next sentence before split)
The constraints on these are pretty easy: either O < L or J > O. Between these two possibilities, we can do this pretty much all the time, and if we have to hold an early T and piece swapping can't get us O < L or J > O, we can still just drop the L or J instead of holding it and still essentially get this position.
I should note another variant that can be done starting from the facedown-J, powerful and still reasonably common: the Izzy-dubdub:
This is much more constrained than the above, with constraints J < Z < S and O < L, so even if this would be better against a given opponent, we still will end up with one of the above variations much more often. I've discussed both the all-clear and the repeat T-spin followup to this elsewhere, so I'll forego further discussion of it for now.
(coming from the facedown J this is...)
(linkes to izzy-dubdub, izzy20, and repeat T-spin patterns)
Beyond
Assuming we didn't do the dubdub and don't just all-clear at 15, we take the T-spin, have a J or L in hand, and we're here:
Beyond
The Trips
We notice that we have a one-high columns 1 and 2 and an empty column 3. So we can just place a Z and L on top and do the Trips:
Note that we need to place at least X squares = blah pieces before we can play our two T-spin triples: a deck and Y pieces. So unless our fourth-deck T comes ath-7th in the deck, we'll end up having to play a T in our build-up. Thus we shouldn't fear playing a T if it's convenient for building, since we may be forced to do so anyway, but should realize we may not be able to play both T-spins until deck 5.
Even so, this bops some popular opening sequences such as Left-O -> Long Combo pretty hard. Before they build high enough to be worth breaking, we can hit them with 14 damage from two T-spin triples... and building only to height 6 so as not to be vulnerable to this just ain't a very powerful combo.
Perhaps this opening's ascendance is part of the formerly incredibly-common and still somewhat frequently-seen opening's decline.
Of course, this is mostly excellent if we've cancelled out their initial attacks with our T-spins, or our opponent has otherwise failed to hit us. If we've traded hits or they have big damage in reserve, then the 6+ lines we have to build before breaking put us in danger of topping out before we can break. Our next alternative, while not quite as excellent at bopping certain openings, lacks this vulnerability and still delivers excellent damage.
Repeat T-spins
We can do a repeated T-spin pattern in column 3 (or column 8 in the mirror image), alternating spinning the T on a L and a S on the left. This quickly builds up our board to the point where we can do a T-spin + Tetris combo attack, and continue onward from there.
This is strong both in TvT and TvP, giving a rapid series of attacks into a larger attack. Additionally, if we need to counterpunch, we can at almost any point clear with a piece in column 3 then start our garbage breaking combo.
Analysis
bah, first
Comparison with other Izzy-alternatives
What we've given up is the high-probability all-clears of the upright-J and faceup-J variants. We still get to all-clear at 15 with some probability, however, and we gain superior followups on miss--the Trips gives us an excellent 14 damage and hard-counters many once-popular openings, and our transition to repeat T-spins is smooth, high on continuous damage, and easily transitioned out of.
Enough talk! Let's grade:
Head-to-head
Conclusion
The power and consistency of the facedown-J variant do much to explain its current popularity. Additionally, its ability to specifically counter various popular openings give it great strength against both the metagame of months ago and our current times. However, its relative all-clear weakness versus other Izzy variants may give us pause, especially in top-level matches and against opponents a quick all-clear is likely to be extremely effective against.
-----
puyo-as-seen-on-ladder
Disclaimer: I've actually read some of what has been written about Puyo Puyo, so I can't guarantee a fully pristine, original perspective here, unlike my Tetris writings.
Playing Tetris versus Puyo on ladder, one often sees fairly consistent formulaic play by Puyo players, especially at the not-quite-top level. Observing this, we can derive fairly consistent "mindless pattern" play, which can be useful to those of us who don't want to learn too many intracacies, but still want to play as Puyo reasonably well in Swap and elsewhere.
Overall Plan
The most common approach I see is to create a full screen combo, going in the general direction of being triggered in the top left, going down through the bottom left corner, and breaking our right side via sandwiches into staircases. Something like this:
We build this around our bottom left bridge, adding drop-tower chains on the left side head and left-to-right staircase drop endings on our right side tail. Generally the first thing built is our bottom left corner bridge. The most common shape is this:
I hear tell this is called by many English speakers a GTR. I'll call it a guitar henceforth, since I prefer pronounceable two-syllable words to meaningless three-letter acronyms, and it looks somewhat guitar-shaped.
Initial build
Our goal is to build our guitar as fast as possible and start adding to our left-side head chain, making our chain resilient against small attacks. The most common two pieces set we start with looks something like this:
Here, both pieces are bichrome, and they share precisely one color. Given that, we place these to build the bottom of our guitar like so:
Using this placement rather than, for example, building the left side of the guitar first, allows us the highest probability of getting into our head building immediately rather than being bottlenecked on finding just the right piece.
We now have the two colors of our guitar defined, in this case, blue on bottom, green on top. Our next piece with a green-plus-notblue goes here:
From there it's really easy to build the rest, potentially just shoving our not-quite-right puyos into column three while putting blues and greens in the right places. The important thing is that we've now built enough that we can start adding to the head. For instance:
Now the starter is robust to early damage. It will take a fairly large hit to bury the (color) which will lead into our green then blue breaking in our guitar.
The bottom bridge.
Most likely, we build aspects of this as we set up our initial guitar and part of our head chain. We notice that after the blue in our guitar breaks, a sandwich can activate in column 3. Something like this:
This leads into some puyos breaking in column 4, then 5, which will trigger any left-to-right staircases we've set up in 4-5 or 5-6, however high up. So our chain can span the entire screen even if there's a bunch of poorly placed pieces in our right side.
The right side:
We can pretty much do what we want here. As a general rule, we place like colors close together while avoiding joining 4, and place the left side of a color higher than the right. If we do this, then voila, everything is likely to magically break once the bottom bridge goes.
Why left-to-right?
For some reason, the left-to-right chain is seen much more than the right-to-left chain on the ladder. We may theorize why this is so. The background here is that in Puyo, pieces come out in column 3 and only column 3 kills you, so we must be careful not to fill it high up unnecessarily lest we die.
We notice above that our left-side drop-chain can easily be only two-wide and still get well high enough to be robust against small attacks. We also see that we have a lot of freedom on the right side to place pieces how we wish in a left-to-right build. We can therefore use this space to build smaller poke attacks (which can sometimes be useful in PvP) that will still likely end up adding to our main chain if we trigger it. Between these two things, our column 3 can be kept more empty, giving us greater survivability versus medium size attacks.
Other starts
The most common start is two bichrome pieces with a matching color. But you'll often encounter starts different from these, and it's wise to map out a brainless decision tree about what to do with the first few pieces. Note that while it's possible to consider our first 3 pieces in our decision tree, since we see two pieces ahead once our current one is dropping, I've written this to require only knowing the next piece. This is because we should almost always be quick-dropping so as to play at maximum speed, which makes it difficult to factor in two pieces ahead before our piece is almost placed.
(graphics) four of a color
Here we just break these right away for an all-clear. Our next attack will have some high bonus damage, which can make for some interesting mindgames in PvP, but this special case is a topic all to its own.
(graphics) Three of a color: monochrome + bichrome
We often see this start, building the bottom and right of a guitar:
The idea here is to use your next monochrome piece to establish the center color of the guitar and build from there. However, in TvP particularly this is vulnerable to early attacks messing up your build if you don't get a monochrome piece right away. My recommendation is to place your first piece flat in 1-2 as usual, then do one of two things. Do the left if your third piece turns out to be monochrome, the right otherwise:
Let's call this shape on the right a corner. The idea is to use the corner like so, allowing us to staircase to the right side of the screen just as with a guitar:
g
b
b
b
gr r
ggrr
Just as with the guitar, we can use this shape to build a head chain tower on the left, and use staircase chains for our tail end to break the right in columns 4-6. Unlike with the guitar build, we're not dependent on a monochrome piece here--we can start building our leftside tower right away.
(graphics) Three of a color: bichrome + monochrome
Well, this knocks out the above decision tree. But we can just again start building a corner like so:
What? The guitar build again? Yes, but if we don't get a monochrome piece shortly, we just play our next g+x piece like so to start our corner:
If you want less to practice, you can just follow this gameplan with both three-of-a-color starts.
(graphics) Bichrome + Monochrome, unmatched
If we're happy mindlessly building guitars over and over, we can just use this to build a more slender guitar as follows:
We extend the guitar handle into column 4, giving us more ability to break sandwiches simultaneously. We sometimes end up extending the handle out even when we're NOT doing this start, since it's often a convenient way to set up our bottom-right chain tail.
(graphics) Monochrome + Bichrome, unmatched
Uh oh, we can't do the above. For those of us who want to keep it mindless, I suggest just dropping the monochrome piece on the right in your preferred position, then going back into this flowchart. You'll probably end up with the most usual start of two bichromes with one matching color, but all cases are handled here.
(graphics) Four different colors.
Again, we can drop our first piece on the right and go back into the flowchart.
(graphics) Monochrome + Monochrome or (graphics) Matching Bichrome
This is pretty easy:
Yeah, it ain't difficult to build a guitar from there:
(gr, bo, gr)
As seen above, we use our first B+X piece to start our leftside tower, and can pretty much fill out our guitar and rightside chains with anything else that comes.
That covers all starts.
Currently seen usage on ladder
The guitar is an extremely common start, especially in the range of top-few-hundred-not-quite-top-tens people, and especially with a starting hand of two bichrome pieces with a matching color. Observing my replays, I notice lots of different ways of handling other starts--the ones I've given above are designed to avoid the need to learn many nuances of many different openings. Once you get to the top-level players, though, you see more variety--most likely because given greater understanding of the game, better plays specific to the pieces received can be found. Most likely also because there's some value to mixing things up, and opening plays which provide for quick attacks, given the right pieces, have great value to those quick enough to use them.
Conclusion
A guitar or simular shape can lead to quickly building a robust, full screen, top-left to bottom-left to right-side combo with almost any start pieces. By refining our map of how to play our first few pieces, we can maximize our robustness and minimize the time to get there, which can be especially useful in PvT. However, we should not mistake this powerful build as the end-all be-all of top level play, despite its incredibly common usage.
-----
tacticalstallingv2
Tactical Stalling
The Tactics of the Stall
The puzzle game I've played with the most epically awesome stalling was Super Puzzle Fighter Turbo 2. In that game, the power of counterattacks could cause stalls that permeated the entire strategic level of the game. In Puyo Puyo Tetris, as Tetris, stalling is more seen at the tactical level.
Most of the time in Puyo Puyo Tetris as Tetris, (link)speed is the name of the game. By putting more pieces on the board faster, we're able to send more damage at our opponent and win. But, in certain specific tactical situations, we're motivated to wait.
<i>Warning: for advanced players only! Until you've mastered playing really fast, you needn't bother learning when to slow down.</i>
Tetris versus Tetris
Background: Attack and Defense
Defensive Stalls
We may be motivated to stall by survival. If one undefended attack by our opponent will kill us, we may need to wait and defend, rather then have our T-spin or Tetris used on an attack which will not kill our opponent and will not save us. For example:
This situation is taken from a game of mine against Wumbo. Wumbo is near the top, and has a T-spin to play. I am lower down, and racing towards a T-spin + Tetris combo. If he T-spins immediately, the combo will kill him.
Wumbo stalls, I don't notice, and I obliviously race towards my combo. Now we're here:
I play my combo, he plays his T-spin, and he lives. He then goes on to win the round.
We normally want to play as fast as possible, and usually don't want to spend a bunch of time thinking about whether to stall. So let's think of some rules of thumb that motivate this stall:
* His next attack will kill me if I don't defend it, and my next move is my onlychance to defend for a while.
* I cannot immediately clear downward by rushing ahead, saving myself with my clears (as implied by above rule).
* Damage from my next attack cannot kill, and is thus best used for defense.
Distilling this still further, we're motivated to stall if we're near the top and we don't have a clearing combo to play after our T-spin. Given that a Tetris will generally clear 4 lines which may be sufficient to save us even without defending, we're much more likely to get this situation with a T-spin double and no multiline clears immediately to come.
Of course, if our opponent notices us stalling, he can counterstall. This would be an example of (*dum dum dum*):
Offensive Stalls
Tetris versus Puyo
Background: the Slippery Slope and the Counterpunch Comeback
<talk about how tetrisgets to come back via counterpunching with garbage, while puyo is rendered increasingly ineffective with garabage.
Pre-Combo Stalls
One time we might wish to stall in order to take damage rather than defend is immediately before doing a high-damage combo. Let's go through a example from a game I played against Coconut, then Michigan's #1 player:
Here, I attack with a T-spin double and Coconut responds by breaking a 3-chain which would otherwise be buried by my attack. While his chain resolves, I build my next attack as fast as I can. Now we're here, with me dropping my S, I in hand, and TISTZ to come:
His chain hasn't quite resolved yet, still needing to pop its last group. So I stall, and once the chain resolves, drop my S. I get 7 garbage lines, not enough to kill me:
Now I attack with my combo, doing a T-spin, a Tetris and clearing 13 lines total with my combo, filling my green bar to 16 and doing 10 lines of damage--essentially, a lethal attack. I add a little icing on the cake by almost immediately clearing the four remaining lines of his garbage for two more counterpunch attacks: (Sequence: TISTI for 4-combo)
This wins me the round. As an aside, we also see one of the great values of mixing in a smaller attack shortly before our big combo: we can put our opponent oni a decision where he must either break his chain immediately to counter our small attack (leaving him without it to defend against our big attack, and possibly giving us material to counterpunch him with) or risk having his chain blocked, again leaving him vulnerable to dying to our big attack. If I had just raced to my big attack without the smaller T-spin first, Coconut could've continued building, used his chain to counter my big combo, and likely ended up in an equal position rather than dying.
Looking at what made this stall so good, we can come up with some rules of thumb for when to pre-combo stall:
* His undefended attack will not kill us.
* We can clear to the bottom, or almost all the way there, on our current screen with our pieces immediately to come. Why? Because this means we can likely use part of his attack to counterpunch immediately.
* We don't have to wait too much longer for the attack to resolve; e.g. it's on the last beat of the chain. Why? We see in the above example that I was able to build up from a just a Tetris to Tetris + T-spin + 4 lines to combo in the time it took his chain to resolve. If his chain has a bunch of time left, it will likely be better for us to just build and break another attack to hit with rather than counterpunch using his current one and give him the opportunity to also build.
Note that all of the above rules don't have to apply for it to be a good time to stall, but if all apply, it certainly is. Note the first rule. As a more general version, realize it's usually not worth stalling for an attack that will do too much more than 7 lines of damage to us--it is likely fatal, and will not immediately fully resolve as we only take 7 lines of damage per drop.
Background: Combos and defense mechanics
(Combo mechanics and counter-damage)
Mid-combo stalls
Executing the Stall
Practical Usage
<talk about how we'll win most rounds without this stuff; erring on the side of speed, etc>
Conclusion
While we will win many a round doing nothing but playing as fast as possible, in some specific situations, a well-timed stall can mean life for us or death for our opponent. For those who can play very fast, recognizing the moment to slow down can turn the tide in games against strong players.
-----
tvpleftolongcombo
TvP: The Left-O -> Long Combo Opening
Let's get one thing straight: this is not my favorite TvP opening; not even close. My favorite is the WumbOJ20->Izzy15 loop, falling back into repeated column 4 or 7 T-spins. That opening continually hits our opponents with attacks, big and small, and can win without taking into account our opponent's play. Left-O->Long Combo, on the other hand, features more flexible but less necessarily frequent attacks, and requires adaptation to our opponent to be viable in high level play. As such, it's worth studying as an example of the interplay between the two players in Tetris versus Puyo.
1. The first two decks.
Above, we have various executions of the Left-O opening, starting with the strongest, which gives a possibility of all-clear at 25. Note that all of these variants will leave us with three squares of material in columns 1-4 after we execuute our two T-spins, giving the possibility of following up with a Long Combo.
2. First decision: to attack or not to attack.
Our first decision point comes as we get our third-deck T. Presuming we're still holding our second-deck T in hand, at this point we can let loose with one or two T-spins. Why wouldn't we do this?
<left-o versus robust gtr>
We see above our opponent has a very nice drop tower leading into his transition and left-to-right. If we do two T-spins, we'll hit him with 3 and a bit lines of damage. We can see this won't even cover his trigger: he can still hit us with an X combo AFTER taking the attack. Worse, he can continue to build, since he can just build a staircase tail combo above his left-to-right--a staircase will trigger at any height as long as his pieces below get broken. So our attack will do all of nothing, and we'll probably die once he's ready to break his chain.
What is our alternative? Obviously, to keep building. We place 1-2 T Tetrominos on our column 5-10 tower, increasing the length of a potential future long combo. We don't want to go much more than two Ts, though, or we'll end up building almost to the top of our screen before we can use our initial two T-spins--bad news.
Once we've gotten to our next after we've done our further building, perhaps with a T in hand, we now probably want to break our attack. By now, they've likely dropped enough pieces that we CAN cover their trigger, and they'll be forced to break in response or risk being buried. Let's look at how this works.
<two 2 spins>
<opponent board>
Opponent options: trigger or not trigger.
Here, there are two main possibilities. Our opponent, seeing our attack incoming, may break his main chain. Or, he can ignore our attack and plan to dig out his chain and hit us back. Either way, we're continuing to build up our tower until we get to our next T. By this time, our opponent has likely made one of the two choices.
A) HE TRIGGERED IT!
Okay, we've built up a big old possible chain. He's triggered his attack. We now need to survive his attack and hit him back, so we're going to start breaking our combo. Ideally, we'd build exactly high enough to survive his attack during our combo execution, take a bunch of not-quite-lethal damage, and continue our combo using the damage he sent us as we reach the bottom of our tower build. Really, we don't need to be THAT greedy--if we manage to survive his attack during our combo, and get a couple extra lines from breaking his garbage, we're probably going to win via hitting him back at this point.
B) HE DIDN'T TRIGGER IT!
Okay, we've now put 3 lines towards burying him. Of course, we could keep building up our combo, then execute it. Or we could execute it right away. Both of these alternatives suck. Why?
We see in this example where my opponent has taken the hit. While his combo trigger is buried (the highlighted blue group on the right), a few good breaks will unbury his trigger. Whether we build up or start executing our combo immediately, we won't hit him until the combo is over. This gives him plenty of time to dig down to his combo trigger, at which point we get hit by at least a 7-chain--likely sufficient to lose us the game even with a pretty big combo.
On the other hand, if we take the initiative and attack right away, just a few more hits of garbage should be sufficient to kill him. He's a few rows from the top, after all. So let's not waste our chance.
Instead, what we do is take a mini T-spin single with our T.
This does 4 blocks of damage, almost a line. We're continuing to harass him and keep his combo buried, not giving him a chance to dig out. We follow our T-spin with yet more T-spins and Tetrises. For example:
Tetris: almost 2 lines of damage.
T-spin single + Tetris: 3 lines of damage.
If he hasn't managed to drill down to his main chain and trigger it by this point, we've won. If he HAS, we've likely harassed him enough that it's not too big a chain, and we have a good shot of winning the incoming back-and-forth midgame.
3) Attacks in the hole.
The above example uses T-spin singles and Tetrises in the hole. This is kind of a happy medium, and generally what we want to do. The other extremes:
Minis galore.
Note the two-line clear here. Two and three line clears are quite viable in this situation, sending 4 or 5 damage blocks respectively and disrupting their attempts to dig out of the top of their screen. We should not insist on only T-spins and Tetrises. But note that after something other than a T spin or Tetris, the 'mini' T-spin singles shown above don't do damage--we rely on the back-to-back bonus to make these damaging.
Greedy trips
You might recognize this from Glitch's Greed. While it's not generally practical as a go-to keep-up-the-attack, it can come in handy if you need to burn a S and a L while building your combo:
We can then later fill in the LIZ on the left, and do the two trips. And continuinue with trips or a double into a Tetris or a two-line clear:
What if we don't get the pieces to be this insanely greedy? Well, we can go back to our happy medium of T-spin singles and Tetrises:
T-spin single
L + Tetris
T-spin single
And hey, we're back here and ready to do whatever, including the above T-spin single+Tetris pattern.
4. Generalized Understanding
Conclusion.
