←back to thread

163 points juancroldan | 1 comments | | HN request time: 0s | source
Show context
jchw ◴[] No.43632004[source]
I have a somewhat contrarian opinion. I think if you're making a clone of Tetris you should actually take design cues from the somewhat more obscure Tetris the Grand Master series instead of the "guideline" or NES Tetris rules. TGM's rotation and kick rules are a lot more elegant and avoid a lot of unneeded complexity. Guideline Tetris kicks let you do absurd and weird things (look up the series of kicks that make up a T-Spin Triple and see if that makes sense to you) and rewards doing canned setups really fast, whereas TGM's game design is all about doing good stacking very fast.

The TGM randomization algorithm is also pretty elegant. 7 bag is a bit extreme, it gives you such a perfect set of pieces at all times that it's genuinely less challenging and fun. TGM's random piece algorithm is a lot simpler: the randomizer has a 4-piece history window and it tries multiple times (IIRC, 6) to find a unique piece that hasn't appeared in that window. It is initialized to SSZZ to lower the odds of starting with an S or Z early on. (~~They also use the Mersenne Twister as their PRNG, which was a pretty good PRNG in an era where many games still used LCGs.~~ edit: Apparently, they do not. Don't ask me where I got this, I have no idea.)

Now of course I'm not sure if it matters at all for this particular game since it isn't really a Tetris clone at all, but while TGM is a well-known cult classic for people deep into Tetris it's relatively obscure outside of that circle (and presumably outside of Japan.) The Tetris Company is very strange about licensing and has apparently, as the legend has it, blocked and forced changes on TGM releases for a very long time due to the fact that it doesn't fit with the Tetris guideline rules they enforce in an oddly totalitarian fashion, probably suppressing the game even further in an era where speed games and competitive gaming is a lot more popular.

replies(8): >>43632098 #>>43632218 #>>43632479 #>>43632584 #>>43632609 #>>43638944 #>>43639185 #>>43640666 #
qsort ◴[] No.43632218[source]
I think it depends on what you want to make a clone for.

TGM rules are more elegant and much more challenging for single-player tetris, but on the other hand multiplayer and especially 1v1 has far greater variety and dynamism with a looser rotation system. The stricter the rules, the more it becomes spamming tetrises against each other until someone misdrops. You have to construct a ruleset where a player who is 0.1pps faster doesn't just autowin.

replies(2): >>43632335 #>>43637651 #
Dwedit ◴[] No.43637651[source]
I watched enough Tetris 99 streamers to know the basic strategy, you 4-wide when on defense, and T-spin/Tetris when on offense.
replies(1): >>43640760 #
1. mitsu_at ◴[] No.43640760[source]
Tetris 99 is a neat example of how the design can result in unintuitive optimal strategies at high level play.

Two mechanics that come to mind are (1) targeting bonus and (2) the finite garbage cap.

1. Targeting bonus makes it so that if multiple players are attacking you, your attack is buffed. Intuitively it makes sense, since it evens the scales a bit: If, by happenstance, you are unluckily targeted by multiple people, at least you have a fighting chance at surviving the 1v2 or 1v3 etc.

One counterintuitive consequence of this, however, is that an optimal strategy to farm KOs is to stack very high on the board so that you are in danger of topping out. A reasonable number of other players will have their targeting strategy set to KOs (since it is 1 of 4 options), and will target you. If a massive number of players are targeting you, a reasonably skilled player is in no danger because they can easily offset incoming damage with singles or doubles, buffed by targeting bonus.

Example of a game using this strategy to achieve 51 KOs: https://www.youtube.com/watch?v=cat9TbAwRWU

Another consequence of this is in the teams mode, if the teams are imbalanced, counterintuitively, a top level player would prefer being on the smaller team over the larger team. To be fair, there is a certain skill floor required to survive the initial onslaught. But if you can survive and get KOs, the attack % buff you get from the KO badges will eventually give you an advantage over the larger team, eventually to the point where victory is almost guaranteed.

Extreme example of a player winning a 1v98: https://www.youtube.com/watch?v=OTE4zEKtxkA Most normal players would not survive the initial influx of garbage. In particular, because the garbage meter is constantly full at 12, you are not even outputting any damage at all unless an individual line clear sends 13 or more (in which case, 12 of that cancels the incoming garbage, and a mere 1 attack is sent out).

But once you start getting KOs, you start getting badges, and eventually by the time the opponent team dwindles enough that you no longer have targeting bonus, you will have a +100% attack bonus from full badges, and you should be able to easily destroy the remaining players.

2. There is a garbage cap of 12. Which means if your garbage meter is full and more attack comes in, it just vanishes.

Intuitively this makes sense for balance reasons, to make the game feel less RNG/frustrating. If you are unlucky and receive damage spikes from multiple players simultaneously, your incoming garbage is never more than 12, so at least you have a reasonable shot at surviving.

A consequence of this however, is if you are on the backfoot and your garbage meter is filling up, it can be better to play slow than to play fast.

For example, suppose you have 8 incoming garbage. Instead of playing quickly, it can be a better strategy is to not place the piece, and wait for your opponent to send. Suppose your opponent sends 10 attack. Here, only 4 of that will go into your meter, and the rest will vanish. Then you can accept the garbage, and you will be in a better position than if you had played fast.

Even if you are at a badge disadvantage, it's possible to catch your opponent off guard with an unexpected counterspike, expecially if they have just sent everything and have no resources to defend.

Example of a Tetris 99 game where the final 2 lasted for 23 minutes: https://www.youtube.com/watch?v=DoLh5w5rwS0 At 19:55 you can see Amemiya stalling until he sees that Doremy has no resources on his board, and then counterattacks fast.

A few years back, HardDrop hosted Tetris 99 snipe lobbies, and those lobbies would stretch excruciatingly long because even with badge disadvantage, reasonably skilled players are very, very good at stalling.