{"id":"eddcf5ed-6871-40f9-981f-258ed4bcc828","arxiv_id":"1908.07109","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":5.0,"correctness_risk":"low","formal_verification":"none","parameter_count":0,"one_line_summary":"A closed-form optimization shows that two coupled unstable second-order systems have maximum unit-energy reachable volume when the ratio of time constants equals the silver ratio, 1+√2.","lead":"This note shows that for two unstable second-order systems driven by one common input, the volume of states reachable with unit energy is maximized when the ratio of their time constants equals the silver ratio, 1+√2. The result gives a simple design rule for inverted-pendulum type systems, though it depends on the chosen coordinates and on input gains that do not vary with the time constants.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Coordinate dependence, not v_i independence, is the load-bearing gap: in modal coordinates the optimum is (√13−2)/3, so the silver ratio depends on the chosen state scaling.","rationale":"The reader's weakest_assumption correctly identified both the v_i-independence assumption and the coordinate dependence of the volume measure. My stress-test confirms the coordinate dependence is the more concrete and load-bearing issue: it survives even when v_i are constants, and it changes the numerical optimum. The paper itself acknowledges coordinate dependence in Section I, so the mathematical derivation is not unsound; however, the abstract and conclusion overstate the rule's scope by omitting the canonical-coordinate qualification. This is precisely a presentation/scope problem, not a mathematical error, so the reader's CONDITIONAL verdict remains appropriate. I would not escalate to REJECT because the core optimization is correct under the stated assumptions, and the paper contains a clear caveat in the introduction. The concrete modal-coordinate check provides a sharp way to demonstrate why the caveat matters and to force a more careful statement of the contribution. The v_i-independence assumption is less damaging because the worked example (20)–(21) has v_1 and v_2 independent of π_i, provided the input is the physical force component as written. Hence I partially agree with the reader: I share the coordinate-dependence concern and the need for qualification, but I downweight the v_i-dependence concern relative to the reader's framing.","tokens_in":5501,"tokens_out":29001,"duration_ms":295145,"concrete_test":"Analytically recompute the volume of X in the modal coordinates q_i=(\\dot x_i+π_i x_i, \\dot x_i−π_i x_i) for v_1=v_2=1 and verify that the ε-dependent factor becomes ε^3(1−ε)^2/(1+ε)^2, whose maximizer is ε=(√13−2)/3 rather than √2−1. If this is confirmed, the abstract and conclusion must explicitly restrict the silver-ratio claim to the stated physical coordinates, or the result must be re-derived for a coordinate-invariant controllability measure.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The derivation of Eqs. (14)–(18) is internally consistent, but the central claim is not coordinate-invariant, and this is the load-bearing issue for the paper's advertised message. The volume in Eq. (15) is computed in the physical coordinates z=(x1, \\dot x1, x2, \\dot x2). If one instead uses the equally natural modal coordinates q_i=(\\dot x_i+π_i x_i, \\dot x_i−π_i x_i), the Gramian for each stable/unstable pair has determinant proportional to π_i^3, and the 4D volume becomes det W ∝ π_1^3 π_2^3 (π_1−π_2)^2/(π_1+π_2)^2; with π_2 fixed, the ε-dependent factor is ε^3(1−ε)^2/(1+ε)^2, whose maximizer is ε=(√13−2)/3 ≈ 0.535, not √2−1. Thus the silver-ratio optimum is an artifact of the chosen state scaling. The paper does state in Section I that the results depend on the choice of coordinates, but the abstract, the conclusion, and the inertia-ratio recommendation in Eq. (22) present the silver-ratio rule without that qualifier. The footnote-1 assumption that v_1 and v_2 do not depend on π_1, π_2 is less problematic here: in the canonical example (20)–(21) the coefficients are v_1=1/(m g_0) and v_2=−1/(m g_0), independent of the inertias, although the single-input condition 'IF1 and IF2 linearly dependent' is an additional unverified restriction for the example.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The note studies two unstable second-order systems coupled through a common input and measures controllability by the volume of the state-space reachable with unit energy while the trajectory starts and returns to the origin. In the normalized coordinates defined by Eq. (10), the volume is reduced to Eq. (15), whose maximizer over the time-constant ratio is the inverse silver ratio. The paper then recommends the silver-ratio inertia ratio for an inverted-pendulum example and concludes that controllability is maximized when the time-constant ratio equals the silver ratio.","tokens_in":5880,"tokens_out":12909,"duration_ms":130683,"significance":"The algebraic core of the paper is correct: the transformation (10), the dynamics (11), and the volume computation (14) are internally consistent, and the maximization of ϵ(1−ϵ)/(1+ϵ) is elementary and correct. The paper also has the virtue of being fully closed-form, with no fitted parameters or postdicted predictions. However, the volume criterion is not coordinate-invariant, and the paper itself concedes this in Section I. The abstract and conclusion nonetheless present the silver-ratio rule without that qualifier, and the design recommendation in Eq. (22) inherits the coordinate dependence. This is a load-bearing issue for the advertised message, not a mere presentation detail.","major_comments":[{"comment":"The optimal ratio ϵ*=√2−1 is not invariant under changes of state coordinates, despite being presented without qualification in the abstract and conclusion. The paper states in Section I that all results depend on the choice of coordinates, but Eq. (17) and the design rule Eq. (22) are phrased as intrinsic properties. The coordinate transformation (10) is not orthogonal, so the Lebesgue measure used for the volume of X changes under reparameterization. For example, in the modal coordinates q_i=(ẋ_i+π_i x_i, ẋ_i−π_i x_i), the volume of the reachable set expressed in those coordinates is proportional to π1^3π2^3(π1−π2)^2/(π1+π2)^2; with π2 fixed, the ε-dependent factor is ε^3(1−ε)^2/(1+ε)^2, whose maximizer satisfies 3ε^2+4ε−3=0, i.e., ε=(√13−2)/3≈0.535, not √2−1. Thus the silver-ratio result is a property of the chosen state realization (10), not of the physical system, and the claims in the abstract, Section VI, and Eq. (22) must be explicitly scoped to those coordinates or supported by a physical argument for why those coordinates are the correct ones for measuring controllability volume.","section":"Section IV, Eqs. (14)–(18); Section I"},{"comment":"The recommended inertia ratio I1/I2=δs² depends on two additional restrictions that the example does not establish. The first is the footnote-1 assumption that v1 and v2 do not depend on π1 and π2, which does hold for the particular example (20)–(21). The second is the statement immediately before Eq. (22) that IF1(t) and IF2(t) are linearly dependent, which reduces the two-axis dynamics to the single-input form (1)–(2). In a general inverted-pendulum system these two force components are not automatically linearly dependent, and the optimal inertia ratio would be different if they are not. The sentence presenting Eq. (22) should therefore state both conditions explicitly and should not be read as a recommendation for the unconstrained two-axis system.","section":"Section V, Eqs. (20)–(22)"}],"minor_comments":[{"comment":"The indices in the displayed inequality appear to be swapped: the term describing the ξ1 direction should use β1², and the term describing the ξ2 direction should use β2².","section":"Section III, Eq. (9)"},{"comment":"The phrase “More concrete examples of such cases cases include” contains a duplicated word “cases”.","section":"Section V"},{"comment":"The quantities π1 and π2 are eigenvalues with units of inverse time, so calling them “time constants” is imprecise; the time constants are 1/π1 and 1/π2. Please adjust the terminology for consistency with the dynamics in (1)–(2).","section":"Sections I–IV"},{"comment":"Equation (12) uses T before T:=diag(T1,T2) is defined; reorder the definitions so that T is introduced before it appears in the formula for X.","section":"Section IV, Eq. (12)"}],"recommendation":"major_revision","confidential_remarks":"The mathematical derivation is sound and the note is clearly written, but the central claim is coordinate-dependent and the abstract, conclusion, and design recommendation go beyond what the mathematics supports. I would be willing to see a revised version that either defends the canonical coordinates as physically meaningful or consistently qualifies the result as a property of the chosen state realization."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe one thing to know: this is a correct, short calculation, and the silver-ratio connection is new, but the headline rule is coordinate-dependent in a way that the abstract and conclusion do not disclose. In the coordinates used in the paper (x_i, xdot_i), the reachable set volume is proportional to (π1 π2 (π1−π2)/(π1+π2))^2, which maximizes at π1/π2 = √2−1. But the same calculation in modal coordinates q_i = (xdot_i + π_i x_i, xdot_i − π_i x_i) gives a determinant proportional to π1^3 π2^3 (π1−π2)^2/(π1+π2)^2, and the optimum ratio moves to (√13−2)/3 ≈ 0.535. So the silver ratio is an artifact of the chosen state scaling. Section I does admit coordinate dependence, but the abstract, conclusion, and the inertia recommendation in (22) present the rule without the qualifier. That mismatch is the main problem.\n\nWhat the paper does well: the derivation is transparent, the calculus is right, and the connection to the golden-ratio/Fibonacci literature is real and not overclaimed. There are no fitted parameters and no postdiction. The v_i-independence footnote is actually satisfied in the pendulum example—v_i = ±1/(m g0) does not depend on I_i. The stable/unstable decomposition in (12) is imported from [2] without proof; it is standard, so this is a presentation gap rather than a correctness gap. More important, the example only maps to the single-input form (1)-(2) if the two components of the applied force are linearly dependent; that condition is asserted but not verified for the systems named in the introduction, and it is not a generic property of a single force in 3D.\n\nWho is this for? A reader interested in numerical coincidences in control might enjoy it, and it is a useful object lesson in how controllability metrics change under coordinate transformations. It deserves a serious referee because the mathematics is honest and correct, but the revision should qualify every claim of 'maximized at the silver ratio' with 'in the canonical physical coordinates used here' and either justify those coordinates as the right ones for the design objective or reduce the strength of the design rule. I would not cite the silver-ratio rule as a design principle, but I might cite the paper as a clean example of coordinate dependence.\n\nRecommendation: send to peer review with requests for revision, not desk reject.","headline":"A correct short calculation whose headline claim is coordinate-dependent: the silver-ratio optimum in the paper's coordinates becomes (√13−2)/3 in modal coordinates, so the rule needs a strong qualifier.","tokens_in":6357,"tokens_out":5966,"would_cite":false,"duration_ms":60062,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["93B05","93C15"],"pacs":[],"model":"deepseek-v4-flash","headline":"For two coupled pendulums, the silver ratio gives the widest reachable state space.","keywords":["silver ratio","controllability","inverted pendulum","reachable set","unit energy","time constants","reaction wheel pendulum","linear systems"],"falsifier":"For a reaction-wheel or dual-pendulum setup with inertia ratio $(1+\\sqrt{2})^2$, measure the unit-energy reachable volume (or the minimal energy needed to reach a fixed state) and compare with nearby ratios; if the maximum is not at the silver-ratio value, the coordinate or input-coefficient assumptions, not the algebra, would be at fault.","tokens_in":5304,"feed_emoji":"⚙️","tokens_out":7163,"duration_ms":70349,"temperature":0.7,"pith_summary":"This note asks which ratio of time constants makes two unstable second-order systems, coupled through one common input, easiest to control. Controllability is measured as the volume of states that can be reached and returned to the origin with unit control energy, a definition that extends ordinary reachability to unstable dynamics. The author derives that volume in closed form in the chosen coordinates and finds that it is maximized when the time constants are in the silver ratio, the larger being $1+\\sqrt{2}$ times the smaller. That gives a concrete design target for inverted-pendulum systems of this type, such as reaction-wheel pendula. The result is conditional on the input coefficients being independent of the time constants and on the coordinate choice.","feed_headline":"Silver ratio gives maximum reachable volume for two pendulums","feed_subtitle":"The unit-energy reachable set peaks when the time constants are in ratio $1+\\sqrt{2}$.","key_machinery":"The load-bearing construction is the set $X$ in (3), the unit-energy reachable set for an unstable system; the author computes its volume by the coordinate change $T_i$ in (10), which decouples each unstable second-order equation into a stable and an unstable first-order mode, reducing the problem to two ellipsoidal constraints described by the matrix $P$ in (13). The determinant of the resulting quadratic form yields the scalar function $\\epsilon(1-\\epsilon)/(1+\\epsilon)$, and the maximizer of that concave function is $1/\\delta_s=\\sqrt{2}-1$. The definition of $X$ itself, following the balanced-realization treatment of unstable systems, is what makes the volume finite and computable.","core_discovery":"The paper derives, in closed form, the volume of the set of states from which two unstable second-order systems (1)-(2) can be driven to the origin and from which the origin can be reached with a single unit-energy input. Using the transformation (10) that splits each second-order mode into stable and unstable first-order modes, the author shows the volume is proportional to $\\left(\\frac{\\pi_1\\pi_2(\\pi_1-\\pi_2)}{4(\\pi_1+\\pi_2)}\\right)^2$ for $0<\\pi_1\\le \\pi_2$, i.e. $\\left(\\frac{\\pi_2^2}{4}\\cdot\\frac{\\epsilon(1-\\epsilon)}{1+\\epsilon}\\right)^2$ with $\\epsilon=\\pi_1/\\pi_2$. The factor $\\epsilon(1-\\epsilon)/(1+\\epsilon)$ is concave on $(0,1]$, so its unique maximizer is $\\epsilon^*=\\sqrt{2}-1$. The claim is therefore that the silver ratio $\\delta_s=1+\\sqrt{2}$ maximizes this measure of controllability; in the example rigid body, this corresponds to an inertia ratio $(1+\\sqrt{2})^2$.","pith_inferences":["If the input coefficients inherit time-constant dependence (for example $v_i$ proportional to $1/\\pi_i^2$), the objective gains new $\\pi$-dependence and the optimal ratio will shift off $1+\\sqrt{2}$; mapping those cases would delimit when the silver-ratio rule is valid.","The same determinant calculation for $N$ coupled unstable modes would produce a rational function of $N-1$ ratios; identifying its maximizer is a natural algebraic extension that may yield other metallic ratios.","Because the volume measure is coordinate-dependent, the practical design rule should be re-derived in the actuator and sensor coordinates of a specific robot; the silver ratio is a canonical-coordinate rule, not a physical invariant."],"forward_implications":["For two unstable second-order modes driven by a common scalar input, unit-energy controllability is maximized when the time constants are in the silver ratio $\\pi_2/\\pi_1 = 1+\\sqrt{2}$.","In the balanced rigid-body example, the corresponding optimal inertia ratio is $I_1/I_2 = (1+\\sqrt{2})^2$.","The optimal ratio is independent of the absolute time-constant scale and of the input gains $v_1, v_2$ as long as those gains are constant.","The reachable volume scales as $v_1^2 v_2^2 \\pi_2^4$, so stronger constant input coupling and a larger overall time constant enlarge the reachable set without changing the optimal ratio."],"supporting_citations":[{"why":"Defines the reachable-set notion for unstable systems as states reachable from and returning to the origin with unit energy.","marker":"[2]"},{"why":"Supplies the standard ellipsoidal reachable-set formula for the auxiliary first-order systems used in the derivation.","marker":"[10]"},{"why":"Provides the dual-pendulum dynamics that fit the two coupled unstable second-order systems.","marker":"[1]"},{"why":"Supplies the reaction-wheel inverted pendulum whose roll and pitch dynamics motivate the system class.","marker":"[9]"}],"fun_headline_variants":["Silver ratio maximizes controllability of coupled pendulums","Unit-energy reachable volume peaks at silver ratio","Two unstable systems: silver ratio gives best control","Silver ratio unlocks maximum reachable states"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The silver-ratio optimum assumes the input coefficients $v_1$ and $v_2$ stay fixed while the time constants $\\pi_1$ and $\\pi_2$ are varied, and it is expressed in the coordinate system chosen in the paper.","fun_headline_variants_meta":{"raw":{"variants":["Silver ratio maximizes controllability of coupled pendulums","Unit-energy reachable volume peaks at silver ratio","Two unstable systems: silver ratio gives best control","Silver ratio unlocks maximum reachable states"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000336,"raw_usage":{"total_tokens":1811,"prompt_tokens":848,"completion_tokens":963,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":464,"completion_tokens_details":{"reasoning_tokens":917}},"tokens_in":464,"tokens_out":963,"duration_ms":7596,"temperature":1.0,"reasoning_tokens":917,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T12:27:15.246768+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For a reaction-wheel or dual-pendulum setup with inertia ratio $(1+\\sqrt{2})^2$, measure the unit-energy reachable volume (or the minimal energy needed to reach a fixed state) and compare with nearby ratios; if the maximum is not at the silver-ratio value, the coordinate or input-coefficient assumptions, not the algebra, would be at fault.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Defines the reachable-set notion for unstable systems as states reachable from and returning to the origin with unit energy."},{"cited_title":"Muehlebach and R","cited_arxiv_id":null,"evidence_quote":"Supplies the standard ellipsoidal reachable-set formula for the auxiliary first-order systems used in the derivation."},{"cited_title":"3 I< HJ1䘸4eV XUm 7 Ni B rR҄SkkߏR9C Rk 5 p 2!q^SZH#^Ã,>8N[Pcʈo","cited_arxiv_id":null,"evidence_quote":"Provides the dual-pendulum dynamics that fit the two coupled unstable second-order systems."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the reaction-wheel inverted pendulum whose roll and pitch dynamics motivate the system class."}],"review_version":1}