Comment Counter Trader: scoring a YouTube crowd

Comment Counter Trader is a tool built on one idea: what do the people who comment under our videos expect Bitcoin to do, and were they right? It reads the comments on the FOMO Factory YouTube channel, sorts each one as bullish, bearish or neutral, and keeps the result.

From that it builds two things. The first is a gauge that shows the overall mood of the comment section right now. The second is a track record: every bullish or bearish comment becomes a call, and 60 days later the call is checked against what Bitcoin actually did.

The name comes from contrarian investing, the habit of asking whether the crowd tends to be wrong at the extremes. The tool does not answer that for you. It collects the evidence so you can look at it yourself. Our glossary has a short entry on contrarian investing if the term is new.

The gauge is a score from 0 to 100. A low score means the comments are mostly bearish, so the crowd is fearful. A high score means they are mostly bullish, so the crowd is greedy. Around 50 the comment section is split. The dial groups the score into three zones: fear below 40, neutral from 40 up to 60, and greed from 60 up.

A signal box next to the dial uses five bands on the same score. The wording in the box (such as BUY or SELL) is a label for the mood band. It describes how the crowd feels, in the contrarian spirit of the tool. It is not an instruction to trade.

A confidence figure sits beside the score. It goes up when there are more classified comments and when the score is further from 50, so a score of 52 from 40 comments carries less weight than a score of 15 from 400.

Every few minutes the tool fetches new comments from the channel's videos through the YouTube Data API. An AI model reads each comment and labels it bullish, bearish or neutral. If the model is unavailable, a simple keyword check does the labelling instead. Comments written by the channel owner are left out, and so are comments that look like spam.

The score is then built from the bullish and bearish counts only. The share of bullish comments is stretched away from the midpoint, so a clear majority shows up as a clearly high or low score instead of a timid 55 or 45. The tool looks at the last 7 days first. If that window holds fewer than 30 bullish and bearish comments together, it widens to 30 days, then to 90 days, then to all history, and the dial shows which window it used.

A chart under the dial draws the score history against the Bitcoin price, with a Raw and a Smoothed view and ranges from 7 days up to all history. Smoothing averages the days around each point, which calms the line without changing the data underneath.

Coordinated campaigns can distort a comment section, for example five or more different accounts all naming one token within ten minutes. The tool looks for that pattern and flags the comments as spam, which removes them from the gauge, along with any calls they created. It cannot catch everything, so treat the gauge as an estimate.

Every bullish or bearish comment is stored as a call on the Bitcoin price of that day. Neutral comments are not scored. Once 60 days have passed, the tool compares Bitcoin's close then with the close on the day of the comment. A move smaller than 10 percent in either direction counts as unresolved, because it is too small to call anyone right or wrong. A bullish call is right if Bitcoin rose by more than 10 percent, and a bearish call is right if it fell by more than 10 percent. The scoring runs once a day.

Those results feed the leaderboards. A commenter needs at least 5 resolved calls to appear on an accuracy board, so one lucky guess does not top the list. Rankings use a cautious estimate that discounts small samples, so a short record cannot jump to the top. The bull and bear boards also need at least 5 bullish or bearish comments. The profile of a commenter shows their calls and record.

Boards include Most Accurate, Least Accurate, Biggest Bulls, Biggest Bears and some for fun, such as Most Active, Essay Writers and One Liners. A board called Who to Fade First ranks commenters who were wrong often, weighted by how one-sided their comments are and how recent their calls are. It is a ranking of past record only.

Two validation panels sit below the chart. The Contrarian Edge panel shows how often the opposite of a bearish or a bullish comment turned out right, with a range that says how uncertain the figure is. The backtest is a deliberately blunt test: it goes long when the gauge drops below 25, short when it rises above 75, and holds for 60 days. A mirror option follows the crowd instead. See our glossary on backtesting for what such tests can and cannot show.

This tool measures one crowd. Read it with these limits in mind.

A made-up example. The gauge reads 22 after a sharp drop, and the comments are full of bearish calls. You open the chart and see that the last time the score fell this low, Bitcoin was already well off its high. You then check the Contrarian Edge panel and notice it rests on a modest number of calls, so you do not lean on it.

You use the tool as one more view on how the crowd feels, next to other sentiment tools such as our guide on DCA during extreme fear. It helped you see the mood. It did not tell you what to do.

This explainer is free for everyone. The tool itself is Premium only. A free member who opens the page sees a Premium card, not the gauge, the charts or the leaderboards, and the data behind them is only served to Premium members. See pricing for what is included today.