How Digg’s “Trending” and “Most Dugg” Feeds Actually Work

We live in a time when algorithms quietly shape what we see, think, and feel, often not for our benefit but to keep us scrolling. Feeds have become engines of outrage and addiction, rewarding the loudest voices over the most meaningful ones.

At Digg, we believe in a different approach. We’re returning to an idea that feels almost radical now: an honest feed. One built to inform and delight, not to exploit your attention. The goal isn’t to make you stay longer, it’s to make your time feel well spent.

Two of the most visible feeds on Digg, Trending and Most Dugg, work in very different ways. One measures statistical confidence. The other measures raw popularity. Together, they create a balance between what’s catching fire and what’s already burning bright.

Most Dugg: Simple Popularity

The Most Dugg feed runs on a simple principle. Count the votes and show what people liked most in the past 24 hours.

Here’s what happens behind the scenes:

1. Look at every post from the last 24 hours.

2. Count how many diggs each one received.

3. Sort them from highest to lowest.

That’s it. No time decay curve. No hidden adjustments. Just a clear snapshot of what the community responded to most that day.

This feed is easy to understand and easy to trust. If you can see the upvote counts, you can instantly tell why a post sits where it does.

Trending: What’s Probably Good, Not Just Loud

Trending is a little more thoughtful. It’s designed to surface stories that look genuinely promising, not just the ones that got lucky with an early burst of attention. To do that, we use a statistical method called the Wilson Lower Bound, often shortened to WLB.

If we ranked stories by a simple ratio of upvotes to total votes, a post with one upvote and zero downvotes would look perfect. A post with 1,000 upvotes and 200 downvotes would look worse if you only looked at the raw ratio. But that simple math can be misleading.

The Wilson Lower Bound fixes that by asking a better question:

Given what we know so far, what’s the most conservative estimate of how many people truly like this post?

It produces a cautious score that accounts for both the ratio of upvotes and the total number of votes. A post with ten votes will always be treated less confidently than one with ten thousand, even if both have the same raw ratio. That keeps the Trending feed fair, stable, and harder to game.

How the Wilson Lower Bound Works

The Wilson method was introduced in 1927 by statistician Edwin B. Wilson to estimate proportions with confidence. It’s used in everything from product ratings to comment sorting systems. The goal is to find out how likely it is that an observed approval rate reflects the true approval rate.

In plain terms, the formula calculates a range of plausible values for a post’s upvote rate based on the data available. We take the lower end of that range—the “lower bound”—as the score. It represents the most conservative estimate of support that still fits the evidence.

The fewer votes a post has, the more uncertain the estimate becomes, and that uncertainty lowers the score. This keeps small sample sizes from overpowering larger ones.

Here’s the mathematical form of the lower bound we use:

score =
(p + z² / (2n) - z * √((p(1 - p) / n) + (z² / (4n²)))) / (1 + z² / n)

where
n = upvotes + downvotes
p = upvotes / n
z = 1.96 (for 95 percent confidence)

This formula adjusts each post’s upvote ratio by its sample size and confidence interval. It gives well-supported posts higher credibility and limits how far small or noisy samples can climb.

Why Wilson Works for Digg

Fairness
It rewards consistency and confidence. Posts that earn steady approval are trusted more than ones that spike briefly and fade.

Resilience
It resists manipulation. A quick burst of votes or a perfect early ratio doesn’t instantly push a post to the top.

Transparency
The formula is public and well documented. Anyone can test it, replicate it, or even run it on their own data.

The Wilson Lower Bound doesn’t try to predict what will go viral. It just makes sure that what’s trending is both popular and statistically reliable.

A Quick Example

Imagine three posts:
1. Post A has 1 upvote and 0 downvotes.
2. Post B has 10 upvotes and 0 downvotes.
3. Post C has 100 upvotes and 10 downvotes.

If we just looked at raw ratios, A and B both look perfect at 100 percent, while C sits at 91 percent. But the Wilson method looks deeper.

It asks: given these sample sizes, which post is most likely to hold up as more people vote?

Because C has far more data, its lower bound is higher and more trustworthy.

So Trending would rank C first, then B, then A.

This keeps the feed aligned with reality instead of rewarding random luck.

How We Apply It at Digg

The process is straightforward:

1. Collect the total upvotes and downvotes for each post.

2. Compute the Wilson Lower Bound score for each one.

3. Sort posts by that score, highest first.

4. Exclude anything that violates community rules or moderation filters.

That’s the entire algorithm. Every post is treated the same way. There are no secret weights, editorial boosts, or human overrides once it enters the scoring system.

Our engineers maintain the implementation in an internal file, but the method itself is open and reproducible by anyone. It’s been part of statistical literature for nearly a century.

How the Two Feeds Work Together

Trending and Most Dugg play complementary roles.

Most Dugg reflects what’s already earned wide approval.

Trending highlights what’s emerging and trustworthy.

Trending rewards early engagement that looks real and consistent.

Most Dugg rewards overall engagement over time.

Together, they let Digg showcase both established hits and promising newcomers without bias or guesswork.

The Takeaway

• Trending uses the Wilson Lower Bound to rank stories by statistically confident approval.

• Most Dugg uses a straightforward count of upvotes in the past 24 hours.

• Both are explainable, transparent, and consistent.

We’ll keep updating this explanation as our systems evolve. The goal isn’t just to show what’s trending, but to make sure the way it trends stays clear and fair.

When you see a story rise on Digg, you shouldn’t have to guess why. The math should speak for itself.