Self-Stake Fourth Reading — 53 Channels, And Three New Single-Video Accounts Below The Analyser Threshold

Our scanner cron stopped firing on 2 June 2026 — the production Docker image for the scanner container was wiped by an unrelated host cleanup job and the cron wrapper had no self-heal. The fault is documented in our methodology page; the result is the same as if no scan had happened. We restored the scanner today (16 June) and re-ran. This piece is the first post-restoration reading. It is also the fourth measurement of the self-stake count we've published since 1 May. The trajectory:

+2 channels and +222,576 LBC in the 16 days since the third reading. The platform's implied 5% cut on the fake-tip rail crossed a quarter-million LBC for the first time. The saturation argument we made in May holds: the count keeps creeping up by ones and twos rather than doubling, but the mechanism is fully baked in. Anchor channels (@Unknown, @ChristinatheAstonishing, @neohumaneve) remain parked. Third-reading detail, second, first.

But The Newest Entries Don't Show Up In The Count

Our analyser requires a channel to have at least three flagged videos before classifying it as likelySelfStake. The threshold was chosen to limit false positives — at 1 or 2 videos it's harder to distinguish a self-stake from a single-anonymous-benefactor donation. But that same threshold misses the cleanest possible fingerprint of self-staking: a brand-new account that posts exactly one video and stakes a round number on it. The 16 June scan surfaces three such channels with new claims since 2 June:

Three channels. One video each. Exactly 100,000 LBC on each. Total: 300,000 LBC of fake-tip activity (Odysee's 5% take: 15,000 LBC) that our official count does not include because the analyser threshold excludes single-video channels. We could lower the threshold to 1 video and capture these — but then we'd also capture all the single-video posts that genuinely received a single 100K LBC tip from a real benefactor. Probability that all three of these are real benefactors, independently choosing the exact same round number on channels with the exact same posting pattern within the same 16-day window: arithmetically possible, editorially indefensible.

What Does Show Up: @OzFlor At 200,000 LBC

One new entry does meet the threshold. @OzFlor previously had three flagged items in our dataset; the 16 June scan adds a fourth — a single video staked at 200,000 LBC titled "AI is Essential for World Government: Documenting the Race to Centralised Control". With the new stake, @OzFlor crosses the analyser threshold: 4 videos, 100% top-1 share on the new one (the prior three combined hold under 0.5% of the channel's LBC pool). Formal classification: likelySelfStake. The framing — AI as the instrument of one-world-government control — is a recurring vocabulary from the Sovereign Citizen movement and crosses into the same conspiracy-history catalogue as the three single-video entries above.

And @CosmicEvent Keeps Building The Same Position

@CosmicEvent — documented in our Trollcow Israel Search-Buy and third self-stake reading — posted another item this round with a fresh 100,000 LBC stake. The new video is titled "ARE THEY POISONING EVERYTHING?". The channel now sits at 20 flagged items with 2,100,080 LBC of accumulated tip activity. It remains below the formal likelySelfStake threshold by top-1-share (48% across the catalogue rather than the 70% cutoff) — but the round- number stake on every new upload is itself the signature. Twenty single-100K-LBC stakes on twenty videos averages out to 48%; the absolute structure is identical to the single-video case above.

What This Round Confirms

• The self-stake mechanism continues to attract new participants on roughly the same cadence — 2 formally counted in 16 days, 3-4 additional sub-threshold examples in the same window.
• The pattern is now visibly tilting toward single-video channels with round-number stakes. The May reading saw multi-video portfolios; the June reading sees one-video accounts. That's consistent with the explanation we floated in May: the founders / experienced operators have their positions; new entrants are testing the mechanism on minimum-viable accounts.
• Two of three new single-video entries are explicitly antisemitic, one is Spanish-language. The self-stake mechanism is being used disproportionately by content categories Odysee already declines to moderate at the title level.
• The platform's implied 5% take crossed ¼-million LBC for the first time in our measurement series. Including the sub-threshold cases, the take is closer to ⅓-million LBC.

What Should Happen

We have served Article 16 notices on all three single-video channels (@anonyme / @III / @uniónhispanoamericana) plus the two new @OzFlor and @CosmicEvent claims to Odysee's support address, with CC to Coimisiún na Meán, ARCOM, and BNetzA. Status checks at +30 / +60 / +90 days will land on the takedown-tests register.

The minimum viable response from Odysee remains what it has been for the last sixty days: label paid-promoted content visibly, publish a quarterly transparency report on tip-flow concentration (the never-filed DSA Article 15 obligation covers this), and cap the self-stake share that influences search ranking. We will publish a fifth reading in mid-July unless the scanner cron silently breaks again — in which case you will see the gap on the public commit log of the data file itself.

Live data — current self-stake count, total laundered LBC, Odysee's implied 5% take, the per-channel concentration scores — refreshes on every deploy from /data/stake_concentration.json and is reflected on the Stats page.