@thomas-heiss (My previous message was sent when incomplete)
Thanks for the thoughtful reply. Perhaps I can make the problem clearer with real data that I collected this morning.
Over the interval of the first 10 cards that I answered:
a. Starting counter:111
b. Answered 10 review questions with unique words 100 percent correctly.
c. Ending counter: 109
d. The interval duration was about 4 minutes.
f. It seems like the counter should have decremented by more than 2 cards.
Over the entire session duration:
a. Starting counter: 111
b. Cards reviewed 83 composed of following
c. 33 repeat cards, 31 answered correctly, 2 wrong (94% correct repeats). The 2 missed words were requizzed before the end of the session and constituted 2 of the total repeated cards.
d. 50 new cards, 32 answered correctly, 18 wrong (64% correct new cards)
e. 11 of the 18 missed new cards were quizzed again before the end of the session, 1 or 2 times, and constituted 16 of the repeated words in the session. All of the repeats were answered correctly. This means I had an ending balance of 7 missed and unreviewed new cards.
f. It also means that excluding the 16 repeats generated by missed new cards, that of the 33 repeats done during the session 17 of them were present in the queue at the beginning of the session. Of these 17 repeats 2 of them were requizzes of repeat card that had been missed and 15 were repeats answered correctly the first time.
g. Ending counter: 123
g. The session duration was 35 minutes (it took some time to record the specific data.)
This data indicates that I finished the session with 7 missed, and unrepeated, new words, which would be additions to the practice queue, and that I cleared 15 old words from the practice queue. If the counter was working correctly it should have decremented by a net 8 words during the session. Instead, it incremented by 12 words. Thus the total counting error was 20 cards. This leads me to suspect that the counter is not working.