## 1. Precise restatement and assumptions Let \(m\ge 1\) be an integer, and let the distinct papers be \[ N=\{a,c,b,d_1,\ldots,d_m\}. \] Define the two routes \[ A=\{a,c\},\qquad B=\{b,d_1,\ldots,d_m\}. \] The model requires these assumptions: - **Exactly these two routes determine success:** an archive \(S\subseteq N\) establishes the discovery if and only if \(A\subseteq S\) or \(B\subseteq S\). There are no additional validation routes. - All papers eventually appear, and all \((m+3)!\) publication orders are equally likely. In particular, there are no chronological prerequisite constraints. - Success is permanent once achieved. - The single paper whose arrival first changes the archive from unsuccessful to successful receives all priority credit. There are no ties or shared credit. - “Standalone success” means whether the singleton archive containing that paper succeeds. - “Final indispensability” means whether removing the paper from the full archive destroys success. Writing \(p_a,p_b\) for expected credit, the precise quantitative claim is \[ \frac{p_a}{p_b}=\frac{(m+1)^2}{4}. \] Thus **the ratio**, not the absolute credit difference, grows quadratically with \(m\). ## 2. Checking the argument and independently deriving the result ### Retrospective assessments For \(m\ge1\), neither \(\{a\}\) nor \(\{b\}\) contains a complete route, so both have zero standalone success. Removing \(a\) from the completed archive leaves \(B\); removing \(b\) leaves \(A\). Hence neither is indispensable to the final archive. These statements are correct. They establish equality on two binary assessments, not equality of all possible contributions. ### Checking the integrals Independent uniform arrival times generate a uniformly random publication order. Conditional on \(a\) arriving at time \(t\), it wins exactly when: 1. \(c\) has arrived, with probability \(t\); and 2. not all \(m+1\) papers of \(B\) have arrived, with probability \(1-t^{m+1}\). The events concern disjoint sets of independent arrival times. Consequently, \[ p_a=\int_0^1t(1-t^{m+1})\,dt =\frac12-\frac1{m+3} =\frac{m+1}{2(m+3)}. \] Likewise, conditional on \(b\) arriving at \(t\), it wins exactly when all \(m\) papers \(d_i\) have arrived but \(a,c\) have not both arrived: \[ p_b=\int_0^1t^m(1-t^2)\,dt =\frac1{m+1}-\frac1{m+3} =\frac{2}{(m+1)(m+3)}. \] Therefore, \[ \frac{p_a}{p_b}=\frac{(m+1)^2}{4}. \] Every step is valid under the stated model. ### Independent derivation: inspect the final publication The two disjoint routes partition the archive. Therefore: - If the last publication belongs to \(B\), route \(A\) necessarily completes first. - If the last publication belongs to \(A\), route \(B\) necessarily completes first. Since the final paper is uniformly distributed over \(N\), \[ \Pr(A\text{ completes first})=\frac{m+1}{m+3}. \] Within \(A\), symmetry gives \(a\) half this probability: \[ p_a=\frac12\frac{m+1}{m+3}. \] Similarly, \[ \Pr(B\text{ completes first})=\frac2{m+3}. \] Its \(m+1\) members are symmetric, giving \[ p_b=\frac1{m+1}\frac2{m+3}. \] This independently proves the formulas. As a normalization check, \[ 2p_a+(m+1)p_b=1, \] as required because exactly one paper receives credit. ## 3. Attempts to break it ### Small numerical example Take \(m=2\). The routes are \(\{a,c\}\) and \(\{b,d_1,d_2\}\), with \(5!=120\) equally likely orders. - The last paper belongs to the three-paper route in \(72\) orders. By symmetry, \(a\) wins in \(36\). - The last paper belongs to the two-paper route in \(48\) orders. By symmetry, \(b\) wins in \(16\). Thus \[ p_a=\frac{36}{120}=\frac3{10},\qquad p_b=\frac{16}{120}=\frac2{15}, \] and their ratio is \(9/4\), as claimed. ### Boundary cases - **\(m=0\):** \(b\) alone establishes success, contradicting the asserted equal standalone insufficiency. Thus \(m\ge1\) is necessary. - **\(m=1\):** both routes have length two, and all four papers receive probability \(1/4\). There is no difference. - **\(m\to\infty\):** \[ p_a\to\frac12,\qquad p_b\sim\frac2{m^2}. \] The ratio becomes arbitrarily large. However, the absolute difference is \[ p_a-p_b=\frac{m-1}{2(m+1)}\longrightarrow\frac12. \] So an interpretation asserting a quadratically growing **absolute gap** would be incorrect. ### Conditions that matter The formula is not implied merely by the existence of the two displayed routes. For example, with \(m=2\), add a third minimal route \(\{a,b\}\). Neither paper becomes sufficient alone or indispensable to the completed archive, but the probabilities change to \[ p_a=\frac5{12},\qquad p_b=\frac14, \] giving ratio \(5/3\), not \(9/4\). Thus “two routes” must mean **exactly two minimal sufficient routes**. Likewise, overlapping routes or nonuniform publication orders need not obey this formula. Identical marginal publication-time distributions alone do not guarantee uniformly random orders; dependence can matter. These are limitations of the model, not counterexamples to the properly specified claim. Under its assumptions, the proof is complete. ## 4. Novelty The mathematical mechanism is a direct instance of the **Shapley–Shubik power index**: a player's index is the probability of being pivotal in a uniformly random ordering. Equivalently, these expected credits are the **Shapley values** of the binary success function. There is also a short derivation using standard unanimity games. Let \[ u_T(S)=\mathbf 1\{T\subseteq S\}. \] Since the routes are disjoint and exhaust \(N\), \[ v=u_A+u_B-u_N. \] In a unanimity game \(u_T\), each member of \(T\) has Shapley value \(1/|T|\). By linearity, \[ \phi_a=\frac12-\frac1{m+3},\qquad \phi_b=\frac1{m+1}-\frac1{m+3}, \] which are precisely the claimed probabilities. More generally, for two disjoint routes of lengths \(k\) and \(\ell\), a member of the first receives \[ \frac{\ell}{k(k+\ell)}, \] while a member of the second receives \[ \frac{k}{\ell(k+\ell)}. \] Their ratio is \((\ell/k)^2\). Accordingly, this is **a special case or application of known results**, rather than a new mathematical phenomenon. I cannot determine from the supplied material whether its particular framing in scientific attribution is new. ## 5. Fairness of the one-sentence claim **Yes, as an existential claim about this stylized credit rule.** The construction demonstrates exactly what it says, provided “differ quadratically” means the credit ratio. Two qualifications matter: - Equal standalone insufficiency and final dispensability are only two coarse binary comparisons; they do not make the papers structurally equivalent. - The result demonstrates a possible mechanism for unequal prominence, not its empirical importance in actual scientific history. The supporting text appropriately avoids concluding that the credit allocation is unjust. {"verdict":"correct with caveats","novelty":"special case or variant of known results","fair_summary":"yes","confidence":5,"one_line":"The quadratic credit ratio is rigorously correct for exactly two disjoint routes under uniformly random publication order, and is a direct application of Shapley–Shubik pivotality."}