Background The incidence of residual paralysis, i.e., a mechanomyographic train-of-four (TOF) ratio (T4/T1) less than 0.9, remains frequent. Routine acceleromyography has been proposed to detect residual paralysis in clinical practice. Although acceleromyographic data are easy to obtain, they differ from mechanomyographic data, with which they are not interchangeable. The current study aimed to determine (1) the acceleromyographic TOF ratio that detects residual paralysis with a 95% probability, and (2) the impact of calibration and normalization on this predictive acceleromyographic value. Methods In 60 patients, recovery from neuromuscular block was assessed simultaneously with mechanomyography and acceleromyography. To obtain calibrated acceleromyographic TOF ratios in group A, the implemented calibration modus 2 was activated in the TOF-Watch S; to obtain uncalibrated acceleromyographic TOF ratios in group B, the current was manually set at 50 mA (n = 30 for each). In addition, data in group B were normalized (i.e., dividing the final TOF ratio by the baseline value). The agreement between mechanomyography and acceleromyography was assessed by calculating the intraclass correlation coefficient. Negative predictive values were calculated for detecting residual paralysis from acceleromyographic TOFs of 0.9, 0.95, and 1.0. Results Group A : For a mechanomyographic TOF of 0.9 or greater, the corresponding acceleromyographic TOF was 0.95 (range, 0.86-1.0), and the negative predictive values for acceleromyographic TOFs of 0.9, 0.95, and 1.0 were 37% (95% CI, 20-56%), 70% (95% CI, 51-85%), and 97% (95% CI, 83-100%), respectively. Group B: Without normalization, an acceleromyographic TOF of 0.97 (range, 0.68-1.18) corresponded to a mechanomyographic TOF of 0.9 or greater, with negative predictive values for acceleromyographic TOFs of 0.9, 0.95, and 1.0 being 40% (95% CI, 23-59%), 60% (95% CI, 41-77%), and 77% (95% CI, 58-90%), respectively. After normalization, an acceleromyographic TOF of 0.89 (range, 0.63-1.06) corresponded to a mechanomyographic TOF of 0.9 or greater, and the negative predictive values of acceleromyographic TOFs of 0.9, 0.95, and 1.0 were 89% (95% CI, 70-98%), 92% (95% CI, 75-99%), and 96% (95% CI, 80-100%), respectively. Conclusion To exclude residual paralysis reliably when using acceleromyography, TOF recovery to 1.0 is mandatory.