Shure's figure of 70 were for the outliers, the main bulk of their measurements was clustered around 10-20cm/sec. That stacks up with what my PPM is indicating on most of my LPs. The PPM I use is a DIY design, using the display of a Velleman VU indicator with the rectifier and time constants changed to match the Type 1 PPM spec. This has a faster rise time than the Type 2 PPM in general use in the UK, and is the more commonly used one in Europe. It is still a quasi-peak indicator rather than true-peak, so will underindicate on very short peaks, especially at high frequencies.
I've set the output level of my phono stages to give 0dBu (0.775v) out at 5cm/sec recorded velocity at 1kHz, and my PPM goes to full scale at +15dBu, and only a few LPs ever get to that level, most peak at around +10dBu, so something around 15cm/sec. It does go to show, however, that unlike digital that has a fixed 0dBFS maximum, LPs can be cut at whatever level the cutting engineer decided was a good compromise between playing time, noise and distortion. I have a few compilation LPs, Sounds of the '60s, that sort of thing, that cram 30 minutes playing time onto one side, and the level is quite low, with the corresponding noise quite high, but it doesn't matter much considering the compressed nature of the music.
As there's no standard for maximum cutting levels, and given Shure's measurements, an absolute minimum of 20dB overload would be indicated as necessary, with 25dB preferable.
S.