--- title: 'Ramsey-Good Graphs: Exact Chromatic Bounds' url: https://www.emergentmind.com/topics/ramsey-good-graphs type: topic --- # Ramsey-Good Graphs: Exact Chromatic Bounds In graph Ramsey theory, “Ramsey-good” most commonly refers to exact attainment of Burr’s lower bound for an off-diagonal Ramsey number: if \(G\) is connected, then \(G\) is called \(H\)-good when \[ R(G,H)=(|G|-1)(\chi(H)-1)+\sigma(H), \] where \(\chi(H)\) is the chromatic number of \(H\) and \(\sigma(H)\) is the size of the smallest color class in a proper \(\chi(H)\)-coloring of \(H\) [1512.07874]. Equivalent formulations in the literature swap the roles of the two graphs and write that a connected graph \(H\) is \(G\)-good if \[ r(G,H)=(\chi(G)-1)(|H|-1)+s(G), \] where \(s(G)\) is the chromatic surplus, i.e. the minimum size of a color class in a proper \(\chi(G)\)-coloring of \(G\) [2506.06724]. This notion, introduced by Burr and Erdős and developed across exact, asymptotic, multipartite, random-host, and coloring-based variants, isolates graph families whose Ramsey numbers are governed by the chromatic obstruction alone rather than by additional extremal structure [1512.07874]. ## 1. Classical definition and the Burr lower bound The standard framework begins with the Ramsey number \(R(G,H)\), the least \(N\) such that every red-blue coloring of the edges of \(K_N\) contains a red copy of \(G\) or a blue copy of \(H\) [1512.07874]. If \(G\) is connected, Burr’s lower bound states that \[ R(G,H)\ge (|G|-1)(\chi(H)-1)+\sigma(H), \] where \(\sigma(H)\) is the smallest color-class size among proper \(\chi(H)\)-colorings of \(H\) [1512.07874]. The usual extremal construction is the disjoint union of \(\chi(H)-1\) red cliques of size \(|G|-1\) together with one further red clique of size \(\sigma(H)-1\), with all cross-edges blue; this avoids a red connected copy of \(G\) and also avoids a blue copy of \(H\) [1512.07874]. A connected graph \(G\) is \(H\)-good if equality holds in this lower bound [1512.07874]. The equivalent formulation used in several more recent papers fixes the first graph and writes, for connected \(H\), \[ r(G,H)\ge (\chi(G)-1)(|H|-1)+s(G), \] with equality defining \(H\) as \(G\)-good [2506.06724]. Here \(s(G)\), called the chromatic surplus, is the minimum size of a color class in a proper \(\chi(G)\)-coloring of \(G\) [2506.06724]. The two formulations encode the same phenomenon: the Ramsey number is as small as the chromatic lower bound permits. The special case \(H=K_p\) recovers the older language of \(p\)-goodness. Since \(\chi(K_p)=p\) and \(\sigma(K_p)=1\), the formula becomes \[ R(G,K_p)=(p-1)(|G|-1)+1, \] and a connected graph \(G\) satisfying this is \(p\)-good [2109.09205]. This clique-versus-sparse-graph regime historically supplied the first major examples of Ramsey-goodness. ## 2. Foundational examples and asymptotic general theorems The model theorem is Chvátal’s result that every tree is \(K_m\)-good [2201.04884]. In the language of paths and trees, the theory was sharpened substantially by Pokrovskiy and Sudakov, who proved that \[ R(P_n,H)=(n-1)(\chi(H)-1)+\sigma(H) \qquad \text{for } n>4|H|, \] so sufficiently long paths are \(H\)-good for every fixed graph \(H\) [1512.07874]. Their proof proceeds through the stronger statement \[ R(C_{\ge n},K_{m_1,\dots,m_k})=(k-1)(n-1)+m_1 \] under the condition \(n>3m_k+5m_{k-1}\), where \(m_1\le \cdots \le m_k\) are the part sizes of a complete \(k\)-partite graph [1512.07874]. This yields a uniform linear threshold in \(|H|\), and for balanced \(H\) the same argument gives goodness already for \(n\ge 8|H|/\chi(H)\) [1512.07874]. The same theme extends from paths to bounded-degree trees. Balla, Pokrovskiy, and Sudakov proved that for all \(\Delta\) and \(k\) there exists a constant \(C_{\Delta,k}\) such that any tree \(T\) with \(\Delta(T)\le \Delta\) is \(H\)-good whenever \[ |T|>C_{\Delta,k}|H|\log^4 |H| \] and \(\chi(H)=k\) [1611.02688]. They also established a stronger many-leaves theorem: if \(T\) has \(\ell\) leaves and maximum degree at most \(\Delta\), and \[ \ell \ge 13\Delta |H|+1, \] then \(T\) is \(H\)-good [1611.02688]. The proof combines a tree dichotomy—many leaves or many disjoint bare paths—with expansion, generalized Haxell-type embedding, and linked systems for long path insertion [1611.02688]. A different asymptotic direction concerns graphs \(F_n\) and \(G_n\) of the same order. For bounded-degree \(G_n\) with \(\alpha(G_n)\le n/4\), it was shown that \[ R(P_n,G_n)=(\chi(G_n)-1)n+\sigma(G_n)+o(n), \] so \(P_n\) is asymptotically \(G_n\)-good under bounded maximum degree and a small-independence condition [1407.7092]. This is not exact eventual equality, but it identifies a broad same-order regime in which the Burr lower bound is asymptotically correct [1407.7092]. These results fit into a more structural picture developed by Allen, Brightwell, and Skokan. They showed that bounded degree alone does not force always-good behavior, but bounded degree together with sufficiently small bandwidth does: if \(G\) is connected, \(\Delta(G)\le \Delta\), and \(\operatorname{bw}(G)=o(|G|)\), then for every fixed \(H\), \[ R(G,H)=(\chi(H)-1)(|G|-1)+\sigma(H) \] for all sufficiently large \(G\) in the class [1010.5079]. They also proved the same conclusion without any maximum-degree restriction when the bandwidth is at most \(\epsilon(H)\log n/\log\log n\) [1010.5079]. This identifies expansion, rather than bounded degree itself, as the core obstruction to goodness [1010.5079]. ## 3. Exact families beyond the classical tree-versus-clique setting A major recent direction studies specific sparse target families against non-complete fixed graphs. One example is the generalized fan \(K_1+nH\), formed by joining one hub to \(n\) disjoint copies of a fixed connected graph \(H\). Let \(h=|H|\) and \(\ell=r(K_p,H)\). It was proved that \(K_1+nH\) is \(p\)-good if either \[ n\ge \frac{2p\ell}{h} \] or \[ n\ge \max\left\{\frac{(p+2)\ell}{h},\,\ell-h\right\}, \] improving an earlier Chung–Lin threshold with constant \(c\approx 52.456\) in \[ n\ge c\,\frac{p\ell}{h} \] down first to \(3\), then \(2\), and finally to the \(1\)-scale in \(p\ell/h\) [2310.13204]. In the special case \(H=K_2\), so \(K_1+nH\) is the fan \(F_n\), the paper gives the sharper bound \[ r(K_p,F_n)=2(p-1)n+1 \qquad\text{for } n\ge \frac{p^2-p-2}{2} \] with \(p\ge 3\) [2310.13204]. Books form another central family. For the \(k\)-book \(B_{k,n}\), consisting of a spine \(K_k\) and \(n-k\) pages adjacent to the spine and independent from one another, Fox, He, and Wigderson proved that \[ r(K_p,B_{k,n})=(p-1)(n-1)+1 \qquad \text{for } n\ge 2^{\,k^{10p}}, \] so sufficiently large books are \(p\)-good [2109.09205]. This avoids the regularity method and replaces earlier tower-type thresholds with an explicit doubly exponential bound [2109.09205]. The same paper also proves an exact multipartite-versus-book result: if \(G=K_p(a_1,\dots,a_p)\) with \(1\le a_1\le \cdots \le a_{p-1}\le t\), \(a_p\le \delta n\), and \(a_1=a_2=1\), then \[ r(G,B_{k,n})=(p-1)(n-1)+1 \] for sufficiently small explicit \(\delta\), with \[ \delta\ge 2^{-\,t^{1000k^2p^2}} \] available from their proof [2109.09205]. A separate line concerns disjoint unions of cliques. For trees \(T_n\), it was proved that \[ R(T_n,2K_m)=(n-1)(m-1)+2, \] so every tree is \(2K_m\)-good [2201.04884]. From this, together with a leaf-deletion criterion, one obtains \[ R(T_n,K_m\cup K_l)=(n-1)(m-1)+1 \qquad (m>l\ge 2), \] so every tree is \(K_m\cup K_l\)-good [2201.04884]. The same paper extends goodness from connected components to disconnected unions: if every component of a disconnected graph \(\mathcal F\) is \(H\)-good, then \(\mathcal F\) is \(H\)-good in the corresponding Gould–Jacobson sense [2201.04884]. Goodness for complete multipartite graphs with one large part has also been classified asymptotically. For \[ K_{p+1}(\alpha;n)=K_{\alpha,\dots,\alpha,n}, \] where \(p\) parts have size \(\alpha\) and one part has size \(n\), and for fixed \(G\) with \(\chi(G)=k+1\), \(s(G)=1\), \(v(G)=m\), the paper “Ramsey goodness of complete multipartite graphs with one large part” gives a full asymptotic if-and-only-if classification for \(p\ge \mathrm{snd}(\alpha)\): \(K_{p+1}(\alpha;n)\) is \(G\)-good for large \(n\) exactly when \[ G \subseteq mT+K_{k-1}(m) \] for every tree \(T\) with \[ v(T)=\mathrm{snd}(\alpha), \] where \(\mathrm{snd}(\alpha)\) is the smallest non-divisor of \(\alpha\) [2605.26826]. The dependence on \(\mathrm{snd}(\alpha)\) shows that divisibility, not merely chromatic data, can control multipartite goodness [2605.26826]. ## 4. Chromatic surplus, non-clique fixed graphs, and new exact phenomena Most earlier star-like goodness results assumed the fixed graph had chromatic surplus \(1\). A recent advance breaks this restriction by studying the Hajós graph \(H_a\), which satisfies \[ |V(H_a)|=6,\qquad \chi(H_a)=3,\qquad s(H_a)=2 \] and therefore induces the Burr bound \[ R(H_a,G)\ge 2(|V(G)|-1)+2 \] for connected \(G\) with \(|V(G)|\ge 2\) [2506.06724]. For the star \(K_{1,n}\), the exact Ramsey number is \[ R(H_a,K_{1,n})= \begin{cases} 2n+2 & \text{for even } n\ge 2,\\ 2n+3 & \text{for odd } n\ge 3, \end{cases} \] so \(K_{1,n}\) is \(H_a\)-good if and only if \(n\) is even [2506.06724]. The lower bound in the odd case comes from the explicit obstruction \[ (\ell K_2)+(\ell K_2),\qquad \ell=(n+1)/2, \] which has \(2n+2\) vertices, contains no \(H_a\), and whose complement has maximum degree \(n-1\), hence contains no \(K_{1,n}\) [2506.06724]. This gives a genuine parity obstruction to goodness in the surplus-\(2\) setting [2506.06724]. The same paper proves that the fan \(F_n\) is \(H_a\)-good for all \(n\ge 111\): \[ R(H_a,F_n)=4n+2 \qquad (n\ge 111). \] Its proof uses the external input \[ R(W_4,F_n)=4n+1 \qquad (n\ge 111), \] a matching analysis in the large-\(\Delta(\overline G)\) case, and a detailed wheel-based decomposition in the complementary case [2506.06724]. The threshold \(111\) is stated explicitly as not best possible [2506.06724]. Odd cycles against complete bipartite graphs provide another exact recent family. For odd \(m\ge 7\), the paper “A study of two Ramsey numbers involving odd cycles” proves that \[ R(\mathbb K_{2,n},C_m)=2n+3 \] whenever \[ n\ge 2m+499,\qquad n\ge 3493, \] showing that \(\mathbb K_{2,n}\) is \(C_m\)-good in that range [2504.15693]. Since \(|\mathbb K_{2,n}|=n+2\), \(\chi(C_m)=3\), and \(\sigma(C_m)=1\), the Burr formula indeed predicts \(2n+3\) [2504.15693]. The upper bound rests on common-neighborhood control in \(K_{2,n}\)-free graphs, specifically \[ |N_G(x)\cap N_G(y)|\le n-1 \] for all \(x,y\), plus cycle-richness lemmas for 2-connected non-bipartite graphs [2504.15693]. These results collectively show that once one moves beyond fixed graphs with chromatic surplus \(1\), goodness can remain exact but the obstruction landscape changes substantially. This suggests that surplus \(>1\) is not a minor perturbation of the classical theory but a distinct regime with its own structural exceptions [2506.06724]. ## 5. Random-host analogues and sparse ambient graphs A different extension asks not for the complete graph \(K_N\) to force the Ramsey alternative, but for a sparse random host \(G(N,p)\). In this setting, the relevant statement is that the random host behaves like a Ramsey-good ambient graph up to a natural additive correction. For paths versus cliques, Moreira showed that \[ G\big((1+\varepsilon)rn,p\big)\to (K_{r+1},P_n) \qquad\text{w.h.p. if } p\gg n^{-2/(r+1)}, \] and also that \[ G(rn+t,p)\to (K_{r+1},P_n) \qquad\text{w.h.p. if } p\gg n^{-2/(r+2)} \text{ and } t\gg 1/p \] [1909.00030]. Since the deterministic good threshold is \(R(K_{r+1},P_n)=rn+1\) for large \(n\), these results show that sparse random graphs on essentially the same number of vertices already force the same pair once \(p\) is above the appropriate threshold [1909.00030]. The paper also proves matching negative statements, including that \(t\ll 1/p\) is insufficient in the near-critical regime [1909.00030]. For bounded-degree trees, Bucić, Letzter, Sudakov, and Tran proved a random analogue of Chvátal’s theorem. Writing \(\mathcal T(n,D)\) for the family of trees with \(n\) edges and maximum degree at most \(D\), they showed that for each \(r,D\ge 2\) there exist constants \(C,C'>0\) such that if \[ p\ge C n^{-2/(r+2)} \qquad\text{and}\qquad N\ge rn + C'/p, \] then \[ G(N,p)\rightarrow \big(K_{r+1},\mathcal T(n,D)\big) \] with high probability [2001.03083]. Since Chvátal’s deterministic theorem gives \[ K_N \rightarrow (K_{r+1},T)\iff N\ge rn+1 \] for every tree \(T\) with \(n\) edges, this is a near-exact random-host analogue with the additive \(C'/p\) term representing the correct sparse correction scale [2001.03083]. The proof develops a stability theorem showing that any bad coloring must resemble the extremal \(r\)-partite obstruction up to an exceptional set of size \(O(1/p)\) [2001.03083]. These random-host results do not redefine goodness, but they transfer the goodness heuristic from complete hosts to sparse pseudorandom ambient graphs. A plausible implication is that the classical Burr–Erdős framework has a robust probabilistic counterpart whenever the deterministic extremal colorings are sufficiently stable [1909.00030]. ## 6. Alternative meanings of “good” and adjacent Ramsey notions The term “good” is not uniform across Ramsey theory. Several papers use it in related but formally distinct senses. In mixed Ramsey theory, an edge-coloring of \(K_n\) is \((G,H)\)-good if it contains no monochromatic copy of \(G\) and no rainbow copy of \(H\). The set of achievable numbers of colors in such colorings is the mixed-Ramsey spectrum \[ S(n;G,H). \] Axenovich and Iverson proved that if \(G\) is not a star and \(\delta(H)\ge 2\), then \(S(n;G,H)\) is an interval for every \(n\) [1005.2629]. They also exhibited the exceptional finite example \[ S(10;C_4,K_3+e)=\{3,7,8,9\}, \] showing that interval behavior can fail once \(H\) has a pendent edge [1005.2629]. Here “good” describes a coloring, not a graph attaining Burr’s lower bound [1005.2629]. A different chromatic-host notion appears in “Good Graph Hunting.” A tuple \((H_1,\dots,H_k)\) is called good if every \(k\)-edge-coloring of an \(R(H_1,\dots,H_k)\)-chromatic graph contains a monochromatic \(H_i\) in color \(i\) for some \(i\) [1508.01833]. In this framework, a graph \(H\) is \(k\)-good if \((H,\dots,H)\) is good, and good if it is \(k\)-good for every \(k\) [1508.01833]. The paper proves that stars are good, that \(P_4\) is \(k\)-good for \(k\neq 3\), that \((P_4,P_5)\) is good, and that \(P_5\), \(P_6\), and \(P_7\) are \(2\)-good [1508.01833]. This is a chromatic-host strengthening of ordinary Ramsey theorems, not Burr-type goodness [1508.01833]. A third usage occurs in extremal graph constructions. In “Doubly Saturated Ramsey Graphs,” a graph \(G\) is called \(R(s,t)\)-good if it contains neither a clique of size \(s\) nor an independent set of size \(t\), equivalently \[ \omega(G)2^{c\Delta} \] for some \(c>0\), while bounded degree together with sublinear bandwidth restores exact off-diagonal goodness [1010.5079]. Their paper also shows that Burr’s diagonal conjecture fails even for path powers \(P_n^k\) [1010.5079]. Another measure of sparse Ramsey structure is the Ramsey density \[ m^*(F,r)=\inf\{m(G): G \text{ is } (F,r)\text{-Ramsey}\}, \qquad m(G)=\max_{H\subseteq G}\frac{e(H)}{v(H)}. \] For even cycles, it was shown that \[ r-\varepsilon \le m^*(C_\ell,r)< r+1, \] and for complete bipartite graphs \(K_{a,b}\) with \(b\ge (a-1)^2+1\), \[ r(a-1)-\varepsilon \le m^*(K_{a,b},r)< r(a-1)+1 \] [1108.1102]. For paths, \[ \left\lceil \frac r2\right\rceil-\varepsilon \le m^*(P_\ell,r) < \left\lceil \left(1-\frac{1}{\lfloor \ell/2\rfloor}\right)r +\frac{1}{\lfloor \ell/2\rfloor} \right\rceil \] [1108.1102]. These are not goodness results in the Burr sense, but they show that sparse Ramsey hosts can be much sparser than complete graphs, especially for bipartite targets [1108.1102]. A further negative perspective comes from algebraic constructions. If an algebraic graph of complexity \((n,d,m)\) on \(N\) vertices is defined by zero-patterns of boundedly many bounded-degree polynomials in bounded dimension, then it must contain either a clique or an independent set of size at least \[ c'N^{1/\gamma}, \qquad \gamma = cnm \min\left\{ d,\ \frac{n\log d}{\log n}\right\} \] [2103.05618]. Thus bounded-complexity algebraic graphs have polynomial-size homogeneous sets, whereas the Erdős random-graph benchmark has largest clique and independent set of order \(2\log_2 N\) [2103.05618]. This shows that any algebraic family with genuinely strong Ramsey-type pseudorandomness must let at least one of the parameters \(n\), \(d\), or \(m\) grow [2103.05618]. Across the recent literature, several open problems recur. The precise smallest \(n\) for which \(F_n\) is \(H_a\)-good remains open below the current threshold \(111\) [2506.06724]. For books, the exact Ramsey-goodness threshold between the lower bound \((k/\log p)^{cp}\) and the upper bound \(2^{\,k^{10p}}\) is open [2109.09205]. For bounded-degree trees, the conjectured linear theorem \[ |T|>C_{\Delta,k}|H| \implies T \text{ is } H\text{-good} \] would remove the \(\log^4 |H|\) factor from the current best general result [1611.02688]. For mixed Ramsey spectra, it remains open whether there exist graph pairs \((G,H)\) for which non-interval behavior persists for arbitrarily large \(n\) [1005.2629]. Taken together, these works show that Ramsey-goodness is neither a single theorem nor a single class, but a broad organizing principle. In its classical Burr–Erdős form, it identifies graph families whose Ramsey numbers are controlled exactly by chromatic data; in modern developments, it interacts with chromatic surplus, multipartite structure, expansion, random hosts, and several adjacent notions of “goodness” used elsewhere in Ramsey theory [1512.07874]. Source: https://www.emergentmind.com/topics/ramsey-good-graphs