Measuring how much I really remember
From: {{{From}}}
Country: {{{Country}}}
Sent: {{{Sent}}}
Subject: {{{Subject}}}
Question:
Summary
Using retrievability sort in the browser, user estimated his retention in a collection that has been neglected for a longer while.
Problem
After a long hiatus, I have a large number of outstanding items that I am slowly but steadily catching up with. If I understand SuperMemo, however, this large number does not represent the quantity of items I have forgotten, but rather, the quantity of items whose probability of being forgotten exceeds my configured forgetting index--which I have set to 10%. Also if I understand correctly, this means that as many as 89.9% of those outstanding items (though likely far fewer) I am STILL able to recall.
Is there any way to determine how many items would be outstanding if my forgetting index were adjusted upward? In other words, I'd like to know the answer to the question, "Among all memorized items, including those that are outstanding, how many of them am I at least 80% (or 50% or whatever) likely to be able to recall?"
I tried temporarily changing my forgetting index hoping that would temporarily change the total of my outstanding items, but unsurprisingly, that did nothing.
Also, the simulator does not seem to help in this regard.
I considered making an estimate by tracking my performance on a random sampling of the outstanding material, but given the huge volume of outstanding material, my sample size would need to be far in excess of what I can review in a reasonable amount of time.
Answer:
You can track Analysis: Efficiency : Recall.
If your material is prioritized and you want to discount the bias: open the outstanding browser, randomize it and execute Learn in the browser (Ctrl+L). In statistics you will see your current forgetting index for the session. This will give you an approximation. Your work will not be wasted. It will clean up some of the backlog.
In browser statistics you can also see a couple of forgetting index statistics, incl. the estimate retention for the browser.
Changing forgetting index will only affect items with the default forgetting index setting and only at the time of their next repetition. No intervals will be changed.
Simulation is useful only to get a general idea about the process. It will not help you understand your memory of a subset.
Outstanding estimate
Q: Is there any way to determine how many items would be outstanding if my forgetting index were adjusted upward? A: if you want to increase retention with the forgetting, it will be hard to make an estimate. If you want to drop it, you could sort the browser by retrievability and get a rough idea.
OP Reply
The statistics panel indicates my retention to be 90.9%, but 80% of my items are past due, and of that 80%, most are past due by several years. So, unless my collection is filled with extremely easy items, I assume that figure relates only to the items in my collection that are not outstanding.
Sorting by retrievability sounds like exactly what I need, but it seems that if I sort by retrievability, there is no table column in which the retrievability values appear, so there is no way for me to identify the number that exceed a particular threshold. Am I correct?
Follow-up: Never mind. I got the retrievability stats from the history panel, and got the estimate I was looking for. 35% of my collection is more likely forgotten than remembered. Unfortunate, but about what I expected. There's probably some way to calculate a precise number, but this is good enough for me. Thanks!
comment
history panel above probably means repetition history window, while a simpler preview would be with the element data window