The condition of fish relates to their energy reserves, and a suite of proxies exist to approximate condition, including biochemical and morphometric indices. Biochemical indices directly measure energy stores but are expensive and sometimes lethal. Morphometrics offer several advantages, but their utility as condition proxies is debated and largely untested experimentally. Here, we manipulated the condition of yellowfin bream (Acanthopagrus australis, Günther, 1859) via food reduction to simulate the effect of poor habitat quality and calculated a variety of condition indices through time. We measured four non-lethal morphometrics (Fulton's K, Le Cren's relative condition, width-to-length ratio and girth), the hepatosomatic index and a biochemical benchmark (liver lipid content). Girth and width-to-length ratio were reasonably well correlated with lipid content (R2 = 0.74 and 0.56, respectively). The hepatosomatic index was only weakly or uncorrelated to other indices, including lipid content (R2 = 0.35). Where precise estimates of body condition are not needed or repeat measures on the same individual are useful, non-lethal morphometrics provide a fast, cheap and non-lethal alternative to biochemical and lethal morphometric methods for this species. We finish by outlining how artificial intelligence-based automation can be combined with morphometrics to further enhance ethical monitoring by eliminating the need to capture and handle fish entirely.