What these tiers are: “Strong”, “Moderate” and “Context” are this project’s research classification of how closely a public statement bears on the documented filing-restriction indicators. They are not judicial findings, not an adjudication of anything said in a video, and not a determination that Frank Tufano is a vexatious litigant — no court in this dataset has made that determination. Court findings, party statements and research classifications are kept in separate evidence layers throughout this site; see Methodology.
How a tier is assigned. Strong — the statement names a tracked case or dispute and describes filing conduct of the kind the indicators document: repeat filing on the same dispute, filing after an adverse decision, or filing framed as imposing cost on the opponent. Moderate — it bears on litigation conduct or chronology in a tracked dispute, but the case link is tentative or the statement describes something other than the speaker’s own filing behaviour. Context — general commentary with no specific tracked case identified.
What the factor labels are. Each statement is also labelled with the filing-restriction factor its own words describe, so the reason for a classification is visible rather than implied. Most factors correspond to a criterion recorded in the indicator data, and the three that do not are marked as such there — a filing that has only been threatened has no docket to record, and an allegation about a court is the speaker’s characterisation rather than a recorded criterion. Repeat filing on the same dispute, filing after an adverse decision and cost imposition as stated motive are the three that can qualify a statement as Strong; threatened or prospective filing, self-reported filing volume, filing announced or confirmed, adverse outcome described and judicial or forum bias alleged record related but weaker signals. A statement that describes no filing conduct carries no factor badge. A factor records what the speaker said — not that a court found it, and not that anyone has been determined a vexatious litigant.
How quotations are marked: An ellipsis (…) inside a quotation marks words left out between two passages of the same video, which can sit some minutes apart. The omitted wording is not reproduced here, and a clip on the statement supercut plays the whole span from the quotation’s first word to its last — so it includes the omitted material. Quotations without an ellipsis are continuous speech.
Scope and evidentiary treatment: Only videos researched as bearing on this litigation appear here; material classified as not litigation-related is excluded from the site entirely. Quotations are transcribed from each video’s own YouTube auto-generated captions and timestamped from the caption cues covering the passage quoted, and every quote links back to that exact moment. Auto-captions are machine-generated and mis-transcribe names and figures, so wording is preserved as transcribed rather than silently corrected. Full transcripts are held in an internal research archive and are not republished here. Nothing in this evidence layer is a court finding.
What changed here, and what it costs. Until October 2026 this page streamed every clip from the speaker’s own upload, and the link on each card opened the full, unedited original at the same moment so any clip could be checked in its own context. That channel has been removed and those uploads are gone. What plays here now is this project’s own compilation, published September 29, 2026: the same 74 clips in the same order, cut at the same boundaries, but re-encoded by this project with each quote and its coding burned into the frame. It is a derivative, not the original, and the original can no longer be produced for comparison — not by this project and not by anyone else. Each frame still names the video id and the offset the clip was cut from, so what you are watching remains traceable to a specific moment in a specific upload; that upload simply cannot now be opened. The verbatim caption transcript each quotation was read from is held, dated and unaltered, in this project’s research archive.
How these clips are cut: Every clip is word-aligned: matched against the video’s own per-word caption timings, so the clip is bounded by the first and last words of the quotation itself rather than by the caption cue around them, with about half a second of lead-in and a second of tail so neither word is clipped. The alignment was read from the caption tracks as they stood on September 28, 2026. The boundary is a request made to YouTube’s player, which seeks to the nearest keyframe, so playback can open or close a moment either side of it. The timestamp shown on each card remains the researched one — a citation should point at what a researcher read, not at a refinement of it. The clips play oldest video first. That ordering is a presentation choice made by this research: it is not evidence that the statements were made in relation to one another, and a statement appearing before another is not treated as a cause of it. Sequence is not treated as proof of motive or causation unless a cited source says so.