It's worth noting that statistics give very mixed reviews of Boston's defensive perimeter players.
According to 82games' Opponent Counterpart Production, Bradley and Crowder do well. Crowder, however, does poorly when moved to the 4. It's worth noting that this statistic doesn't attempt to differentiate between the times when Bradley guards 1s and when he guards 2s; it more or less considers him a full-time 2 and lumps all of that data into one stat. Marcus Smart is roughly average at either PG or SG, according to this metric.
According to basketball-reference's box plus/minus stat, Smart rates fairly well but Crowder is average and Bradley is actually a bit below.
Huge caveat here: defensive statistical measurements are in their infancy and should be taken with a large grain of salt. Specifically, they currently struggle when attempting to separate a player's defensive impact from the influence of his teammates. For example, Gobert's presence is going to have a positive effect on Hayward's defensive metrics, while IT is going to make Bradley look worse. I chose to share BPM because it attempts to adjust for this effect, but I'm not sure how well it performs this task.
For what it's worth, both Hayward and Morris grade as roughly average according to BPM, while 82games loves Hayward's defense and thinks Morris is average.