Puzzle Bobble

- 5 mins read
Two Bubs --- the protagonist green dragon --- load and aim bubbles into their cannon.

It’s hard to imagine that Puzzle Bobble arose from nowhere in particular, but none of the post-Tetris random-piece puzzle games that came prior resemble it: it’s match-three, but the board is designed by-hand and drops from the ceiling, crumbling only as the player shoots bubbles into it. One might surmise that, in an alternate universe, a continuous lineage from Breakout and Arkanoid to Puzzle Bobble exists, with some intermediary game that blends the real-time paddle of the former with the match-three of the latter, yet none exists. The key genetic trait retained here are the ricochets, a dollop of the basic kinetics from the block-breakers injected into the more modern puzzle concepts. Puzzle Bobble combines this with a hexagonal grid to create craggy playfields where one can sneak bubbles into crevices.

In the most traditional case, this can play out on a solid playfield of alternating colors, letting the player pick away at the front. This alone reveals the base system’s most intriguing dynamics. For instance, the staggered grid alternates seven and eight-wide rows of bubbles, and firing into the eight-wide row will inevitably leave one of its bubbles obstructed. Avoiding this requires some guesswork on what bubbles will show up later in the queue, which is no small feat given that the only one upcoming bubble is previewed.1 This results in juggling between the exposed instances of a given color across the board, both at the front and open via holes eaten away from previous clears. Choosing when to cover one to set up a pair for another, especially when considering the difficulty of making a given shot, generates enough decision-making to stress one in the 10-second window for each.

The game’s two-player mode is also especially loved, combining the game’s mechanics with a garbage-send concept similar to Puyo Puyo and switching out combos for numbers of bubbles dropped after completing a match. Later games would expand on this and add a single-player gauntlet against computer-controlled antagonists. [src]

Another particular archetype that stresses these dynamics are the string/stalactite levels, where a single-wide column of bubbles zig-zags down from the top. The early-game versions of these pin each column to the ceiling via two identical bubbles apiece; given that gravity is in effect, destroying the base automatically drops every bubble below it that it supported. This aspect crops up in many levels, but in these it’s especially prominent. As more columns are added (three in level 15, and four with no bases in the infamous level 30), it becomes harder to avoid tying the columns together with stray shots, especially in the latter stage, where many colors are inaccessible at the front. Tying the strings together effectively double-secures them against the ceiling, which explicitly prevents removing them, but it’s impossible to avoid doing so given that the hexagonal grid makes the space between strings the only area one can place chaff. To succeed, the playfield needs to be carefully partitioned — spare colors in a drop-pile under a pair, and the rest allocated to specific available matches — without attempting to seize on every opportunity, lest one accidentally overcrowds lanes to important matches.

With the prior pattern, the designers understood that escalating the difficulty required removing easy-clear bases at the top, and yet the back half of a run is cluttered with gimme stages that trivialize clears thanks to prominent premade matches anchoring everything to the ceiling. Presumably their reasoning for this was to give survival players ample downtime (perhaps an exceeding amount of downtime) in exchange for score-focused players to push themselves on these levels. Each bubble dropped when a match is made on its supporting pillar scales the score exponentially, and thus score players are obligated to overfill the screen with garbage before finally severing the whole board in one go. With that in mind, these layouts make more sense: level 27, for instance, has two same-colored antennae for its bug-shaped board binding it to the top. The goal here transforms from hitting those two as quickly as possible to taking one out and turning its side of the playfield into a garbage dump, leaving enough room on the other side to wipe everything at once. Others, like level 23, concentrate their clear points in the middle, asking the player to preserve a center lane or attempt to carve out a single side while keeping the other clear. Although lengthy on typical survival play, these levels take a different approach to decision-making and optimal play.

As this score run from essekappa demonstrates, Level 27 can be optimized for score by accumulating garbage on the one side before eliminating the whole thing via the antenna on the other. [src]

Of course, if the game was eminently solvable, it wouldn’t hold its flavor across multiple attempts, so just enough dynamism is infused into the premade levels to make each run a little different. While the levels have core layouts, they vary in how they randomize the colors, with some being fully random, some using random colors but following a pattern, and many using a hybrid approach of fixed and non-fixed colors. The fully random ones occasionally skew towards enabling free clears thanks to premade matches showing up where they likely shouldn’t, which diminishes the power of the system. What’s more interesting is the descent of the board downwards after a certain number of shots. The game has a proper built-in timer (the 10 seconds per shot) and leverages that to drop the board closer and closer to the player, killing them once it crosses the bottom threshold. Interestingly, the number of shots before the board lowers is based on the number of colors left; the more colors completely removed by the player, the quicker the whole thing will lower. This has the additional effect of changing ricochet angles and potentially obstructing the player into uncomfortable circumstances if they over-stacked the bottom, a keen, in-system method of scaling difficulty towards player achievement.


  1. The interview in this compilation with Seiichi Nakakuki, creator of Puzzle Bobble, mentions something about the random selection being weighted based on the “current situation,” but it’s not clear to me exactly how this works. I haven’t found any other information online that sheds light on this either. Also interesting from this article: supposedly the whole game was made in a month and a half! ↩︎



Comments

You can use your Bluesky account to reply to this post.