The past does not assign probabilities
Saying that something «looks like 2008» identifies a possible precedent, not the probability of another recession. The US Great Recession ran from December 2007 to June 2009, according to the chronology discussed by the Federal Reserve. That describes a completed episode. A new situation has its own conditions.
A useful analogy states what matches, what differs and what remains unknown. If it only lists similarities, it may become a persuasive story instead of an analytical tool.
A forecast must be able to fail
«Will there be a crisis?» is too vague. Define the event, place, indicator and deadline. Then keep the forecast and its date so that it can eventually be compared with what happened. Updating a forecast should not erase its earlier version.
Our laboratory shows similarities between energy episodes using published criteria. Those scores are not probabilities. The first editorial forecast has a date, conditions and a test: its quality can be assessed only by preserving the question, versions and outcomes over time.
A better reading habit
Begin with a specific question about a current story. Read a source on the earlier event and one on the present context. Look for a difference capable of changing the outcome. If the comparison still helps after that check, you have found a precedent worth studying.