时间线
- 1974布达佩斯的建筑学教授 Ernő Rubik 用木头和橡皮筋制作出第一个能用的原型，以向学生教授三维几何。他称之为“魔方”（Bűvös Kocka）——据说他花了一个多月才还原自己的发明。
- 1977魔方在匈牙利玩具店开售。起初它传播缓慢，在铁幕之后的数学家和好奇玩家间口耳相传。
- 1980Ideal Toy 公司取得该益智玩具的全球分销授权，并将其更名为“Rubik's Cube（魔方）”。它在国际上推出并赢得多项年度玩具奖。
- 1981魔方狂热达到顶峰。一年之内售出数千万个魔方，解法手册登上畅销榜，魔方成为 1980 年代流行文化的标志性符号。
- 1982首届世界魔方锦标赛在布达佩斯举办。Minh Thai 以 22.95 秒的成绩夺冠——这个数字在现代选手看来几乎是悠闲的。
- 2003世界魔方协会（WCA）成立，统一了规则和纪录，并开启了现代竞技时代。常规比赛在全球恢复。
- Today如今，像 CFOP 和 Roux 这样的方法、更快的硬件以及全球社区已把单次还原世界纪录推进到四秒以内，机器人则能在不到一秒内还原。
速拧的兴起
早期的解法者依靠简单的逐层法。随着社区壮大，出现了高级体系——最著名的是由 Jessica Fridrich 推广的 CFOP（Cross、F2L、OLL、PLL）——它们最小化步数，把还原变成一项识别与手速的运动。
世界纪录与现代时代
平均还原时间已从 1982 年的 20 多秒骤降到如今顶尖选手的 6 秒以内，单次还原降到 4 秒以下。专门打造的机器人已把绝对纪录推进到数百毫秒。
超越 3x3
魔方的成功催生了一整个扭转益智玩具家族——从 2x2 口袋魔方到巨型 NxN 魔方和异形魔方——每一种都有自己的解法和竞技圈。
How we got the language of the cube
Before solvers could share solutions, they needed a shared language. In the late 1970s the British mathematician David Singmaster created one in his Notes on Rubik's Magic Cube, and it became the standard the whole world still uses today. His insight was to name the six faces not by their colours — which differ from cube to cube — but by their fixed positions relative to the solver: U (Up), D (Down), L (Left), R (Right), F (Front), and B (Back). A single letter means a quarter-turn of that face clockwise; R2 means a half-turn, and R' means a quarter-turn counter-clockwise. As the Handbook of Cubik Math notes, the system “has been accepted internationally by most students of the cube.” Every algorithm on this site is written in Singmaster notation.
The birth of the layer-by-layer method
The first widely taught way to solve the cube is the one beginners still learn today: the layer-by-layer method. Rather than trying to fix everything at once, you build the cube one layer at a time and use a small set of repeatable move-sequences to place pieces without wrecking what you have already done. Singmaster observed early on that “many people get the whole bottom layer correct, then the middle layer, then the top layer,” and built a step-by-step algorithm around exactly that idea — one he reported could restore any cube in under 200 moves and about five minutes. This intuitive, layer-first approach is the direct ancestor of every modern beginner's guide, and the foundation that faster systems like CFOP were later built on top of.
The math of the cube: 43 quintillion positions
The cube has exactly 43,252,003,274,489,856,000 reachable positions — about 43 quintillion. The number comes from counting the pieces independently: the 8 corners can be arranged in 8! ways and the 12 edges in 12! ways; each corner can be twisted into 3 orientations (3⁸) and each edge flipped 2 ways (2¹²). But not every combination is reachable by turning faces — you can never swap a single pair of pieces, the last corner's twist is forced by the other seven, and the last edge's flip is forced by the other eleven. Those three constraints divide the total by 12 — which is why, as Singmaster put it, a randomly reassembled cube has “only a 1/12 chance of being able to get back to START.” In group-theory terms this set is the Cube Group, and that headline number is its size. To put it in scale: counting one position every microsecond would take roughly 1.4 million years.
God's Number: every cube is 20 moves from solved
No matter how badly a cube is scrambled, it can always be solved in at most 20 face turns — a value cubers nicknamed God's Number. Mathematically it is the “diameter” of the cube's puzzle graph: the moves needed in the worst case with perfect play. Finding it took decades. Morwen Thistlethwaite pioneered the computer-assisted approach in the late 1970s, solving the cube in stages (his algorithm guaranteed a solution in around 52 moves). Researchers steadily tightened the bounds: by 2008 the hardest known position (the “superflip”) was shown to need exactly 20 face turns. The two ends finally met in 2010, when Tomas Rokicki and his collaborators used massive computation to prove the answer is exactly 20. (Measured in quarter-turns only, the equivalent number is 26.)