High-level chunking

High-level chunking allows to operate with up to 80 object in your working memory, using computer-like data structures. Since the size of working memory is correlated with IQ, this actually make you smarter.

When dealing with large amount of information you may choose several strategies to encode it. High level visualization is probably the most accurate, but also the less analytical and more time consuming method. Hyperlinking is very fast and analytical, but it works best with stylized markers. Eventually I failed to describe the method in the middle, which I call here high-level chunking.

Simple chunking

In low level chunking we detect similar markers and chunk them in groups of 3-5 markers per chunk.In this way we effectively increase our working memory from 7 objects to 20 objects (5 chunks of 4 objects). Moreover we can perform marker manipulation between the chunks and between markers within each chunk.

Intermediate chunking

In mid-level chunking we work with data structures. These include linked lists, trees and maps.

Linked lists

Linked lists are ordered chunks of markers connected by some sort of linking “animation” that allows us effortlessly and accurately transverse forward and backwards between them. A good example of a list are task lists, like what to buy in grocery store.


A great way to remember texts are trees of markers. We have markers per articles, sections, paragraphs and sometimes sentences and terms within sentence. The linking here is hierarchical from parent to kids and similar to zoom-in animation on Prezi. Often we find it useful to connect the children of the same node as linked lists to increase accuracy. It is also a very natural way to follow the text logic.


Maps are fully associative ways to remember pairs of markers. We can use them to connect object of seemingly unrelated trees. This is a great way to encode out-of-the-box thinking or interdisciplinary connections. References we see in any scientific article are examples of such maps.

High-level chunking

In high level chunking we want to operate with 80 objects in our working memory. To do that we construct several data structures based on various types of objects and connect these data structures by mental links to allow passage between them. This is generally very similar to hyperlinking, but the markers may be significantly more complex and the data structures are much more organized. The organization of 5 tree data structures with 16 objects each enables memorization of 80 objects and active manipulation on these objects in working memory. This is approximately the maximal working memory utilization verified by the science.

Personally I find high-level chunking complex to master and use it only when I need to deal with complex multidisciplinary projects. Typically when I manipulate on these chunks in my head, I back it up by a spreadsheet or presentation to ensure I do not miss anything.



Get 4 Free Sample Chapters of the Key To Study Book

Get access to advanced training, and a selection of free apps to train your reading speed and visual memory

You have Successfully Subscribed!

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.