Choice is good. 
different protocols approach these issues differently. So you may find one works better for one particular installation while the other works better for another.
The zigbee home automation standard requires that every device check in at least once a minute. If not, it will be marked as non-responsive and eventually dropped from the network.
In contrast, Zwave has no requirement that a device check in. And in zwave Classic, it is never marked as unresponsive unless you run the repair. (Zwave plus handles this housekeeping stuff somewhat differently)
From the beginning, zwave was designed for very low cost residential lighting systems with some (cheap) sensors. People could and did do some tricks with Layout to get the maximum efficiency in each zone. Allowing for automatic update of the address tables would interfere with those tricks.
It’s harder to do the same tricks in a zigbee install because of the check in requirement.
The end result of all of this can mean that zigbee May mark a device as unavailable which in fact is available, while zwave May mark a device as available when it is in fact unavailable.
It’s not that one method is always better than the other: they are different, and you need to choose the one that best matches your own needs.
Running the zwave repair utility Makes sure that all of the device statuses are up to date, so if batteries have died or a device has failed, the network tables will get updated.
A lot of people think of a Z wave repair just as a wellness check: and they do it every day to check the battery status on everything while avoiding excess polling during the busy times on the network.
So if you think of it that way, I think running it once a day will make more sense. It’s a wellness check, which makes sense when you have battery operated devices.