How we count, and where the limits are.

A channel is on the panel while we follow it and it has named at least one ticker in the last 30 days. Each channel gets a tier, A to D, by its subscriber count. The tier sets the order in which channels appear in the brief, never how much a mention weighs. We do not read videos shorter than a minute, or content marked as made for kids.

In total, the archive spans 8,055 videos and 39,987 ticker mentions — well beyond the 30-day window shown above.

A mention is one ticker in one video, however many times it is named there. For each mention we note its direction (bull, neutral or bear), the kind of argument, how convinced the creator sounds, and how much of the video the ticker took up, which is the mention's weight. The classification is done automatically, from the video's transcript.

BubbleIndex is a number from 0 to 100 for how loud the panel is about a ticker right now. It is built from three things, heaviest first. We publish what goes in and in what order; the exact weighting is our own.

1. Channels — how many channels in the panel talk about the ticker.
2. Conviction weight — how much each of them says about it, and how firmly: each mention counted by how much of its video the ticker held and by the creator's conviction.
3. Hype share — how much of it is retail hype. A ticker whose hype over the last 30 days runs clearly above the panel's own baseline carries the ⚠ hype flag beside its number.

The number does not depend on the other tickers or on the window you pick, so it can be compared from day to day. It moves a little when the panel itself changes size.

Hype raises the number rather than lowering it, so a high value alone cannot tell sustained analysis from a burst of hype: read it with the ⚠ flag and the signal types on the ticker's own page. The number measures the conversation and says nothing about the price.

BubbleIndex v2 · YouTube only

The stage and the arrow read the change over the past week; BubbleIndex reads the level. A ticker can have a high number and be Cooling: the panel still talks about it a lot, but less than a week ago. A ticker nobody mentioned this week is Flat, whatever its number: silence is not a trend.

Every ticker is placed into one of five stages every morning, from how much its BubbleIndex has changed over the past week — not a subjective call. A ticker nobody mentioned this week is Flat.

Emerging

BubbleIndex has risen over the past week — early signs of a building conversation.

Bubbling

A clear rise over the past week — more channels, saying more.

Boiling

A clear rise on a ticker that is already loud, or where hype is driving it — the most intense stage.

Cooling

BubbleIndex has fallen over the past week while the panel still talks about it — the conversation is thinning.

Flat

About as loud as a week ago, or not mentioned this week.

A ticker is flagged ⚠ when its share of hype-classified mentions runs clearly above the panel's own rolling baseline, and never on the strength of a single video.

The Hype Watch rail ranks flagged tickers by hype share, highest first — and only tickers that cleared the flag appear on it at all.

A ticker clears the data floor when at least two channels have mentioned it, or at least three videos have. A ticker below it is kept, because that is usually where a name first appears, but it is marked "Low data" in every listing, ranked below the tickers that clear the floor, and left off the bubble map, where a dot built on one sentence would sit beside dots built on hundreds and look like the same kind of thing.

A ticker's own page goes further. It carries a caveat whenever it is built from at most two videos, whoever made them — including the cases that clear the floor on two channels. The floor decides what can be ranked against the rest of the panel; a page has to say what it is built from before it says anything else.

• Not a forecast, and not financial advice. The numbers describe the conversation, not the price.
• Not calibrated against prices. The weighting is a judgement, not a model fitted to outcomes.
• Only the panel, not all of YouTube. A ticker the panel does not discuss does not appear, however much it is discussed elsewhere.
• The classification is automatic, and its accuracy has not yet been measured against a hand-checked sample.
• The three-day comparisons in the movers and the brief follow the weekday: after a weekend, more tickers look like they are falling. The 7-day arrow beside BubbleIndex compares a day with the same weekday a week earlier, so it does not.
• The panel grows and shrinks, and the numbers are relative to it.