I should also mention the issue of mesh latency which is one of the reasons why central scene commands were created in the first place.
Mesh topologies, including zwave and Zigbee, do not guarantee forced sequencing.
Messages can and do bounce around the network, and they can take different paths to the hub at different times.
It’s entirely possible that if you open and close a door very quickly, the “sensor closed” message might be received by the hub before the “sensor opened” Message. It happens.
The same thing can happen with multiple clicks on a single switch. There isn’t really any way to guarantee that the first message will be received by the hub before the second one. And you can’t get super precise timing, since the messages may take different paths on different attempts.
All of this is why the official SmartThings documentation tells you that you can’t schedule something closer than one minute apart and be certain that it will occur as you expect.
Core does allow you to design a piston with shorter increments, there’s just no guarantee that it will work that way every time.
So a person might tap the switch in exactly the same way four different times but the Precision timings received by a smartapp in the cloud might be different for each of those four attempts.
Again, you can try it and see what happens, just be aware that the performance overtime may vary.