Defensive Play

This might be the last major article I ever write for Tetris. It details my style and the thoughts I have come upon over my years as a competitive Tetris player.

I never wanted to be a defensive player. When I started out and when I gained my first Cases championship, I played exceedingly aggressively. This, generally, is how new people play. I was a fast pure player - perhaps the fastest of all, which might now be hard to believe.

Still, I was not unbeatable. I do not believe in anyone being unbeatable, but I lost fairly regularly - about 10% of the games, after I had gained the championship. I wasn't pleased. Sure, I could regain the championship if I was ever to lose it. But then, there came other players, and I got tired. It's hard to keep up the speed over the years, it's stressful. The new ones were faster than I, so I had to learn to deal with people possessing physical advantages over me.

This lead to what would be the transformation of my game. Before, I would instinctively play for an aggressive downstacking - using as many enemy lines as possible to throw back. Now, downstacking became my life and blood, and the method of my victories became survival. I would sacrifice speed to great amounts in order to find the perfect place for the block, in order to accurately judge the field weight of the different permutations and in order to assess the potential danger of every move contrasted to the situation of the opposing player, going sometimes as low as 20 blocks per minute.

There is one simple way to distinguish between average, good and great players and that has all to do with the piece preview function.

An average player is one who has learned to place one block. He will place this block so that the result is favourable to subsequent ones.

A good player is one who can place two blocks, using the preview function. He will place the first block so that the second one will have a decent anchor, if at all possible.

A great player is the person who can calculate each decent placement combination of two blocks, the current and the preview, as to how benevolent the end result would be to further blocks.

Here, we are only talking about what I like to call the "surface stacking". These distinctions also apply to the second great rule and theory of Tetris, the "depth stacking". This involves placing each block in a way that takes into account the entire state of the field. Then, we have this:

A good player is one who has learned to place one block so that the resulting permutation will not carry any needless unpleasant side effects on later term; simplifying it, he will place the block on stable ground. There are a number of those players.

A great player is one who can routinely judge and place two blocks, the current and the preview, in the same manner. There might be two or three such players in the world today.

The truly supreme player is one who can calculate each decent placement combination of two blocks as to how benevolent the end result would be to further blocks, on the merit of the field weight and depth stacking issues. There are no players in the world today able to play at this level constantly.

Making judgments regarding the priority of deep versus shallow stacking concerns is very important for a pure player's game. There is insufficient room in this text to go into great detail, so I'll just show you a very simple example.

Current: Box
Preview: Stick


 

The following notation is given in the x:y format.

Here, surface stacking clearly dictates that the current piece (Box) should be placed at [4-5]:12. It is the only fit. However, as most half-decent pure players will realize, it is also a mistake. Box at [4-5]:12 will hide well at 4:[04-07].

The worst placement in this situation, clearly, is [11-12]:12, hiding x:12 for the preview stick. The next worst is [3-4]:13, slightly worse than the logical surface stacking measure, [4-5]:12. Fourth worst is [1-2]:14 or [2-3]:14, obscuring small hole 2:06, which has to be cleared to proceed to well 7:[01-03]. It therefore follows that the fifth worst is any placing obscuring this well, namely [6-7]:14 or [7-8]:14.

The three best placements are [8-9]:14, [9-10]:14 or [10-11]:13, with perhaps the second being preferable. [9-10]:14 will make it possible to fill resulting small hole 10:13 with a variety of blocks. Since the necessary slide will be made on solid ground, it is an acceptable risk even for an instant-add client.

Clearly, the example above is exceedingly simple. Yet, consider how many alternatives must be taken into mind even in making such a simple decision.

As you noted and for reasons that should be clear to anyone, I recommended to make a priority to avoid covering the upper holes in the field. This is, of course, because you will have to clear the upper holes in order to proceed to the lower, and because the upper holes are the more urgent. However, this does not mean that the upper holes are not sacrifiable. If you remember, I stated in my pure strategy guide; "Try to build even, but not too even". This was a simplified rule, which should be expanded upon.

Try to stack evenly on all unsteady ground in rows you are interested in converting over the near future. For example; if you have one clear 2y deep well at x:12, one 1y hole underneath at x:2 and one 5y hole underneath that at x:8 (see diagram below), it is hardly good politics to stack 6y of debris on x:8 just to leave x:12 and x:2 free. In that case you will have to clear 6y in the near future just to get down. Stacking 3y over both x:2 and x:8 would probably be much better. Even covering the clear well should be considered.
 

Current: Z
Preview: Box

Personally, I try to be concerned mostly with the top 6y of my upfield, stacking the unsafe ground in this region fairly evenly. This ensures that I have an escape route at all times and accounts for the difficulty most people have in killing me. But; that is a matter of taste and personal style. Sometimes, when I feel that the game will likely draw out or I see an excellent opportunity for long-term strategic advantage, I will consider as much as 15y if that is available. Other times, for example if I need a fast 2+ add within 30 seconds for a certain kill, I won't concern myself with it at all.

There is clearly a good argument to be made for a gradient in the stacking based on the depth and position of the holes in the upfield. Merely being aware of the fact that such a gradient should exist should be enough to lead your game on the right path. At least, try to remember never to stack more debris on an upper hole than a lower one.

I will say with the authority of being the single most experienced regular pure player in the world, that these are by far the most valuable skills to possess. This is what I have concluded from playing and watching tens of thousands of games. When you play fast Tetris, you are always likely to some time run into one player who applies an overwhelming amount of pressure, one who is immensely faster than yourself, one far superior adder. That is when the strategic approach to playing becomes hard to carry through, but even then, the skill of predictive judgment remains the most important one you could ever hope to acquire.

In regular TetriNET and because of the piece delay, there is no way to apply such pressure. The game is one of paradigms clashing with each other, not one of brute force and speed. In this sense, regular TetriNET is the most qualitative of all online Tetris games, akin to how chess by correspondence represents a higher quality of game than what does live chess. Consider also how it is fairly easy to invent a Tetris bot able to beat any human player by moving at a couple of thousands of blocks per minute, but thus far, no-one has been able to invent one capable of defeating even an intermediate player by using the same speed as its opponent.

In my opinion, purely defensive play is clearly superior to purely offensive in TetriNET. That might seem a bold statement, so this is the rationale.

Sending four lines by way of constructed field takes at least 12 blocks in a 12x well. Clearing four lines with one hole each in them rarely takes that many if you have developed a healthy sense for it - if you're lucky, you can do it with one! Thus a defensive player can outspeed the offensive one's adds even if he is slower in bpm, providing the aggressor does not care for his upfield. In practice this has proven to be true. Add to that that the defender will send lines while clearing them, and you have a very uneven scenario indeed.

Nearly all offensive players want you to bang with them (well, unless they are so poor they can't stand being hit). That's their forte, that's what they like. They have no answer to a defensive strategy, except to add, add and add, and all defensive players expect adding, while no offensive players expect the opponent to simply stay alive. An offensive player rarely has the same endurance as a defender, and aggressive players always get frustrated by defensive ones, while the reverse is rarely true. Thus, they lose.

But it should be noted, a defensive player has to add too. Actually, it is most necessary in order to stay alive, and this is usually the defender's failing. The offensive strength of a defender must be boosted by keeping a high yield in the survival play, and occasionally even upstacking, since a good offensive player will force a defender to attack by going for large initial stacks.

Following the outlines for the downstacking laid down above will facilitate increasing the yield. Also, offensive players can gain great power by being aware of defensive techniques and incorporating them in their game, thus increasing their offensive strength while at the same time improving their durability.

In the end, the outcome of the game will naturally simply depend on who is better - with "better" being used in the broadest sense of the word.

spindizzy/HELLFIRE, 06/04

Main.