Glass-forming ability ͑GFA͒ as defined by a critical cooling rate R c to vitrify a liquid upon solidification is a complex function of many parameters. Some of the parameters, such as liquid-crystal interfacial energy, temperature-dependent liquid viscosity, and influence of heterogeneities, are crucial but their accurate experimental determination is challenging. Here, instead of relying on the experimental data, we draw random values for the difficult parameters and use the classical theory to examine probabilistic distributions of R c for two well-known metallic glasses. Direct random parameterization produces extremely broad distributions spanning tens of orders of magnitude. Dramatically sharpened distributions are obtained around experimental R c upon guiding the random parameterization with limited calorimetric data. The results suggest that it is plausible to determine GFA even in absence of data for crucial parameters. © 2010 American Institute of Physics. ͓doi:10.1063/1.3462315͔Glass, as representative of a diverse family of materials distinguished by an amorphous or disordered structure, is widely used in various industrial and engineering applications. Being in a metastable thermodynamic state, glass always has a tendency to transform to the more stable crystalline state, even during a formation process. To take advantage of the unique properties of glassy materials requires the ability of such materials to form a sufficiently large quantity of glass without crystallization. The intrinsic parameter that describes such glass-forming ability ͑GFA͒ is the critical cooling rate, R c , needed to prevent formation of detectable amount of crystals upon quenching the materials from liquid state. A reliable assessment of R c is mandatory for both practical and fundamental reasons.Since glass formation is equivalent to the suppression of crystallization, classical crystal nucleation and growth theory has been widely quoted as a theoretical description of glass formation. [1][2][3][4][5][6] According to the theory, GFA depends on many thermodynamic and kinetic parameters. Some of these parameters, including the liquid-crystal interfacial energy , temperature-dependent liquid viscosity ͑T͒, and influence of heterogeneities, are considered critical for glass formation but are challenging to determine experimentally. Therefore, the classical theory has rarely been used in its comprehensive form for the quantitative assessment of GFA, despite that a large number of simplified criteria ͑e.g., Refs. 7-10͒ featuring one to three parameters have been proposed based on qualitative or semiquantitative analyses of the theory.In this study we intend to explore the plausibility of using the comprehensive theory for GFA assessment without specific information on the difficult-to-measure parameters. We randomly draw values for the parameters from vast domains, and examine the probabilistic distributions of R c calculated with the classical theory for two bulk metallic glasses, namely, Pd 40 Cu 30 Ni 10 P 20 ͑PCNP͒ and Zr 41.2...