After pondering this MO question > https://mathoverflow.net/questions/193848 <, I wonder whether a Brownian motion can be fast (i.e. beats the law of the iterated logarithm) at its extrema?  Is it necessarily fast?