We checked intraoperative patellar tracking with both ‘towel clip technique’ and the ‘no thumb technique’ on 354 patients (571 knees) who underwent primary total knee arthroplasty to decide whether to do or not to do lateral retinacular release. All surgical procedures consisted of medial parapatellar arthrotomy and patellar resurfacing. Patellar tracking was assessed under pneumatic tourniquette with the no thumb technique first and reevaluated with the towel clip technique. The tracking was graded as total contact, good contact, lateral contact, and subluxation. The knees graded as total or good contact with the no thumb technique were classified into group A; those graded lateral contact or subluxation by the no thumb technique but total or good contact by the towel clip technique were classified into group B; and those graded lateral contact or subluxation by both techniques were classified into group C, in which lateral releases were performed. We classified 371, 148, and 52 knees into groups A, B, and C respectively. Patellar lateral tilting in the Merchant view was reviewed preoperatively and 2 weeks, 6 weeks, 6 months, and 1 year postoperatively. There were no statistical differences on postoperative patellar tilting among the groups. Assessment of the patellar tracking using only the no thumb technique may overestimate the need for lateral retinacular release. The use of the no thumb technique as a screening test, and reevaluation with the towel clip technique, may reduce unnecessary lateral retinacular release.