For centuries, the cube remained the everyday die. It was easy to make, easy to read, and familiar wherever games traveled. More complicated probability often required tables, cards, chits, spinners, or several six-sided dice interpreted together.

Then a standards committee in postwar Japan gave the icosahedron a new job.

In 1950, the Japanese Standards Association formed a committee concerned with quality control and measurement. Industrial researchers needed random digits for sampling and statistical work, but printed random-number tables were cumbersome. Ishida Yasushi of Toshiba proposed an icosahedral device. Because an icosahedron has twenty faces, the digits 0 through 9 could appear twice. Roll differently colored dice together and the results could form two- or three-digit random numbers.

The instrument was developed in 1950 and 1951 and advertised for sale by June 1952. It was not a fantasy accessory. It was a compact scientific tool intended for quality control, biometrics, Monte Carlo methods, and other work that needed decimal randomness.

It was clever, useful, and not perfectly made. Mathematician C. B. Tompkins physically examined Japanese Standards Association sets in a 1961 review. He noted imperfect casting, then reported no meaningful bias for their intended pace of human use. The modern argument about whether a die’s manufacturing flaws affect its fairness had already begun.

The decimal icosahedron moved from technical circles toward games through magazines, military simulation, and hobby networks. Scientific American described the Japanese dice in 1968. Wargamer’s Newsletter mentioned commercially available twenty-sided random generators in 1969. The appeal to wargamers was immediate. A d20 marked 0 through 9 twice could generate percentage results far more cleanly than awkward combinations of ordinary cubes.

This mattered because postwar wargamers increasingly wanted detailed models of weapons, armor, training, terrain, and battlefield outcomes. A designer might need a five-percent chance, a thirty-percent chance, or a table with twenty distinct results. One icosahedron could deliver those steps directly.

American hobbyists still had a supply problem. In 1968, Len Lakofka discussed the potential of the shape while also explaining how readers could construct one from cardboard or wood. The idea had arrived before the merchandise.

Tractics helped close that gap. Published in 1971, the modern-warfare rules by Mike Reese and Leon Tucker, with Gary Gygax, used percentile-style resolution. Gygax wrote about the expected use of an icosahedron with the game. Before the d20 symbolized a dragon, it helped determine whether a tank round struck its target.

In 1972, Creative Publications of Palo Alto offered an inexpensive educational set based on the five Platonic solids: tetrahedron, cube, octahedron, dodecahedron, and icosahedron. Don Lowry stocked the sets for Tractics customers. Suddenly American hobbyists did not have to build a d20 or import a scientific instrument. They could buy a bag containing several unfamiliar shapes.

The economical set created a design opportunity. If the wanted die was the icosahedron, what could a game do with the other four?

Gygax explored that question in his June 1973 article “Four & Twenty and What Lies Between.” During the same period, he and Dave Arneson were turning experiments in fantasy wargaming and character-centered play into a publishable game.

The polyhedral set did not spring from a dragon’s hoard. It traveled from geometry to statistics, from statistics to military simulation, and from military simulation to fantasy.

The next step would make those shapes inseparable from role-playing.

AI-Assisted: This article was developed with AI-assisted research, fact-checking, and editorial support. Ken Whitman reviewed and approved the final text.

Copyright: Copyright © 2026 Ken Whitman. All rights reserved.

Sources Used

• Peter D. Evan, “An Elegant Little Instrument” (2024).
• C. B. Tompkins, Mathematics of Computation 15.73 (1961).
• Ian Tonat, “The 20-Sided Die,” American Historical Association (2022).
• Jon Peterson, “How Gaming Got Its Dice” (2013).