Extended Data Fig. 5: Failure of the SCTIF to reproduce both the MR and Hall response.

Simulations of the MR and Hall responses within the SCTIF given different parameterizations of v_F(ϕ) and τ⁻¹(ϕ, T). In each simulation, the experimentally determined Fermi surface (k_F(ϕ)) has been used. Note that the SCTIF is slow to converge at low fields and so the simulations do not extend all the way to H = 0. Simulation 1, ADMR-derived parameterization of overdoped Tl2201 albeit with no anisotropy at T = 0 and an anisotropic term that increases strictly linearly with T. Simulation 2, A scenario incorporating the v_F anisotropy derived from tight-binding modelling of ARPES measurements³⁷. Simulation 3, A scenario in which the anisotropy ratio of τ⁻¹(ϕ) is strictly T-independent in order to generate an MR with a maximum slope that is also independent of temperature (reminiscent of quadrature scaling). Simulation 4, A scenario in which a similar τ⁻¹(ϕ) parameterization to that used to model Nd-LSCO⁵³ is applied to overdoped Tl2201. Simulation 5, Simulation for overdoped Bi2201 with an enhanced anisotropy in v_F and τ⁻¹(T = 0) consistent with ARPES⁵⁴. Column 1, The Fermi surface parameterizations k_F(ϕ) and v_F(ϕ). Column 2, τ⁻¹(ϕ, T). Column 3, dρ/d(μ₀H) versus H. Column 4, dρ/d(μ₀H) versus H/T. Column 5, R_H(H, T).