Sometimes experimental or observational evidence from credible physicists points to new physics and then turns out to be wrong due to statistical flukes, experimental error or a theoretical analysis mistake. What cases of this happening do you find most notable, what showed that the hints were unfounded, and why do you find them most notable? I'll start with four, to which you can add your own (or which you can discuss in their own right). Also, what lessons should be learn from these experiences as new hints of BSM physics come up over time? Ultimately, I find that these false starts strengthen my confidence in the scientific process, because they were doggedly analyzed and ultimately debunked, suggesting that the scientific community is appropriately skeptical even in the face of announcements from authoritative sources. (For clarity, this thread is not about cases where experiments actually established what constituted new physics at the time or currently unresolved anomalies. It is only about new physics "fails" that at first looked like proof of new physics.) 1. Opera's announcement that it had observed superluminal neutrinos. In 2011, the Opera collaboration had detectors at a collider and then another detect hundreds of kilometers away. By coordinating observations from those detectors and using precision clocks, they reported that neutrinos seemed to be moving slightly faster than the speed of light, which is prohibited by Special Relativity. The problem in their instrumentation was revealed nine months later. This was notable because it was a widely reported result of a major big physics collaboration and went to the very fundamental laws of general and special relativity illustrating how the physics community might respond if a real breakthrough did happen. In the end, it turns out to be due to a wiring problem in the measurement equipment. 2. The purported discovery of cold fusion. In 1989, Martin Fleischmann (then one of the world's leading electrochemists) and Stanley Pons reported that their electrolysis-like setup was generating modest amounts of net energy, which they attributed to a low temperature fusion process. This was notable because is was proposed by respected scientists despite a lack of any well articulated theoretical basis why it should happen, was widely publicized, and would have had profound economic effects if true. The result was not replicated (something that should have been easy to do in their table top setup) and when studied more carefully did not show signs of fusion byproducts. Within a few months the discovery was discredited. In some ways, this is the most puzzling of the four incidents that I list here, because the scientists involved knew that they were making extraordinary claims, but did not exercise the extreme care that they should have to make sure that their evidence was similarly extraordinary before announcing it. It may be that the commercial potential of this claimed discovery put pressure on Fleishmann and Pons to announce their results early in the face of the likelihood that someone would leak this inexpensive to carry out process, potentially depriving them of an inventor's profits from the discovery. The other three examples that I cite had no real commercial application in the short to medium term, even if they were true, and were difficult or impossible for more than a handful of research collaborations to replicate or stumble upon in the meantime. 3. The Pioneer anomaly. A lot of the research discovering and than explaining this anomaly came from high prestige and authority NASA scientists. As the linked Wikipedia article summarizes: Thus, in this case the measurements showing the anomaly were accurate but the theoretical analysis used to determine the predicted value missed one very subtle point. While this was a less high profile issue in the popular press than Opera's superluminal neutrinos or Cold Fusion, it too went to very fundamental points of the law of General Relativity in a very clean solar system test never before conducted where there was anomaly that was correctly measured and statistically significant. Also, while the other two were ruled out in a matter of months, this anomaly wasn't conclusively explained for 32 years, and it took four years of serious investigation before the result that ultimately came to be accepted was proposed. 4. The 750GeV diphoton excess. In 2015, the LHC saw events which pointed to a new particle with a mass of 750 GeV in the very clean diphoton channel in which the Higgs boson was discovered, and many new physics theories, such as "two Higgs doublet" theories predicted a Higgs boson-like particle for which that mass would have been plausible. It turned out that this was a statistical fluke and when more data was collected in 2016, the combined data actually disfavored the existence of this resonance. This is very much an "inside baseball" incident that attracted a massive rush of papers from physicists but received much less attention in the general public. But, it is particularly notable because the high prestige physics collaborations in question in this case made no experimental or theoretical mistakes, they just happened to be suckered by a fluke of random chance. Also, it is hard to fault the physicists in the "gold rush" of papers written to explain this phenomena, because the practical reality in the scientific community is that to receive credit for making important discoveries, you have to articulate theories describing anomalies in published papers long before they reach the generally accepted five sigma discovery threshold. So, you have little choice but to take the risk that an anomaly will turn out to be a fluke if you want to have any hope of ever being considered an important figure in discovering BSM phenomena.