Sunday 17 March 2019

Tetris versus Tetris: the strategic overview

    

Honing our speed and tactics pays by far the most dividends, in terms of our win rate, and this blog largely has focused on these. But, just as we took a step back and talked about TvP (Tetris versus Puyo) mechanics and strategy, let's talk about TvT (Tetris versus Tetris).


1. Basic mechanics: attack and defense

In Tetris, we do damage only when breaking lines. Depending on what we break, and how we break it, we send a particular amount of garbage to the opponent's screen. We win by topping-out our opponent: either he places a piece entirely above the 20th line, which instantly kills him, or a piece intersects material on his board (largely via material being above the 20th line) which also instantly kills him.

When we send garbage, we send a particular number of lines at an opponent, represented by symbols on his board. A little square represents 1 line, and a big square represents 6. Beyond these, a red symbol can represent 30--though this is rarely seen in TvT. Our opponent takes damage once he drops his next piece, but if he himself sends an attack, the attacks cancel out--he defends. If he uses a break that would attack for 4 damage, and is about to take 10... well, he ends up taking 10 - 4 = 6 lines of garbage. If, on the other hand, you send him 4 and he responds by doing an all-clear on that move for 10 damage, he sends you 10 - 4 = 6 lines, and takes no damage himself. Situations around attack and defense can produce intentional tactical stalling to either intentionally defend or intentionally prevent defense by waiting to break one's own attack until after an attack has been received. That said, it's optimal to play as fast as possible the vast majority of the time, and players outside the top level of play need not worry much about intentional stalls.

2. Basic mechanics: the attacks

While all attacks come from breaking lines, different breaks do different amounts of damage. We can divide the different attacks into 5 basic categories:

1) The Tetris


A Tetris is only achieved by breaking 4 lines at once using an I. Doing this does 4 lines of base damage, but can do an additional 1 line of damage via the Back-To-Back bonus, which occurs if our previous break was a Tetris or T-spin. While Tetrises are the most well-known attack in the game of Tetris, the T-spin becomes more important in high level play.

2) T-spins
        

A T-spin does 2, 4, or 6 damage, depending on whether it breaks 1, 2, or 3 lines. In other words, it does twice as much damage as lines broken. It may also do an additional 1 damage via the Back-To-Back bonus, again occurring only if our previous break was a Tetris or T-spin. Puyo Puyo Tetris also has the concept of the "mini" T-spin, which is good only for the back-to-back bonus, but not the double-our-lines-broken attack damage. Mini T-spins are much less useful in TvT than in TvP.

Opening with a T-spin or two T-spin combo is quite popular in TvT, much more popular than opening with a Tetris. Moving into midgame, one sees repeated T-spin patterns at the top level of play, where Tetrises are mixed in as the board builds up, but one also sees more balanced and more Tetris-heavy patterns.

3) The all-clear
  

If one completely clears the board as Tetris, the message "PERFECT CLEAR" shows on the screen, and exactly 10 damage is sent to one's opponent. This damage is the same regardless of how we cleared our board--whether we did a one line clear, or a Tetris with a back-to-back bonus, we still send our opponent exactly 10 damage. Combos, which add bonus damage to a Tetris or T-spin, do not add bonus damage to an all-clear.

All-clears occur primarily in the beginning of a match, on the tenth, fifteen, twentieth, or twenty-fifth piece. Largely, we see practiced and memorized all-clears used as openings. The standard 10-piece all-clear, the Ace, dwarfs the others in popularity, though I personally do not use it much.

While mid-game all-clears do occur, taking an odd number of damage rows entirely prevents an all-clear, and taking any damage at all makes one much less likely, zoning out any pre-memorized all-clear sequence, except for the occasional case of taking exactly 4 lines of damage and clearing with a Tetris. On the other hand, we can often follow up one all-clear with another, since a completely clear board is the perfect starting point for either a pre-memorized or improvised all-clear.

4) The combo

As I alluded to above, combos add bonus damage to existing attacks. A combo is a series of line breaks without an interruption, clearing at least one line with each tetromino.  In Tetris vs. Tetris, every hit of the combo from the 2 combo onward adds bonus damage, increasing every two drops.

Opening with a Long Combo is often seen, though substantially rarer than T-spin openings in top level play. We also see intentional combo building after deck 3 or so, where players will go for a side long combo after defusing their opponent's initial attacks by trading their own defensively. Besides intentional Long Combo setups, combos often occur when handling garbage--a player will make a series of breaks in a row to get rid of garbage and do a huge amount of damage to their opponent in the process. Sometimes, these two combine, where a player builds a very flat (but not necessarily very tall) Long Combo setup, clears the built-up lines, then continues with a garbage clear for massive additional damage.

The possibility of a Long Combo being built is one of the defining tensions of TvT. It creates a back-and-forth where large attacks are held in reserve to bop potential combos, and one constantly is required to create potential pressure.

5) Two and three line breaks.

Finally, one can do one line of damage with a 2-line break, or two lines of damage with a 3-line break. These are far less important in TvT than in TvP. In TvT, these largely exist as icing-on-the-cake when doing a significant clear, sometimes putting our opponent over the top with the just-one-more-line we get from these. However, they do have tactical significance in a few rarely used openings such as the T-block Ace, where the 2-line clear sends a row of damage preventing all-clears by our opponent.


3. The Opening

Roughly speaking, we generally see four classes of openings used in TvT, with a fifth that exists primarily as a fallback.

1) Quick T-spins

Here, we do a T-spin with our first deck (our first 7 pieces), then plan to keep doing T-spins each deck unless we switch to doing an all-clear or counter-punching our opponent with received damage. The most commonly used opening for this is the Izzy:


This opening offers substantial all-clear possibilities, especially on piece 15 or 20, but is usually seen instead branching into a repeated T-spin pattern in column 3 or 8.

2) 10 piece all-clears

Most commonly, the Ace:
  

Here, the idea is to hit an all-clear as fast as possible, then follow up either with another all-clear, or a switch into a quick T-spin game. The 10 or 11 lines we do can disrupt our opponent's planned combo building, as well as putting us at an advantage in terms of efficiency.

No 10 piece all-clear is reliable--all require luck as to one's second deck pieces, and all require not taking damage before we clear. If we miss the all-clear, we fall back to a different sort of opening. In the Ace, we most commonly see a fall-back to our fifth class of opening, the TetrisTea opening, which in TvT primarily occurs as a fallback rather than intended path. Other possibilities include falling back to a Long Combo, a not-that-quick T-spin, and building a multi-spin. We also see 10-piece all-clears used opportunistically rather than intentionally, for example branching off the Left-O opening, if one holds the T and gets J, L, O, I as the beginning of the second deck. This happens a couple percent of the time, but only a small proportion of players train to spot such things and exploit them.

3) The Long Combo
        

As an initial opening, by far most commonly we see the central Long Combo. However, side Long Combos are also worthwhile, particularly for countering central Long Combos. The existence of the Long Combo opening puts substantial pressure on all other openings--it has the potential to hard counter anything too slow to handle it, such as some multispins. Besides being opened with directly, we also see Long Combos frequently after certain multispin openings, particularly the Left-O.

4) The Multispin

Here, we don't T-spin in our first deck. Instead, we set up a combo for intended execution typically during our third deck, with at least two T-spins. The most commonly seen multispin opening is the Left-O:


Such openings delay our initial attack, and sometimes win against quick T-spin openings by simply replying back to their early damage by taking their two T-spins and counterpunching the quick T-spinner to death using the damage blocks received so far. They also have some utility against side Long Combos, as the side combo cannot safely build higher than the multispin's expected damage and thus must start breaking, which can then be exploited by the multispin side of things, e.g. by building a bigger combo.

5) Tetris Tea
  

This is often used as a go-to opening in TvP. In TvT, however, we largely see this as the primary fallback on a missed Ace or after messing up a pre-practiced opening. Essentially, we build Tetrises as fast as possible while opportunistically combining them with T-spins.


4. Midgame perspectives.

After the first two to four decks of the game, we generally move beyond pre-practiced pre-memorized piece placements, though some openings lead to themes that persist many more decks in, such as repeated T-spins, Long Combo building, the specific multispin followups of the Music, or repeated all-clear sequences. Usually, both players continue to try to attack each other and clear their own boards as fast as possible. From there, it's often a war of attrition.

1) Attrition

In a basic attrition match, players attack each other with T-spins and Tetrises as fast as they can, while at the same time clearing their board to avoid themselves dying. Heavily offensive players sometimes ignore the clearing part of this, focusing entirely on sending repeated attacks at high speed, but tend to gradually get topped out. Heavily defensive players may focus excessively on keeping their board as low as possible, to the detriment of killing their opponent. But of course, purely defensive play that focuses entirely on keeping a low board loses quite trivially to a Long Combo transition, against which it has no defense.

2) The Long Combo Transition

If our board is very low, and our opponent's ammunition is limited, we have the time and space we need to build a Long Combo. This is most often seen on either the left or right side of the screen:

  

If our opponent is focused on a defensive, board clearing approach, we have plenty of time to build 11+ lines. Breaking 11 lines in a combo will do a 10-combo for 25 damage--and of course, by building higher, we can do even more.

While the side combo is by far the most common as a midgame transition, I also sometimes opportunistically transition to a central, an off-center in columns 3-6, or a three-wide hole in columns 4-6. These all have fairly good survivability if we take substantial damage, but transition less easily and tend to build a little slower.

3) The Bop

We can counter a side long combo with a bop: a quick set of attacks which, combined with the opponent's high build, tops him out. One of the strongest bops is the Trips:

  

This gives us two back-to-back T-spin triples, for 13 or 14 damage. If they've built their combo up to 7 or 8 lines by the time we're ready to break, they just die instantly to our attack. Thus, they have to stop building up, and perhaps start breaking their combo.

If they start breaking, we can actually save up our attack and continue to build up with our own counter side long combo:

    

The T-spins can be broken later to counter the later hits in one's opponent's combo, preventing us from dying to it and allowing us to build a quite high potentially lethal combo ourselves. Of course, rather than breaking his combo to avoid death, our opponent could theoretically instead build his own bop in the hole, thus forcing us to break our combo:

    

  

    

Other than the Trips, we frequently see the weaker but much easier 9-10 damage bop of a T-spin double + Tetris:

  

And, of course, other midgame multispins, double Tetrises, and combinations thereof.

        

While the slightly lower damage offered by the above allows a combo to build higher than a Trips would let it, the idea of forcing a combo break with our potential attack and potentially using the window to build our own combo still applies.


4) The Counterpunch

While the long combo transition and the counter of building a bop hold great sway when opening attacks cancel out, in situations where damage has been exchanged, we can do more damage by using the damage blocks we've received.



Here I'm at the end of the second deck of the WumbOJ, about to take my second T-spin for 7 damage. I have taken 9 damage, all oriented in column 9.

          

Because the damage is oriented in column 9, I do the T-spin triple third-deck followup. From there, I can use the damage blocks to counterpunch with two Tetrises and a T-spin double. Starting with the first diagram, I've done 7+7+5+5+5 = 29 damage within 10 or so pieces, well enough to kill.

We can combine the ideas of the midgame combo and the counterpunch with the counterpunch combo. Here, I have a bunch of damage entirely oriented in columns 5-10:



I build up columns 1-4 with my next few pieces, then I'm ready to start breaking:

           Sequence: TOSSILJZJZSI for 12-combo.

Here I have a continuation up to 12-combo, but my opponent died at 9-combo. We see that the material stacked up on the left, not oriented over any damage holes, helps us continue to combo by absorbing the extra material of tetrominoes dropped to break damage lines. If we have 6+ squares of material on a given line, we essentially get a combo hit for free, and lines with fewer than 6 can absorb material from lines with more. However any sticky-uppy spare material over a damage hole will end our combo once we get to the corresponding line of garbage. In the above combo, there was no damage oriented under the sticky-uppy columns 1 and 2, and I was able to continue my combo almost to the bottom of the screen.


5. Endgame

In chess, endgame is a final phase of the game. Often, one player or the other is essentially assured of winning given even moderately good play. In TvT, the finality may be in doubt, since it's generally possible to get back to midgame via both players clearing their screens and cancelling out each other's attacks, but the chess-like property of moderately good play ensuring a win is still there. Another stand-out property of TvT endgames is that it very often becomes unnecessary or even undesirable to play as fast as possible--stalling or pausing to think is the correct play in a variety of situations. Let's look at some basic endgame situations.

1) Attrition: T-spin + Tetris versus Tetris, high up.

Often, a lead in an attrition match will take us to a place where we have sufficient damage to finish off our opponent, assuming they don't defend. Aye, there's the rub: sometimes we have to play around potential defenses.

      To play: L In hand: T To come: IJZLJI

Above is Wumbo on the left versus me on the right. I am racing towards a lethal Tetris + T-spin combo. Wumbo intentionally stalls his Tetris to prevent my coming 10 damage from being fatal. If I immediately drop my L and take the Tetris and T-spin, Wumbo will only take 5 damage due to the cancel-out, and with his 4 lines cleared from the Tetris, will have room to come back via some fast board-clearing. On the other hand, if I stall dropping the L until after his I drops, I'll take 5 lines of damage then immediately send him 10, killing him. Of course, once he notices my stall, he's incentivized to drop the I immediately and try to clear his board as fast as possible, before I hit him with the 10 damage, but it is unlikely to be sufficient to save him without the 5 damage cancellation.



Suppose I don't notice his stall until after I drop the L? In that case, my alternative is to drop the I on the right then race to my upcoming T-spin + Tetris + T-spin before he clears his board sufficiently. The moment I make this drop, bypassing my combo till my next I, Wumbo again is incentivized to drop his I instantly and clear his board as fast as possible, before my attack comes.

At this point it's 4 pieces till my next I for my next Tetris. In this time he could likely clear 7 lines and uncover his next Tetris. So while this lesser stall might be good, it gives him a lot more chances than just stalling the immediate Tetris + T-spin would have.


2) Long Combo versus insufficient attack: just closing it out.
    

    

Rei (レイ) on the left versus me on the right. I disrupt Rei's opening Ace with a quick T-spin, then take the opportunity to build a side long combo. 7 breaks later, I'm at 6 Combo and will win if I simply continue the combo.

      To play: O In hand: T To come: LITZJ (6 combo)


I have no need to stall either to cancel out Rei's attacks or bait additional damage to counterpunch with. However, taking some extra time to think and not hurrying is the right move here: if I can extend the combo a few more hits, I win no matter what Rei does, whereas if I play too fast and accidentally end my combo early, the game goes on and may stabilize to equality.

It turns out there is a continuation to 13 combo findable with some brief seconds of thought, though my opponent dies at 11 combo:

    Sequence: TLOTIZJ for 7 more hits.


3) Long Combo versus Long Combo: twists and turns.

    

In Long Combo versus Long Combo, things get a little weird. Above, we have a side Long Combo versus a central. While the side long combo is higher because it can be placed slightly faster, it's more dangerous. If the central player starts breaking first and plays as fast as possible, he may be able top out the other player before he can clear his screen. In fact, this is true even in central versus central, if the players have built to height 15+. Here, if the central player gets three breaks ahead, his opponent's screen will be going upward even if they break at the same speed... and the central player can break slightly faster, as pieces start in the center.

However, suppose the central breaks first, but the side responds and starts breaking in time, and is only one or two breaks behind. Then after a lot of breaks, we might be something like here:

          9-combo versus 10-combo.

Side still has material to burn, but central is going to run out and end his combo, and end up two combo hits down, taking three undefended hits back (due to being a break ahead), and dying. What might he do here? Well, if he stops hard-dropping and his opponents breaks two more combo hits, he'll receive damage before his combo ends, and that damage can be used to hit back:

          Sequence: SOLJZT for 16-combo.

Of course, if side noticed in time, he could have also slowed down, preventing sending two attacks at once and allowing central extra damage lines to extend his combo. As is, though, central ends up getting the extra breaks needed to survive. If side slowed down, but not in time, he might as well then go back to going as fast as possible and hope to get to clearing his garbage for counterpunch damage before central can get on the ball, going from the endgame to entering the midgame at advantage.

As we can see, zany things can happen here, where both players end up playing as fast as humanly possible to try to top the other out before combo end, or avoid being topped out, then suddenly flip to playing as slowly as possible, even using hole-bottom rotations and last second piece swaps to delay things further, in order to either get more material or deny their opponent more material, then suddenly go back to playing as fast as possible. When you take into account reaction time, and players slowing down and speeding up to fake out their opponents, things get even more gnarly.


Conclusion:

While improving our speed and tactical execution often pays the biggest dividends, understanding the broad overview of the game can be helpful in decision-making, and can give us ways to win beyond playing faster and executing better than our opponents--it can allow us to play on a higher level. Understanding the overview of openings can allow us to gain an advantage or even immediate win in the great rock-paper-scissors. Understanding multiple perspectives on midgame can allow us to break out of basic attrition and find win conditions we may otherwise miss. And understanding endgames can allow us to close things out, not throwing away wins that should be ours for the taking, and conversely to snatch victory from the jaws of defeat via exploiting our opponent's lesser understanding.

No comments:

Post a Comment