This website is dedicated **the search
for the smallest
Kochen-Specker system**. See the
section *The experiment* below,
for an explanation what Kochen-Specker systems are and why they are
interesting. At the moment, this website primarily lists
the results from a paper by Uijlen
and Westerbaan. Please feel
free to contribute: send
an e-mail or contribute
directly on github.

Consider the following experiment. Measure the SPIN of a spin 1 particle along a direction. Or more concretely: shoot a deuterium atom through a certain fixed inhomogeneous magnetic field. Depending on the direction it was shot through the field, the particle will move undisturbed or deviate.

Quantum Mechanics only predicts the probability, given the direction, whether the particle will deviate. Its probabilistic prediction has been tested thoroughly. One wonders: is there a deterministic theory predicting the outcome of this experiment?

Kochen and Specker have shown that any such deterministic theory must be odd: it cannot satisfy the plausible (non contextual) SPIN axiom, that is: in exactly two out of three pairwise orthogonal directions the particle will deviate.

They have proven this by giving 117 points on the sphere, representing 117 directions, for which there can be no special {0,1}-valued coloring (called a 010-coloring), which corresponds to the predictions of a deterministc theory.

After Kochen and Speckers original paper,
several people have tried to give a smaller set of points to demonstrate
this fact. Such a finite set of points is called a **Kochen-Specker
system**.

The smallest *known* Kochen-Specker system is due to Conway.
It has 31 points and is shown on the right.
We also know that the smallest Kochen-Specker system must have
at least 21 points. A short history:

by | year | bound |
---|---|---|

Kochen and Specker | 1975 | ≤ 117 |

Penrose, Peres (independently) | 1991 | ≤ 33 |

Conway | ~1995 | ≤ 31 |

Uijlen and Westerbaan | 2014 | ≥ 21 |

Arends, Ouaknine and Wampler | 2009 | ≥ 18 |

Arends, Ouaknine and Wampler observed that a *n* point
Kochen-Specker system corresponds to a certain *n* vertex graph:
a graph that is both *(sphere) embeddable* and
not *010-colorable*.
They have shown there are no such graphs on less than 18 vertices.
Uijlen and Westerbaan demonstrated that there is also no such graph
on less than 21 vertices.