---
title: "Fraction comparison trick revisited"
slug: fraction-comparison-trick-revisited
url: https://listedarticles.com/articles/fraction-comparison-trick-revisited
canonical_url: https://www.johndcook.com/blog/2026/11/03/fraction-comparison-trick-revisited/
content_type: blog_post
language: en
published_at: 2026-11-03T14:41:25.000Z
updated_at: 2026-11-04T14:09:18.071Z
author: "John D. Cook"
authored_by: human
publisher: "John D. Cook Consulting"
topics: ["mathematics"]
license: all-rights-reserved
word_count: 371
reading_minutes: 2
citation: "John D. Cook, John D. Cook Consulting. \"Fraction comparison trick revisited.\" 3 Nov 2026. https://www.johndcook.com/blog/2026/11/03/fraction-comparison-trick-revisited/ (all-rights-reserved)"
# The full text follows. The web page shows an extract and sends readers
# to the source above; quote the citation and link the canonical URL.
---

# Fraction comparison trick revisited

> John D. Cook revisits the result that a/b is usually greater than c/d when a + d exceeds b + c (with probability 11/12), restates it for uniform random variables on [0, 1], and extends it to the probability as a function of how much larger a + d is.

# Fraction comparison trick revisited

[A few years ago](https://www.johndcook.com/blog/2021/02/25/fraction-comparison-trick/) I wrote about a curious result: *a*/*b* is usually greater than *c*/*d* when *a*+ *d* is greater than *b*+ *c*. And “usually” means with probability 11/12.

As stated in the post, you can make this statement rigorous by choosing integers up to size *N* at random and taking the limit as *N* goes to infinity.

We can also choose *a*, *b*, *c*, and *d* uniformly from [0, 1]. This is similar to choosing four integers up to *N* and dividing all the integers by *N*, which doesn’t change the probabilities.

Now *a*/*b* > *c*/*d* if and only if *ad* > *bc*, so another way to state the theorem is that

Prob(*ad* > *bc* | *a* + *d* > *b* + *c*) = 11/12.

Or to state it yet another way, pick two pairs of numbers at random. The probability that the pair with the larger sum also has the larger product is 11/12. That seems more intuitive to me than the original statement.

If *a* + *d* is greater than *b* + *c*, it’s likely that *a*/*b* is greater than *c*/*d*. It seems plausible that if *a* + *d* is *much* greater than *b* + *c*, it’s *more* likely that *a*/*b* is greater than *c*/*d*. Or in terms of cross products, the more the sum of one pair is bigger than the sum of the other, the more likely the product of the former is to be greater than the latter. That is

*f*(*r*) = Prob(*ad* > *bc* | (*a* + *d*)/(*b* + *c*) ≥ *r*)

should be an increasing function of *r*. We know the probability above equals 11/12 when *r* = 1. We expect that the probability should approach 1 as *r* goes to infinity.

This is indeed the case, as shown in the plot below.

![](https://www.johndcook.com/f_of_r.png)

In particular, *f*(2) ≈ 0.99 and *f*(6) ≈ 0.999. So if *a* + *d* is two (or six) times bigger than *b +* c, there is a 0.99 (or 0.999) probability that *ad* is larger than *bc*.

You can calculate *f*(*r*) analytically or estimate it by simulation. I had Claude do both and they produce comparable results. There is a closed-form expression for *f*(*r*) but it’s really complicated.
