Playing this video in slow motion reveals many details.  It starts out showing several sequential changes in speed commands. At about 2 seconds we turn on the engine #3 Light and quickly turn it back off.  The speed command of step 7 of 128 is continually being sent so it rises to the top of the display right after the Light is turned off.

At about 13 seconds the DCC Monitor loses power as the Command Station switches into Programming Track Mode. At about 15 seconds power is back up but the DCC Monitor reports there was a timing error for a DCC bit ("Wacky bit time").  That error display stays up long enough for a human to read (the duration of the error screen is the Sketch #define 'ERROR_DISPLAY_TIME'. At about 20 seconds the command display returns... but note that (because of the power outage) it has forgotten all the previous commands. The only command it has seen is a large number (more than 99) of  Reset ('Rst') commands. The Command Station does this per the NMRA specification.

The Sketch version number is displayed at the bottom of the screen because it is receiving no Idle Packets.

At about 27 seconds the Command Station sends a "Service Mode" command to set Register 6 to select CV page 0. It then quickly starts guessing at the current value of CV3. It's first guess is 255. It then guesses 0, 1, and so on until 254. Nothing responds because there is no train on the track.  The Command Station then, per the NMRA specification, sends a bunch of Resets.

 Note that no Command Station commands are dropped at interrupt level. They are all correctly stored in the Sketch's 'history[]' linked list. However, the camera sometimes catches the OLED display as it refreshes and it may look like some commands were dropped. Simply advance the video a few frames and you will see the correct command order appear.