Self-Stake Third Reading — 51 Channels, 4.85M LBC, Same Anchors Still Parked

On 1 May we published Self-Stake Fraud: 29 Channels Caught Tipping Themselves. On 13 May we ran the same analyser two weeks later and reported the count had risen to 50. Today, eighteen days after that second reading, we've re-run it for the third time. The count has risen to 51. The trajectory is no longer exponential. The mechanism has saturated.

Three Readings, One Trajectory

The channel count went up by one between the second and third readings. The total LBC pool grew by 169,672 LBC over the same eighteen days — a 3.6% increase, almost entirely from new entries replacing dropouts rather than from the existing anchors adding to their positions. Odysee's implied 5% take rose from ~233,900 to ~242,345 LBC — an extra ~8,445 LBC in fees from the fake-tip rail over the same window.

The flattening is the story. The exponential-growth pattern of the first two readings (29 → 50 in 12 days) was always going to be unsustainable: there are only so many channels with the capital, the technical know-how, and the editorial motivation to run the playbook. We are now at the near-saturation point — the operational ceiling of users who actively want to self-stake. From here forward, the count will move with the churn rate: a new arrival here, a wallet drained there, the same dozen anchor channels parked indefinitely on the same videos.

The Anchors Have Not Moved

Four channels have been in the top tier for all three readings and now appear to be permanent fixtures of the self-stake architecture. Their positions on 31 May:

@Unknown has held a position above 1,000,000 LBC for sixty days now. The stake on its single anonymous "WEF Covid Acts declared criminal" video has not moved. @ChristinatheAstonishing has held a 99%-top-1-share position on its single "Where Have I Been??" video for six months at this point (per the original 1 May piece). Each of these is unactioned in our takedown-tests register with notices served in May and re-served in this round.

The Martin Luther Antisemitism Stake Is Still Live

@louismarschalko still holds 100,100 LBC staked on a single video titled "On the Jews and their lies - Martin Luther" — Martin Luther's 1543 antisemitic treatise, which we named in the 13 May follow-up. The 100,000-LBC stake was the cleanest documented case in that piece of self-staking applied to explicit antisemitic content; eighteen days later the stake is still active, the video is still in search results, and Odysee has continued to accrue the implied ~5,005 LBC platform fee on the wash transaction. Under France's Loi Gayssot (1990, codified as Article 24bis of the 1881 Press Law) this content constitutes contestation de crime contre l'humanité adjacent material; under Germany's §130 StGB (Volksverhetzung) it is direct incitement to hatred. Both jurisdictions extend hosting-provider liability on notice. Notice has been served three times.

Four New Entries This Round

The channel count went from 50 to 51 because four channels entered the self-stake set and (per our threshold rules) three dropped out. The new entries:

The CEO Self-Stake Closes the Honesty Argument

We named Jeremy Kauffman's personal handle @kauffj in The CEO Tips Himself earlier today. The LBRY co-founder holds 88,232 LBC across 29 flagged videos, with three near-identical round-number stakes (10,208 / 10,159 / 10,117 LBC) on apolitical vanity content. He doesn't qualify for the formal self-stake threshold in this scan because the individual top-1 share (~12% by item count) sits below our 70% threshold — the analyser is tuned to catch the egregious single-video-park pattern, not the multi-video portfolio approach the founder uses. Both mechanisms are the same architecture, applied at different scales. The 51 channels above are the single-video-pinned-promotion edge of the distribution. The CEO is at the other edge.

What Saturation Means

The wrong reading of this article would be: "the self-stake problem is contained at ~50 channels, the mechanism is stable, no further action is needed." The right reading: the mechanism is now baked-in. The anchor channels are unactioned for the third consecutive measurement. The platform fee accrues every month. New entries (this round: PrivacyGuides, amaterasusolar, Qwinten, thisworldworks) cycle in at roughly the rate that wallets get drained or operators move on. The total LBC pool is stable because new entries enter at the same ~100,000-LBC stake tier that dropouts depart from — the same paid-promotion template, executed by different operators, producing identical fingerprints.

Saturation also means that future readings will only show meaningful change if one of two things happens: (a) Odysee actions the anchor channels, in which case the count and total LBC both drop sharply; or (b) the underlying ranking algorithm is changed so that self-staking no longer purchases search placement, in which case the new-entry rate goes to zero and the existing positions get unwound over time as operators realise the return is gone. Neither has happened in two months. We will publish a fourth reading in mid-June.

What Should Happen

The minimum viable response from Odysee — the same as in our two prior pieces — would be to (a) label paid-promoted content visibly, (b) publish a quarterly transparency report on tip-flow concentration (the never-filed DSA Article 15 obligation covers exactly this), and (c) cap the self-stake share that influences search ranking so the same wallet can't buy its own placements. None of these requires legal or moderation expertise; all of them are product decisions. None have been taken in the sixty days since the original measurement.

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.