Contents

How to Do Randomized Subset Review in SuperMemo

Contents

Photo by Brett Jordan from Pexels

Abstract
From the Knowledge Tree: Right click on a branch or folder –> View –> Branch to bring up the browser window–> Browser Context Menu > Child > Filter to bring up Element filter –> Next repetition: Choose today’s date for Maximum–> Click OK –> Randomize the elements or sort by priority –> Review

./images/KnowledgeTreeSubsetReview.png

./images/ElementBrowser.png

./images/ElementFilterWindow.png

Normally I don’t do subset review, and neither should you. There are at least two problems with subset review: limited interleaving and sequential order of review.

First, doing subset review already diminishes the interleaving among your whole collection. Originally your “outstanding queue” spans across your entire collection but now it’s limited to a particular folder or branch.

Second, the default setting for subset review is sequential. This is likely to happen in a sequential subset review:

1
2
Q: Group [...] has antigen A on blood cells.
A: A

Then when you click Next:

1
2
Q:Group A has antigen [...] on blood cells.
A: A

The previous item is an unnecessary cue for the subsequent item. It makes recalling the subsequent answer immensely easier. But it’s not desirable difficulty; this “cheating” doesn’t help you learn. This SuperMemo session is easier to “get by”, but it hurts long-term retention; possibly more forgetting, hence, paradoxically, needing more time and repetitions.

If you want to do a subset review for a particular branch without such sequential order, follow the steps outlined above. After filtering, you can sort according to priority or press that dice icon to randomize it.

In an ideal world this is not necessary: you would let SuperMemo’s algorithm handles all of your impending elements (items/articles). However in certain cases and real-life circumstances, doing randomized subset review may still be beneficial.