We use cookies to keep the site running, for analytics, and to remember your settings. You can choose which categories to allow. Cookie Policy
Guide
Back to homeThe rating here is about quality, not popularity: real work, community trust you actually earned, and a profile that gives both some context.
A like counter is trivial to game. We weigh several signals at once — here's the contrast.
How many people pressed a heart. A couple of accounts and ten minutes will inflate it.
Profile, project depth, community trust and activity — each with its own weight. One fake like won't move the formula.
The platform keeps two independent scores. They are easy to confuse, so here is the line between them.
How well a specialist presents themselves and their work.
This is a person's score. It grows from the quality of your work and from how the community answers it. Filled-in profile fields give that work context, but on their own they do not make a rating.
How well one specific project is presented.
This is a single piece of work's score: its description, media, stack, and the votes cast on it. Every project carries its own — and the strongest ones pull the profile rating up with them.
How well a specialist presents themselves and their work. It is not a reward for filling in fields: three quarters of the score come from the work itself and the response to it, and only a quarter from profile completeness.
Each column sums to 100. You can see exactly what shifts between boards: the 30-day one leans harder on freshness and trust, and lighter on completeness.
The average quality of all your projects plus your single best one.
→Polish 1-2 key projects to the max. Keep the rest at least decent.
A weighted scale across profile blocks: bio, avatar, skills, education, experience, contacts.
→Click the "Profile X%" pill in your space — the modal shows what's still missing. It is a one-off job: fill it in and this factor is spent.
How others react to your profile. A few likes from acquaintances won't tip it.
→Publish strong projects and articles. Multi-accounts don't work — see below.
Projects published and media inside them. The first 5-10 give the biggest lift.
→5-10 projects with real work is already a strong signal.
The date of your newest published work. Logins, page views, and likes are not counted here at all.
→On the main board this factor is switched off — a quiet month costs you nothing. It only wakes up on the 30-day board, and what lifts it is a new publication, not activity for its own sake.
How well one specific project is presented. A separate formula for each piece of work, also from 0 to 100 — and the strongest ones pull the profile rating up with them.
Each column sums to 100. You can see exactly what shifts between boards: the 30-day one leans harder on freshness and trust, and lighter on completeness.
Project votes adjusted for confidence — a couple of likes won't trigger a jump.
→Real work → real votes. Just not from fake accounts.
Field completeness: description, role, status, links, problem/solution/results, cover.
→Don't leave empty blocks. A "Results" field with metrics lifts the signal most.
Screenshots and videos attached. Growth flattens after a handful of items.
→3-6 screenshots or a short video — more persuasive than words.
How many technologies this piece of work actually involved. It belongs here: tags describe one project rather than a person — the profile rating does not count technologies at all.
→5-8 tags you really used. A longer list buys nothing — the saturation curve flattens it.
Measured from the publish date: the older the project, the lower the weight. Edits don't reset it.
→Aiming for the 30-day board? Publish new projects regularly.
Three techniques make the formula resistant to gaming and fair to different kinds of contributors. Under each chart is the same thing in plain terms.
3 likes with 0 dislikes is a weaker signal than 100 with 5. The formula scores our confidence in the ratio, not the ratio itself.
→A handful of votes from friends buys nothing — the formula is waiting for a real volume of reactions.
A raw ratio calls 3 out of 3 a perfect 100%. The trust score sees that three votes aren't enough and returns less than half. Two hundred votes with a few against land near the top.
Old votes and projects matter less every week. Consistent quality still holds the top.
→Nothing burns off on the main board. For the 30-day one, publishing something new every few weeks is enough.
Freshness weight by age. The half-life is 45 days on the all-time board and 20 days on the 30-day one.
10 projects beat 5. But 50 are only slightly better than 30. A flood of template work won't beat a handful of strong cases.
→Past roughly 10 projects, improving the ones you have pays better than adding another.
How project count feeds the score. At 8 projects you already collect half of this factor; past ~20 the curve is nearly flat.
Badges are not arbitrary: each has a clear criterion and is awarded automatically the moment you meet it. Together they add at most +5 points, so farming them isn't worth it.
All-time keeps the history, the last-30-days board gives newcomers a shot. A low all-time score doesn't lock you out of the monthly top.
The full history, freshness excluded. Established members hold the top until their signal is beaten.
Freshness weighs more and old votes fade faster. A few weeks of steady activity is enough.
Trying to bypass the formula doesn't get you the rating you wanted — it gets you moderator attention.
Publish 5-10 projects with media, a description, and a real role — that is the largest share of the score. Fill in the profile so that work has context. After that, simply be in the community: follow people, comment honestly. Badges show up on their own once you meet their criteria.
