Auto-ranking based on quiz performance?

[Anonymous]

I must be old-school or something. When Anki first came out, they had a "wrong" button, which, when pressed, would slide a card to the back of a deck so it would be asked again. I like this because, in one session, you couldn't end a session unless you finally got all cards correct. Now, Anki v1.6's "wrong" button is simply for statistical purposes.

I've been trying to do this with Pleco's ranks, but it seems that you need to manually adjust ranks based on what cards you feel need to be quizzed at certain freqencies, and then you need to define what ranks need to be quizzed when you start a session.

Can Pleco be set up to do these menial tasks automatically? Maybe I'm missing something. For example, a fresh deck is:

word rank correct
----- ------ --------
apple 1 0/0
dog 1 0/0

After the first 5 sessions, if I never can remember "dog", the history would be:

word rank correct
----- ------ --------
apple 1 5/5
dog 1 0/5

But, why can't the ranks automatically change to (assuming rank 5 means higher quiz frequency):

word rank correct
----- ------ --------
apple 1 5/5
dog 5 0/5

So, then the next few sessions might not even ask "apple", but definitely will ask "dog"?

Can this be done?

 

[Anonymous]

Sorry, I'll write the tables in Code so they display better:

[b]Pre-session:[/b]
word   rank  correct 
-----  ----  -------
apple  1     0/0 
dog    1     0/0 

[b]After 5 sessions (current model):[/b]
word   rank  correct 
-----  ----  -------
apple  1     5/5 
dog    1     0/5 

[b]After 5 sessions (desired model):[/b]
word   rank  correct 
-----  ----  -------
apple  1     5/5 
dog    5     0/5

 

[Anonymous]

Please disregard this. Although I can't set up ranking to work exactly as Anki 1 did, I've adapted it to benefit my study.

 

[mikelove]

Glad to hear it. But certainly this is an area that needs improvement. I think we may end up needing to adopt two different "rank" systems - one that's heavily customizable and one that (like SuperMemo) basically just does everything automatically and doesn't bother you with the details. The current system tries to be both of those, but the rank system is a little too tightly bound to the idea of SuperMemo-like frequency adjustment and doesn't allow you to do seemingly obvious things like cycling through cards until you've gotten all of them correct.