Fibonacci & Lucas Convergence Reference

Almost every popular claim about the golden ratio and art, architecture, or the human body turns out to be exaggerated or unsupported once someone checks it against the historical record — covered at length elsewhere on this site. This table is the opposite case: a golden-ratio claim that is not only true, but provably, exactly true, with no measurement, photograph, or overlay involved anywhere.

The ratio column shows F(n+1) divided by F(n) at every step, and the error column shows exactly how far that ratio sits from φ = 1.6180339887… The early rows oscillate — above φ, then below, then above again — before settling into a steady, monotonic shrink from roughly n=5 onward. That shrink isn’t gradual arithmetic; each step’s error is smaller than the previous step’s by a factor very close to φ² (≈2.618), which is why the error collapses from a difference you could measure with a ruler at n=1 down to a difference smaller than a computer’s floating-point precision by n=30.

The Lucas column tells the same story from a different starting point. L(n) satisfies the identical recurrence as F(n) — each term the sum of the two before it — but starts from 2 and 1 instead of 0 and 1. Its own consecutive-term ratio (not shown as a separate column here, but computable the same way from the L(n) values above) converges to the exact same φ, at essentially the exact same rate, which is the clearest evidence that this convergence is a property of the addition rule itself and not a special feature of Fibonacci’s specific starting numbers.