The characterisation of inorganic nanoparticles (NPs) by single particle inductively coupled plasma mass spectroscopy is possible only if the spectrometer is capable of measurement with high time-signal resolution. The latest generation of spectrometers allow for measurements with dwell times (dt) shorter than the 100 μs gold standard, i.e. as low as 10 μs. The statistical behaviours of signals obtained with dt values of 10, 20, 50, and 100 μs were tested for 40, 60, and 100 nm silver NPs. Very low measured signals (units of counts) led to the occurrence of zero signal values inside the peaks corresponding to individual NPs. The probability of the occurrence of a zero signal inside the peak increased with decreasing dt and decreasing NP size. The standard approach to the bordering of the beginning and end of the peak by one zero signal point failed here and lead to the false detection of a larger number of smaller peaks. For example, in the case of 40 nm NPs a quadruple number of peaks were detected for a dt value of 10 μs compared to the 100 μs dt value; the mean peak width at 10 μs dt was approximately 220 μs, while at 100 μs dt it was 550 μs. The results tended to be less distorted when dt was longer and the NP size was larger. Low dt values also led to a distortion of the peak area distribution. For 40 nm NPs and 10 μs, the most frequent peak area and the width of the peak area distribution were not evaluated due to a non-Gaussian course; 20 μs dt caused (compared to 100 μs) a decrease in the most frequent peak area by approximately 35% (33 counts for 100 μs dt vs. 22 counts for 20 μs dt) and an increase in the width of the peak area distribution by 70% (10 counts for 100 μs dt vs. 17 counts for 20 μs dt). Therefore, new approaches to bordering peaks were tested, which consisted of searching for an uninterrupted zero signal point sequence with a total length of 50 μs or 100 μs. Only the criterion of a 100 μs delay between the two adjacent peaks resulted in values of the number of detected peaks, the most frequent peak areas, and the width of peak area distribution virtually independent of dt.