Researchers at Tomsk Polytechnic University (TPU) have developed an algorithm that calculates the Newtonian gravitational constant (G) with 2.5 times greater precision than the current method used by CODATA, the international committee that standardizes physical constants. This fundamental constant underpins calculations in space navigation, celestial body mass estimation, and tests of gravity theories, yet it remains one of physics’ most stubbornly elusive numbers despite centuries of measurement efforts.
CODATA updates recommended values of fundamental constants roughly every two years by averaging results from multiple laboratories using a weighted mean. This approach, however, is sensitive to outliers-if some experiments report significantly different values of G, the final average can be skewed. TPU’s team suggests an alternative statistical method called the preferential median, which better handles measurement scatter and hidden errors without overreacting to anomalies.
Sergey Muravyov, professor at TPU’s School of Information Technology and Robotics Engineering, clarified that their method doesn’t replace established calculation techniques but serves as a complementary tool, especially suited to scenarios requiring robustness against anomalous data. Crucially, the new approach doesn’t revise physics itself; it simply refines the statistical processing of existing measurements.
Compared to constants like the speed of light, whose value is fixed exactly, the gravitational constant remains one of the least precisely known fundamental constants. Different labs worldwide often produce results for G that diverge more than expected, a problem dating back to Henry Cavendish’s famous 18th-century experiment. Even modern torsion balances, atomic interferometers, and vacuum setups have yet to finalize a universally accepted value.
The practical importance of improving G’s precision is far from academic. Beyond theoretical physics, more reliable values are essential in celestial mechanics, geophysics, and orbit and mass modeling of large objects. With the global space sector booming-the Space Foundation estimated the worldwide space economy surpassed $570 billion in 2023-the demand for more accurate models to predict spacecraft trajectories and orbital motion is only growing.
Next, TPU’s algorithm will need to be tested against a broader range of experimental data and gain adoption beyond its institution. The real test will come as new measurements of G are incorporated into international datasets, allowing for comparisons of TPU’s preferential median method alongside CODATA’s traditional weighted averages to gauge its stability and reliability.