Twitter Algorithm Explained for Beginners: How X Picks What You See

How does the Twitter algorithm work, explained for beginners?

The Twitter algorithm, now called the X algorithm, is the system that picks and orders posts in your For You feed. It gathers posts from accounts you follow and accounts you don't, guesses how likely you are to reply, like, share or skip each one, and shows the best guesses first.

Open the For You tab on a brand-new account and it can feel random. Sports, crypto, a cooking video. Like and reply to a few topics for a week, and the feed starts to look very different. Nothing magic happens. The system is learning from what you do.

That's the core idea behind the Twitter (X) algorithm. It's a prediction machine. Once you see it that way, most of the confusing advice online starts to make sense, and you can tell the solid tips from the myths.

What is the X algorithm in simple words?

The X algorithm is a set of computer programs that decides which posts appear in your For You feed and in what order. It does this fresh every time you open the app, using what it knows about you and about each post.

Think of it like a helpful shop assistant. You walk in, and they quickly pull a pile of items you might like, then put the best ones at the front. They learn from what you pick up and what you ignore. Over time, they get better at guessing.

Here's something most people don't know: you can read a lot of the real code. Twitter published its recommendation code in March 2023 in the twitter/the-algorithm repository. Then in January 2026, X published the code for its newer For You feed in the xai-org/x-algorithm repository. The newer code is the one that matters today, though X keeps a few files private.

What's the difference between For You and Following?

For You mixes posts from people you follow with posts from people you don't, picked by the algorithm. Following shows only accounts you follow. Since late November 2025, Elon Musk says, the Following feed is ranked by Grok by default too.

X's Help Center says For You picks posts using signals like popularity and how people in your network interact with them. The Following change came from a post: Elon Musk wrote that posts from people you follow "are now ranked by Grok" and that you can still see an unfiltered chronological view, as Social Media Today reported. X also said the change may matter less if you follow fewer than 200 accounts.

For beginners, this is the main thing to remember. For You is where new people discover you. Following is where your existing followers see you.

How does the algorithm choose posts? The four steps

The X algorithm chooses posts in four steps: gather, learn about, score, and filter. Every For You feed goes through all four, in that order, each time it loads.

1. Gather. It collects recent posts from accounts you follow. It also finds posts from accounts you don't follow that seem close to your interests.
2. Learn about each post. It looks up details like the text and who wrote it.
3. Score. A model called Phoenix guesses how likely you are to do things like reply, like, repost, share, or hit "not interested."
4. Filter. It removes posts you shouldn't see, like posts from accounts you blocked, posts with your muted words, posts you've already seen, and posts older than 48 hours.

If you want the technical version with file names and exact values, my guide on how the X algorithm works in 2026 walks through the code.

Which words do beginners need to know?

Beginners only need about six terms to follow most algorithm talk. Each one has a simple meaning, and none of them needs a tech background.

One more term worth knowing is "impressions," which is how many times a post was shown. My explainer on Twitter impressions vs engagement covers how that differs from likes and replies.

Which actions count most?

Replies, quotes and shares count much more than likes in the current code. Negative actions, like muting an author or tapping "not interested," count heavily against a post. That's the single most useful thing a beginner can learn from the code.

In the published 2026 settings, a reply's default weight is 5.0 and a like's is 0.66. Sharing a post's link has a weight of 20.0. Muting the author sits at -58.8. These numbers multiply the system's guesses, not your raw counts, and X can change them at any time.

What this means for you is simple. A post that makes people want to answer beats a post people quietly like. A post that makes people annoyed hurts you more than a post nobody notices.

Common beginner myths about the algorithm

Most beginner myths come from old screenshots or guesses that the published code doesn't support. A few are worth clearing up early so you don't waste time.

• "The algorithm hates small accounts." The 2026 README actually describes a new-author boost. Posts from authors with impressions below a threshold get lifted toward a target position. X doesn't publish the threshold.
• "Old posts keep getting pushed." The For You code filters out posts older than 48 hours, so most reach happens early.
• "Posting 20 times an hour floods everyone's feed." The code shrinks the score of each extra post from the same author in one feed.
• "There's one secret trick." There isn't. The score adds up many small predictions.

What should a beginner do first?

A beginner should pick one or two topics, post about them consistently, and reply thoughtfully to others in that space. That gives the system clear signals about who your posts are for.

Based on what I've seen, clarity beats volume early on. If your posts jump between ten topics, it's harder for the system to find your people. My page on improving content relevance on X goes deeper on this, and the Twitter growth strategy for beginners turns it into a weekly plan.

If you'd like to see these signals in your own feed, I built a free extension, XVision, that shows each account's follower count next to their name and marks posts that are reaching far beyond their author's followers. It's on the XVision home page.

The simple version to remember

The Twitter algorithm explained for beginners comes down to one line: X shows people the posts it predicts they'll engage with, and it learns from every tap. You don't need to beat it. You need to give it clear signals by posting useful things for a specific group of people. Do that for a few weeks, and the feed starts working with you instead of against you.

Frequently Asked Questions

Is the Twitter algorithm the same as the X algorithm?

Yes, it's the same system under a new name. When Twitter became X, the For You feed kept running on a recommendation algorithm. The ranking model has changed since the 2023 code release, and X described its January 2026 ranking model, Phoenix, as Grok-based.

Can I turn off the algorithm on X?

Partly. The For You feed is always algorithmic. For the Following feed, Elon Musk announced in late November 2025 that posts are ranked by Grok by default, and he said an unfiltered chronological view is still available. X's help page doesn't describe Following as chronological.

Does liking a post change my For You feed?

Yes, likes shape your feed. The 2026 README says your recent engagement history is the main input to the ranking model. Likes, replies, reposts and even how long you pause on posts all teach the system what to show you next time.

How do I tell X to show me less of something?

Use "not interested," mute, or block on posts and accounts you don't want. The 2026 code predicts these negative actions and gives them large negative weights, such as -52.272 for "not interested." Muted words also filter posts out of your feed before ranking.

Do I need to understand code to grow on X?

No, coding knowledge is not needed to grow on X. The useful lessons from the code are simple: replies and shares count more than likes, annoying people costs a lot, and most For You reach happens within 48 hours of posting.