INTRODUCTION A bowler’s ability to manipulate line and length of a delivery is a key performance attribute in cricket. 3D-motion capture (3DMocap), the gold standard in quantifying ball trajectories1 lacks ecological validity2. Fulltrack AI app may provide a feasible alternative to quantify and provide immediate ball trajectory feedback. This study explored the validity of Fulltrack AI to measure ball landing position compared with 3DMocap. METHODS 836 deliveries under various conditions (pace, spin; bowled, thrown (SidearmTM); batter, no batter), were recorded using a Qualisys 8-camera 3DMocap system and iPad running Fulltrack AI (version 1.13.1). Line and length were extracted from Quintic Spline filtered 3DMocap data and tabulated with Fulltrack AI data. Statistical analyses were conducted in R Statistical Software following removal of outliers. Bland Altman’s, 95% limits of agreement (LOA)3 were calculated, with line and length interaction further explored using 95% confidence ellipse area (CEA) to assess practical difference between 3DMocap and Fulltrack AI relative to ball landing position. Validity between Fulltrack AI and 3DMocap, and the interaction of conditions, were assessed with generalised additive models (GAMs). RESULTS Whilst LOA (line= -0.15 to 0.10; length= -0.88 to 0.27) demonstrated good agreement, CEA ranged from 0.17m2 to 0.42m2 depending on ball landing location relative to the stumps (2-4m or >8m, respectively). GAMs ‘average model’ established no significant (p>0.05) difference between 3DMocap and Fulltrack AI, or the interaction effects of training conditions. CONCLUSION Although Fulltrack AI appears statistically and ecologically valid in identifying cricket ball landing position, practitioners should be cognisant that practically, error varies relative to where on the wicket a ball lands. Thus, hindering the accuracy of feedback if needing to precisely quantify ball landing position when using Fulltrack AI app.