It’s well known that Tag Change Scripts defined on the tag itself can sometimes miss events and fail to execute reliably. Using a Gateway Tag Change Script has generally been the bullet-proof alternative.
Has this behavior improved in Ignition 8.3?
We often need to synchronize values from OPC Device A to OPC Device B through Ignition, which can result in a long list of Gateway Tag Change Scripts. Any improvements in reliability or alternative approaches would be appreciated.
They don't "miss events". The queue can overflow when there are too many tag changes too quickly on a single tag, and there is a flag that indicates this happened.
The behavior hasn't changed.
On a single tag? 
It’s on the tags that change only once we see this happening.
What is the name of this flag you mention? 
Sensor values etc, no worries - its ok the next second.
Pssssst!
See my RePublish Tag Actor in my Integration Toolkit.
If you cannot use that, study this topic and the linked topics: