Skip to content

· Social Media · Feed ranking

Algorithm Whisperer: free game on how feed ranking works

Watch time, shares, saves. Tune your post and watch the reach meter climb.

What is Algorithm Whisperer?

Feedscore is a simple, made-up feed ranking model: it shows each post to a test group, earns more waves with early engagement, and grows each wave with watch time, shares, saves and comments. Pick 3 edits for each of 8 posts, watch the signals move, publish and watch the reach meter climb, or stall.

Algorithm Whisperer is a free game about how social feeds rank posts. It uses Feedscore, a simple model made up for the game: every post is shown to a test group first, reactions in the first hour earn more waves of testing, and each wave grows with a weighted mix of watch time, shares, saves and comments.

Each round gives you a draft post on a made-up app, like a 40-second recipe on ClipTok or a budget template carousel on Snapgram, and a list of edits. Pick three, watch the signal bars move, publish, and watch the reach meter climb wave by wave. Engagement bait and bought likes look tempting; Feedscore punishes both, and every result explains why.

At a glance

Category
Social Media
Difficulty
Medium
Play time
3–5 min
Price
Free, in your browser, sign-up optional

How to play

Read the app
Each made-up app weighs the signals differently: short videos lean on watch time, carousels on saves, text posts on comments.
Pick 3 edits
Cut the intro, add a checklist, post when followers are online… The signal bars move as you choose.
Publish
The post goes to a test group, then wave after wave. Bait counts 40% less; fake engagement stops it after the test group.
Score
Up to 100 for how close you climb to the best possible reach, +50 for the best mix. Under halfway costs one of 3 lives.

Scoring

Each post scores up to 100 points for how close its reach climbs to the best possible reach (measured on the reach meter), plus 50 for finding the best mix of edits.

Climbing less than halfway costs one of 3 lives. There are 8 posts per run; your best score is kept in this browser.

What it trains

Feed ranking signals
How watch time, shares, saves, comments and the first hour work together.
Format thinking
Why short videos, carousels and text posts lean on different signals.
Avoiding shortcuts
Why engagement bait and bought likes backfire.

Strategy and tips

Check what the app weighs
Short videos lean on watch time, carousels on saves, text posts on comments. Boost what counts most.
Don't forget the first hour
Without early engagement, even a great post only reaches the test group. Posting when followers are online matters.
Cut, don't stretch
Removing a slow intro raises watch time; adding filler lowers it.
Skip bait and fakes
"Comment YES" and "tag 10 friends" count less, and bought likes stop the post cold.
Look at the best mix
When you miss, the result shows the best three edits. Compare them with yours.

Why play Algorithm Whisperer?

Real feeds are complex and secret, but a few general principles show up again and again: posts are tested on small groups first, early reactions matter, and signals of real interest (watching to the end, saving, sharing, replying) count for more than empty likes. Algorithm Whisperer turns those principles into a small, honest, made-up model you can play with, so you learn to make posts people genuinely want rather than chase tricks.

Frequently asked questions

Is Algorithm Whisperer free?

Yes. It is free and runs in your browser, with no download. Signing up is optional.

Is Feedscore how a real app ranks posts?

No. Feedscore is a simple model invented for the game to show general principles. Real apps don't publish their ranking systems, and they change often.

Why do bought likes stop my post?

In Feedscore, fake engagement ends the post after the test group, to show why buying likes or joining like-for-like groups backfires: the people liking it don't care, so nothing real follows.

Can I play with a keyboard or a screen reader?

Yes. Every edit is a toggle button, each choice announces the new signal values, and every wave of the reach meter is read out.