Softening of articular cartilage, mainly attributable to deterioration of superficial collagen network and depletion of proteoglycans, is a sign of incipient osteoarthrosis. Early diagnosis of osteoarthrosis is essential to prevent the further destruction of the tissue. During the past decade, a few arthroscopic instruments have been introduced for the measurement of cartilage stiffness; these can be used to provide a sensitive measure of cartilage status. Ease of use, accuracy and reproducibility of the measurements as well as a low risk of damaging cartilage are the main qualities needed in any clinically applicable instrument. In this study, we have modified a commercially available arthroscopic indentation instrument to better fulfil these requirements when measuring cartilage stiffness in joints with thin cartilage. Our novel configuration was validated by experimental testing as well as by finite element (FE) modelling. Experimental and numerical tests indicated that it would be better to use a smaller reference plate and a lower pressing force (3 N) than those used in the original instrument (7-10 N). The reproducibility (CV = 5.0%) of the in situ indentation measurements was improved over that of the original instrument (CV = 7.6%), and the effect of material thickness on the indentation response was smaller than that obtained with the original instrument. The novel configuration showed a significant linear correlation between the indenter force and the reference dynamic modulus of cartilage in uncontined compression, especially in soft tissue (r = 0.893, p < 0.001, n = 16). FE analyses with a transversely isotropic poroelastic model indicated that the instrument was suitable for detecting the degeneration of superficial cartilage. In summary, the instrument presented in this study allows easy and reproducible measurement of cartilage stiffness, also in thin cartilage, and therefore represents a technical improvement for the early diagnosis of osteoarthrosis during arthroscopy.