El titular es mi versión personal extrapolada de la siguiente cita:
“Over the past three decades, the progress in algorithms for the simulation of the Ising model has outperformed Moore’s law: running modern algorithms on 30 year old computers would be faster than running 30 year old algorithms on the fastest supercomputers of today!” L.D. Landau.
Extraída del artículo de Vinay Ambegaokar, Matthias Troyer, “Estimating errors reliably in Monte Carlo simulations of the Ehrenfest model,” Am. J. Phys. 78: 150-157, Feb. 2010 [gratis en ArXiv]. Por cierto, ya que estamos, me permito extraer también lo siguiente:
“Stanislav Ulam invented the Monte Carlo method in the 1940s when playing Solitaire while lying sick in bed. He wanted to know the probability of winning in Solitaire but was faced with the problem that with 52! ≈ 1068 different ways of arranging the cards he could never exactly calculate the chance of winning. He realized, however, that by just playing 100 games and counting the number of wins he could already get a pretty good estimate.” R. Eckhardt.