An arXiv preprint reports an updated estimate of the magnitude of the CKM matrix element Vcb from semileptonic B to D and D* decays, placing it at 39.18 plus or minus 0.47 times 10 to the minus 3. Vcb is the CKM matrix element governing the b-to-c transition. The same analysis predicts R(D) at 0.3069 plus or minus 0.0080 and R(D*) at 0.2548 plus or minus 0.0043. These ratios compare tau and muon decay rates. But the result leaves the central theory-experiment tension unresolved. The reported experimental averages remain above the theoretical predictions, particularly for R(D*).
Three streams of evidence
The fit draws on three classes of published input: LCSR theory predictions, lattice-QCD form-factor results, and binned differential decay-rate measurements from Belle and Belle II. The result is built by making theoretical calculations and measured decay distributions constrain one joint description of the B to D and B to D* modes. The analysis compares these inputs in nested stages, so the reported numbers can be read alongside the constraints used to produce them.
To describe how the decay shapes vary, the authors use the model-independent BGL dispersive parameterization, a flexible way to represent the form factors that govern those shapes without fixing them to a single model. They add strong unitarity bounds, which limit the allowed coefficients across four heavy-to-heavy transition channels, with HQET power corrections included. The fit uses frequentist maximum-likelihood minimization, a method that chooses parameter values that best match the inputs, and fully keeps the correlations in the SCET sum-rule inputs rather than treating linked uncertainties as separate ones.
What changes when the fit expands
The staged comparison starts with a lattice-only fit, then adds the correlated SCET/LCSR sum-rule predictions, and ends with a global fit that includes Belle and Belle II data. The first two theory-only scenarios do not determine Vcb. Once LCSR and experimental inputs are added, many high-order BGL coefficients carry smaller reported uncertainties. The exercise shows how the fitted form-factor description changes as the information set expands.
In the complete global fit, simultaneous B to D and B to D* constraints give Vcb at 39.18 plus or minus 0.47 times 10 to the minus 3. The plus-or-minus amount is the reported fit uncertainty. Because the calculation combines published theory inputs with binned measurements, this is a joint extraction rather than a number from one decay channel alone.
The ratios at the center of the test
R(D) is the branching ratio for the tau mode divided by the corresponding muon mode in B to D decays. R(D*) is defined the same way for B to D* decays. In the combined lattice-QCD and LCSR fit, the predicted values are 0.3069 plus or minus 0.0080 for R(D) and 0.2548 plus or minus 0.0043 for R(D*).
The central values shift when the fit scenario changes. In the displayed order of lattice-only, lattice-plus-correlated-SCET-sum-rule, and final global fits, R(D) is 0.3018 plus or minus 0.0144, 0.3069 plus or minus 0.0080, and 0.2983 plus or minus 0.0034. R(D*) is 0.2628 plus or minus 0.0082, 0.2548 plus or minus 0.0043, and 0.2538 plus or minus 0.0011. The reported uncertainties narrow across those three scenarios, showing how the added constraints influence the fitted ratios.
A gap that remains
The comparison with experimental averages is less settled. For R(D), the experimental average is above the arithmetic average of diverse Standard Model predictions, a pattern the paper describes as persistent theory-experiment tension. For R(D*), the reported experimental average is 0.281 plus or minus 0.011, compared with a theoretical average of 0.254 plus or minus 0.005. The study's prediction, 0.2548 plus or minus 0.0043, is below the central region of that experimental average.
The study also gives momentum-transfer and angular differential distributions for both muon and tau modes. For the muon channel, the theoretical central values are broadly consistent with Belle II within quoted uncertainties, although mild local deviations appear in some angular regions. The broad agreement in these distributions does not remove the separate difference seen in the averaged ratios.
Keeping the correlations between fitted quantities, the theoretical contour in the R(D), R(D*) plane shows substantial deviation from the experimental average at the 68% confidence level. The paper nevertheless says the current local angular discrepancies do not establish statistically significant evidence of new physics. The analysis therefore adds a sharper, correlated comparison, but it does not identify the cause of the mismatch.
A result that stops short of a verdict
The work is a preprint, identified in its front matter as arXiv:2608.24332v2 and dated 29 August 2026.
Paper data and sources
Original title: Exclusive Determination of $|V_{cb}|$ from Semileptonic Decays $B\to D^{(*)}\ell ν_{\ell}$
Authors: Xue-Wen Chen, Bo-Yan Cui, Jia-Wei Zhang et al.
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-25
DOI: Not available
Original paper · Full text