Electroweak loop corrections and oblique parameters

Gauge-field self-energies, vertex diagrams, box diagrams and counterterms can all contribute to a measured process. Their separate gauge or scheme dependence cancels only in an appropriately complete prediction. Infrared singularities from virtual photons also require the relevant real-emission contribution in an inclusive observable. An ultraviolet cutoff inserted into one selected diagram is therefore not a physical measurement of heavy particles or a complete estimate of the uncertainty.

Heavy states need not always decouple as an inverse mass squared. A familiar Standard Model example is the leading custodial-breaking contribution from a strongly split top–bottom weak doublet, which grows with top mass squared when the bottom mass is neglected. It reflects the relation between mass and Yukawa coupling, rather than a rule that every heavy particle yields such growth. A new-physics oblique fit subtracts a reference Standard Model prediction that already includes its chosen top and Higgs contributions.

In a commonly used convention, new-physics T is the difference between charged and neutral transverse self-energies at zero momentum, each normalized by the corresponding mass squared and divided by the electromagnetic fine-structure constant. We use self-energies with energy-squared dimensions and the sign convention of the cited electroweak review. S and U describe additional derivative combinations of vacuum polarizations. Their definitions require a fixed reference prediction and a consistent normalization of gauge couplings.

The traditional three-parameter description assumes corrections are small and predominantly universal through gauge-boson propagators, with an analytic momentum expansion appropriate to sufficiently heavy new physics. Significant vertex or box effects, light thresholds, or additional momentum dependence require a larger description. Setting U to zero is an extra model assumption, often useful in a heavy effective expansion, rather than a universal theorem about new interactions. A fitted ellipse has meaning only together with these choices.

A positive T in this convention raises the leading predicted W mass when S and U vanish and cosine squared exceeds sine squared. The equation gives the fractional shift of mass squared. Converting it to a mass shift requires dividing the small fractional response by two, a frequent source of factor-of-two mistakes. When the corrections are not small, a consistent nonlinear or model-specific prediction should replace this linear response.

Neither a nonzero best-fit T nor a single mass residual identifies a unique particle. Different spectra can generate similar self-energy combinations, while correlations with S and the chosen inputs can rotate the allowed region. Combining results from papers that use different reference masses or floating-input conventions requires translation. The practical value of the parameters is that they expose a common response pattern, allowing several observables to test it together rather than treating every discrepancy as independent evidence.