In certain fields, professional experience really does convert into reliable intuition. An anaesthetist notices something wrong within seconds, a firefighter assesses structural risk before entering a building, a strong chess player sees the shape of a position at a glance. These abilities are real and they survive testing under controlled conditions. What they have in common is feedback that arrives quickly, unambiguously and in volume. The verdict follows the judgement closely, errors announce themselves without needing to be looked for, and similar situations recur often enough that the pattern behind them can actually be learned.
Markets provide none of that. Whether a judgement was sound may take years to establish, and by the time it does, conditions have usually changed enough to strip that verification of most of its value. The proportion of noise makes matters considerably worse. A poor judgement can produce a profit for entirely unrelated reasons, and a sound one can look wrong for a very long stretch before resolving. Under these conditions the link between years of experience and accuracy of judgement is far weaker than people assume, and in some studies it disappears altogether.
The difficulty is that confidence of expression does not weaken alongside it. Confidence comes from the practice acquired by expressing views publicly over many years, and that practice is entirely independent of whether the views turn out to be correct. Somebody can publish opinions for three decades, be quoted throughout, steadily improve how clearly they explain themselves, and improve their accuracy not at all. Nothing in the feedback they receive would inform them otherwise. From the listener's position, such a person sounds exactly like somebody who is genuinely accurate.
This creates a practical problem for anyone on the receiving end. Outsiders cannot infer the quality of a judgement from the quality of its expression, and expression is the only thing directly observable. The criteria used to select which experts to follow therefore slide towards clarity, consistency and persuasiveness, none of which requires any deliberate deception to develop. In a field with poor feedback, those three qualities bear no stable relationship to being right, and they may even correlate with overconfidence.
A more workable approach changes the role expert opinion plays in one's own thinking. Treating a statement as a source of conclusions outsources the accuracy problem entirely, and the accuracy problem is the one part that cannot safely be outsourced. Treating the same statement as a source of arguments extracts only the parts open to examination. What data was seen, how the reasoning proceeds, and which single link would collapse the whole thing if it failed. Far less survives that treatment, and what survives can be checked independently.
Keeping a record costs almost nothing and helps considerably. On encountering a forecast, note the date, the substance and the exact wording, then return to it six or twelve months later. Most people have never done this, and consequently hold no evidence whatever about the sources they have been following for years. The exercise is unusually cheap because the material arrives free and the checking takes minutes. After two or three rounds of it, the list of sources tends to shorten on its own without any deliberate decision.
None of this suggests expert opinion lacks value. On questions with determinate answers, professional knowledge holds an enormous and genuine advantage. Where to obtain data, how an industry is structured, what a particular regulation actually requires: no amount of independent reading replicates that. Reliability breaks down specifically around prediction. Prediction happens to be the output most often requested and most readily supplied, and it is also the one most easily confused with the parts that are dependable. The confusion is understandable, since both arrive from the same person in the same tone on the same afternoon.