Cotton productivity in the Indo-Gangetic loamy sands has remained stagnant, largely due to subsoil compaction, which restricts water infiltration and root elongation. This study quantified the effects of one-time sub-soiling, applied as single or cross passes at different spacings, on soil hydraulics, root and fruit development and seed cotton yield (SCY) using analytical approaches beyond conventional mean comparisons. Five tillage treatments, including conventional tillage (CT) and 4 subsoiling configurations, were evaluated across3 site-years on fixed plots in Punjab, India. Treatment responses were assessed using effect size estimates (Hedges g) for key traits, while a piecewise structural equation model traced the causal pathways linking tillage intensity (TI) to infiltration rate (IR), root length (RL), fruiting capacity and yield. A quadratic response surface regression was used to identify the joint optimum of the IR and RL. Cross sub-soiling at 1 & times; 1 m spacing increased infiltration by 10-13 mm h-1, enhanced rooting depth by 31-33 % and improved SCY by approximately 33 % relative to CT. Structural equation modelling indicated that 57 % of the total tillage effect on yield was mediated through improved soil hydraulics, with an additional 34 % being mediated through enhanced fruiting capacity. The response surface explained 80 % of the yield variability and identified an optimum near a 40 mm h-1 IR and 110 cm RL, with small deviations causing substantial yield losses. Overall, deep, closely spaced cross sub-soiling improved soil-root functioning and translated these gains into higher fruit retention and yield, providing a robust framework for optimising sub-soiling strategies in semi-arid cotton systems.