# The probability cheatsheet changelog stops in 2015 and the tree moved in 2022

> A ten page introduction to probability, written in TeX, published as two PDFs, one in colour and one in black and white, and co-authored since its second version by the professor whose course inspired it. The source is in the repository and the build output is committed beside it. The licence is stated in the readme body and there is no licence file, and two planned topics have been on the list since 2015.

**wzchen/probability_cheatsheet** — GitHub describes it as A comprehensive 10-page probability cheatsheet that covers a semester's worth of introduction to probability.. The repository metadata lists TeX as its primary language. This article stays within the project description and details documented in the GitHub repository README.

- Repository: https://github.com/wzchen/probability_cheatsheet
- Website: http://www.wzchen.com/probability-cheatsheet
- Stars: 3,170 · Forks: 695
- Language: TeX
- License: not declared
- Published: 2026-08-13 · Updated: 2026-08-18 · Language: en
- Canonical page: https://hysenlabs.com/projects/wzchen-probability-cheatsheet

## Two PDFs, one TeX file, and no licence file

The repository is five entries: a readme, a figures directory, the TeX source, and two compiled PDFs. Two PDFs from one source is the interesting structural choice, and it is not redundancy. One is the colour version and one is the black and white version, which is what you want when the cheatsheet is printed, photocopied, or read on a screen where a coloured figure disappears. Because the source is committed next to the output, either PDF can be rebuilt, and the figures directory holds the images the TeX pulls in. The licence is the gap. It is stated in the readme as a Creative Commons attribution, non-commercial, share-alike licence with a link to the deed, and the metadata for the repository records nothing identifiable. There is no licence file at the root. So the terms are discoverable if you read the document and invisible to tooling, which is the same pattern that makes a package look unlicensed to a dependency scanner.

## The changelog stops in 2015 while the tree moved in 2022

Four changelog entries, all from 2014 and 2015, and then nothing. The first is version 1.0 in July 2014, adding the cheatsheet itself. Version 1.1 follows in February 2015 with a new page of content and a new section on problem hazards. A patch follows in March 2015 with a diagram, a matching problem, and a corrected formula. Version 2.0 lands in September 2015. After that the changelog has no entries at all, while the default branch received a push on 15 June 2022. That is seven years of work on the tree with nothing in the history describing it. Nothing here says the changelog was declared finished or moved elsewhere, and no release was ever tagged, so the file's own record of what changed stops two years after the first version and three months after the last one it mentions. For a document people cite, that is the number to keep in mind: the description of its contents is shorter than its age.

## Two planned topics have been planned for eleven years

There is a section listing planned additions, and it has two items on it: the Monty Hall problem and the St Petersburg paradox. Both are classic probability paradoxes and both are exactly what a one-page-per-topic cheatsheet would want, which is presumably why they were written down. Neither has been added, according to the file's own list, and the list is positioned above the changelog, so a reader meets the statement that these are pending before meeting the last thing that happened. Neither item appears anywhere in the change history either, so there is no record of an attempt. The most likely reason is the one the changelog itself reveals when it records what had to be cut for space, but the file does not say so. What it does say is that the two names have been on a to-do list in plain view for eleven years, which is a more honest artefact than a silently missing section.

## Version 2.0 is where it stopped being one person's document

The fourth changelog entry is the only one that credits another person by name, and it credits him for more than edits. It records thanks for polishing the cheatsheet, for adding colour and figures, and for officially using it in his own course, and then adds the sentence that changes what the document is: the cheatsheet is now a joint work. The parenthetical matters as much as the thanks, because the course named is the one that inspired the cheatsheet in the first place. So the sequence is that a teaching fellow wrote a reference for himself, a professor adopted and improved it for a class, and from the second version the two of them own it. That is worth holding in mind when you read the licence paragraph, because a non-commercial share-alike document co-authored with an academic and distributed as a course handout has a different provenance from a personal note-taking file, and it is being redistributed by the hundreds of students who took that class.

## A section merge renamed classic problems and broke the references

The third entry contains a rename rather than a rewrite, and renames are the part of a changelog that breaks people. A section called Classic Problems was merged into a section called Example Problems. Nothing about the content changed, so anyone who had linked to the old section name, or told a student to look up a problem in it, or built a revision around it, now has a dead reference and no way to tell from the document that the material still exists. The same entry reorganised the material on one named distribution, which has the same effect on anyone who had memorised where it lived. This is the cost of a document that is reflowed rather than versioned per section, and it is the strongest practical argument for the two planned additions staying unmerged: adding two paradoxes to a ten page budget means displacing something else, and the file records that the last time it made room it did so by cutting left margins and moving a section.

## Space was found by cutting margins

One line in the second changelog entry is the most telling sentence in the file, because it describes how this kind of document is actually maintained. Alongside adding a diagram and a matching problem, the entry says that removing some left margins allowed room for both. That is the whole revision process for a ten page reference: content is not added because there is space, it displaces something, and the cheapest displacement in a dense single-page layout is typographic rather than conceptual. The same entry fixed an error in one formula and cleaned up sections, and the third entry fixed another formula error, which tells you that formula errors are found and corrected by readers, one at a time, and each fix produces a version bump. That is why there are four versions in fourteen months: this is not a document with a release cycle, it is a document that gets patched when someone notices something wrong with it.

## The contact page is two handles and a personal site

The file closes with a contact section that offers three routes for one person and two for the other, and they are all social. The first author is reachable through a question and answer site, a social handle, and a personal website, which is also the repository's homepage. The second is reachable through a single social handle. Below them is a route that is arguably the most useful of all: submit a pull request, or open an issue. So a document that is a static PDF plus a TeX source has an issue tracker and a merge queue attached to it, and the file says so plainly for comments, suggestions and corrections. What is missing is any statement about how long an issue stays open, who reviews, or whether a correction that arrives years later is still accepted, which for a document whose last recorded revision was in 2015 is the question worth asking.

## Conclusion

Use this if you teach or revise introductory probability and want a compact reference you can read in ten minutes, and remember that it is a cheatsheet rather than a text: it gives you the formulas, the distribution properties and a set of worked problems, not the derivations. Two things to check before you redistribute it. The licence is a share-alike, non-commercial one stated only inside the readme, with no licence file in the repository, so a fork has to carry that licence forward explicitly even though the metadata does not record it. And the changelog stops in 2015 while the repository itself was worked on in 2022, which means the document has had seven years of changes that the history does not describe. For anything you intend to build on, read the document rather than the version history.

## FAQ

### What is the probability cheatsheet repository?

A ten page reference covering a semester of introductory probability, based on a Harvard introductory probability course and co-authored by a former teaching fellow for that course and the professor who teaches it. The repository holds the TeX source, a figures directory, and two compiled PDFs, one in colour and one in black and white.

### How do I get the probability cheatsheet PDF?

The repository links to a hosted PDF from the first author's personal site, and both the colour and the black-and-white versions are committed to the repository alongside the TeX source, so either can be rebuilt or read directly. The homepage is a personal site rather than a project page.

### What licence is the probability cheatsheet under?

A Creative Commons attribution, non-commercial, share-alike licence at version 4.0, stated in the readme with a link to the licence deed and a badge. There is no licence file in the repository and the license field comes back empty, so the terms are discoverable only by reading the document.

### What changed in version 2.0 of the probability cheatsheet?

It became a joint work. The entry credits the professor for polishing the document, adding colour and figures, and using it in his own course, and adds figures, a section on the probability of an intersection or union, a distributions section, and a distribution properties section that absorbed an inequalities reference. Repeated definitions were removed and sections were reordered.

## Sources

- [Official documentation](http://www.wzchen.com/probability-cheatsheet)
- [Official README](https://github.com/wzchen/probability_cheatsheet#readme)
- [Project repository](https://github.com/wzchen/probability_cheatsheet)

---

Hysen Labs editorial analysis, written from the project's own repository and release notes. Cite the canonical page: https://hysenlabs.com/projects/wzchen-probability-cheatsheet