The number of large metropolitan areas is only relevant if it makes up a larger share of the general population. His argument is that if there are more large cities where the violent crime rate is higher, then, the nations violent crime rate should be higher.
A more relevant statistic would be the number of people living in large metropolitan areas vs total population.
This would mean that, using his cited sources, we have: (35,070,220 Americans in cities of 250,000 or more people) / (311,591,917 Americans). Giving us around 11.3% of our population living in dense cities.
Whereas, the UK and Wales have 64,641,000 + 3,064,000 = 67,705,000 total population, of which 25,250,215 are in urban areas. This gives them about 37.2% of their population living in dense cities.
So yeah, he was trying to make a sound argument there but made a serious mathematical error. Their crime rates are at best similar to ours, given that the selection of crimes that are considered violent is similar, which he never established.
This guy made some pretty serious statistical errors either accidentally or intentionally.