AM–GM Inequality¶
For any nonempty finite list of nonnegative real numbers, its geometric mean does not exceed its arithmetic mean, with equality exactly when all entries are equal.
Core Idea¶
For \(n\geq1\) nonnegative real numbers \(a_1,\ldots,a_n\), the AM–GM inequality says $$ (a_1\cdots a_n)^{1/n}\leq\frac{a_1+\cdots+a_n}{n}, $$ with equality exactly when all entries are equal. Zero entries remain within the theorem, although a proof taking their logarithms would not be valid. This is a typed comparison of the geometric and arithmetic means of the same equally weighted list.[^ref-0088f7fdfc36]
Scope of Application¶
With a fixed nonnegative sum \(S\), the product cannot exceed \((S/n)^n\); equality requires equal entries. For a rectangle of perimeter $20$, sides \(x+y=10\) yield \(xy\leq25\), attained by the $5\(-by-\$5\) square. This is a direct two-variable derivation from the sourced inequality, not a general rule that every equal allocation remains feasible under other constraints.[ref-864647b77bd5][ref-0088f7fdfc36]
In a separate financial-summary setting, use nonnegative gross period factors \(g_i=1+r_i\), not possibly negative net rates themselves. Their geometric mean minus one describes an equivalent compounded net rate; their arithmetic mean minus one is the ordinary average net rate. CFA's historical return tables report both, including US real equities in 1900–2015 at 6.4% geometric and 8.3% arithmetic. This is descriptive mathematics, not investment advice.[ref-09f712c47a67][ref-0088f7fdfc36]
Clarity¶
For two nonnegative entries, \((x+y)/2\geq\sqrt{xy}\) follows from \((x-y)^2\geq0\), with equality only at \(x=y\). In the full finite-list result, any zero factor makes the geometric mean zero; equality with the nonnegative arithmetic mean then requires every entry to be zero. A log-concavity proof must first assume strictly positive entries, then treat zero separately or by a limit.[ref-864647b77bd5][ref-0088f7fdfc36]
Manages Complexity¶
AM–GM replaces many possible products at a fixed sum with one sharp bound and a named equality case. In the rectangle, the two side lengths have a fixed additive total, so the area product is bounded without checking every shape. For sequential gross returns, it condenses a product path into a comparable per-period factor, while leaving the interpretation of that factor to the financial setting.[ref-0088f7fdfc36][ref-09f712c47a67]
Abstract Reasoning¶
Identify the exact nonnegative list, verify both means use it with equal weights, and apply \(G\leq A\). If solving a fixed-sum or fixed-positive-product problem, rearrange only after checking the domain and whether the equal-entry equality case is feasible. An illustrative pair of gross factors $1.5$ and $0.5$ has arithmetic gross mean $1$ but geometric gross mean \(\sqrt{0.75}\approx0.866\): net summaries are $0\%$ versus roughly \(-13.4\%\) per period. This constructed calculation is not a historical observation or a forecast.[ref-0088f7fdfc36][ref-09f712c47a67]
Knowledge Transfer¶
The identical order/equality pattern travels from rectangle sides to compounded gross factors, but the applications ask different questions. A fixed perimeter is a design constraint; a realized return sequence is a measurement record. Weighted AM–GM is a broader generalization, and Log-Sum Inequality is a distinct relation. General prime Comparison does not itself supply AM–GM's numeric domain and sharp order, so the staged DAG node is unparented. A more portable additive-versus-multiplicative aggregate pattern remains a future-prime question.[^ref-0088f7fdfc36]
[^ref-0088f7fdfc36]: Manuel Eberl, “Pólya’s Proof of the Weighted Arithmetic–Geometric Mean Inequality”, Archive of Formal Proofs (2022), abstract and unweighted corollary. [^ref-864647b77bd5]: Anders Björner and Richard P. Stanley, A Combinatorial Miscellany (2010), Chapter 6, Eq. (6.3), PDF p. 55. [^ref-09f712c47a67]: David Chambers and Elroy Dimson, eds., Financial Market History, CFA Institute Research Foundation (2016), Chapter 1 appendix introduction, PDF p. 37, and Appendix 1.2, PDF p. 39.
Neighborhood in Abstraction Space¶
AM–GM Inequality sits in a sparse region of the domain-specific corpus (71st percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.
Family — Codes, Matrices & Combinatorial Problems (30 abstractions)
Nearest neighbors
- Maclaurin's Inequality — 0.86
- Gram Matrix — 0.83
- Sum-product number — 0.83
- Reduction operator — 0.83
- Log-Sum Inequality — 0.83
Computed from structural-signature embeddings · 2026-10-08