REP v3 · New scoring · Live since 26 September

How the leaderboard changed under the new scoring

The new scoring is live. Every September post has been scored under it, and the board and this month’s rewards now use it. Below is how the top 25 changed, with the reason for each.

Why

Many of you said the board was filling up with AI-written posts. The old scoring rewarded clean writing and good structure. A tool can produce both in seconds from a Somnia announcement, so those posts scored as well as real work.

How it works

The judge reads what Somnia’s official accounts posted before your post, then scores your post on what it adds beyond that. A reply is judged against the post it answers too.

It is not AI detection. Nobody is accused of anything. It stops giving credit for the part a tool can do for you.

What scores well

  • Your own use: a trade, a screenshot, your own numbers.
  • A claim, prediction or comparison you stand behind.
  • Specific criticism, a bug, a question that asks for proof.
  • A faithful translation earns a middle score.

Rewording or summarising an announcement scores low, however polished. An image adds to your score only when your words add something too.

4,964
September posts scored under the new rules
3,124 (63%)
rated lower
940 (19%)
rated the same
900 (18%)
rated higher

Most totals went down because the new scoring is stricter overall, by about 17% across the board. So the minimum to qualify moved from 10 to 8 REP to match, and the same people qualify as before. The pool is shared, so what matters is your place, not the size of your total. How far a total moved depends on which posts were rated lower. "Why" is the judge’s own words on that Guardian’s biggest drop.

Before #Now #MoveGuardian BeforeNowChangeRated lowerRated higherWhy (biggest drop)
11·@thomasthomnb13751065-23%682 / 85581Agrees with parent's point and adds one specific metric (weekly cohort retention) but mostly restates the parent's argument in compressed form.
22·@bhupixhot482428-11%247 / 35456Adds no new information, analysis, or first-hand experience.
33·@jundohud461368-20%246 / 37457Generic enthusiasm about "gamifying agent performance" with no specific reference to what integration or product is actually being discussed.
44·@djaikoku428356-17%171 / 31866Generic validation of parent's detailed analysis.
56▼ 1@lesliemen418329-21%241 / 37759Translation of official announcements into Vietnamese.
65▲ 1@censarboyzzz389347-11%123 / 17014No added analysis, product use, or original contribution beyond repackaging the announcement with slightly different wording.
77·@arunikaliem342265-23%148 / 22931Well-structured explainer, but adds no first-hand use, testing, or critical evaluation beyond reframing existing material.
88·@chau09_dr311255-18%176 / 25448Restates the parent's core point (USDso as a circulating financial layer rather than isolated stablecoin) in shorter form.
99·@vothaithinh102311254-18%119 / 18946Adds no new data, product understanding, or specific insight beyond the official post's own framing.
1010·@plzgivemehopium303242-20%128 / 17329The author restates the core concept (reserve-generated value flowing back into DeFi), repeats the framing about ecosystem strengthening and sustainable capital…
1111·@vasyl_khlibko244216-11%57 / 9019Structured summary of Lipsky's article and official announcements (vesting extension, Somnia Markets positioning, dreamDEX/Prophecy Social numbers).
1212·@mr_samuel01191163-14%81 / 19465A recurring series post (Day 372) that restates official features: dreamDEX for trading, Prophecy for prediction markets, Somnia Agents for AI apps.
1313·@0x_empire159144-9%50 / 9624Generic reframe about the rewards program's mechanics.
1418▼ 4@denis71036993148101-31%71 / 888Restates official announcements about Somnia's milestones (Mainnet launch, Agents/Reactivity, dreamDEX volume, Prophecy users).
1515·@mogiwarad129106-18%47 / 10228Repackaged with polished rhetoric but adds no original observation, data, experience or criticism.
1614▲ 2@carlxbt122115-6%75 / 20163Agreement with parent's point about authenticity over REP farming, wrapped in generic reframe about AI use stages.
1716▲ 1@eleremen7115102-11%37 / 8221Reorganizes and expands the explanation (how yield reaches makers, why fees are zero, the flywheel) but adds no new information beyond what was already announced.
1821▼ 3@gbongpink10874-31%36 / 5713Vietnamese summary of Sept 3 Somnia AMA already extensively covered in official posts.
1917▲ 2@shinyviq104102-2%44 / 9825The question is direct and targets a concrete detail, but adds no new information itself, it simply asks for clarification of something already alleged.
2019▲ 1@trong_hatachi9985-14%19 / 396The addition of ecosystem metrics (dreamDEX volume, Prophecy users, product names) is thin, all already in official posts.
2122▼ 1@meowrypto9471-24%54 / 725Generic agreement wrapped in familiar framing ("solid test").
2220▲ 2@justhappysergio8784-4%35 / 7215The closing observation that "this month will show us how well it works" is generic speculation, not a specific prediction or grounded claim.
2323·@vilvati7162-13%44 / 6011Author rephrases the core concept with minor rewording but adds no new evidence, use case, explanation, or insight.
2424·@dioooya4435-20%12 / 162Agrees with parent's point ('Build where attention already is') and restates it in generic terms ('ecosystem needs to prove' / 'infrastructure can handle').
2526▼ 1@vithamongomerry3932-18%14 / 233Adds only generic reasoning ("prevents supply surges", "ensures commitment") already implicit in official announcements.
Cycle 2026-09 · top 25 before the change · scores are cycle totals, before is the board as of the switchAs of 26 Sep 2026, 07:07 UTC

This is a community-built tool created by an independent member of the Somnia community. It is not affiliated with, endorsed by, sponsored by, or officially associated with Somnia or its affiliated entities. All outputs are generated independently and are provided “as is” without warranties.

← Full rules