It’s entirely possible for there to be a large discrepancy between IQ and SAT scores. I know someone with a professionally tested IQ of 135 and an SAT score in the 1500s on the 2400 scale. I don’t see why it can’t go both ways.
The research carried out by Frey and Detterman is only ONE study! The studies correlating SAT/ACT and IQ are scarce and limited. Also that study analyzed data from the 1979 National Longitudinal Survey of Youth so it used the pre-1994 test which was much more like a group-intelligence test at the time. Since then and 2005 the SAT has undergone some considerable changes to content. It is much more like an achievement test now and assesses one’s grammar skills and vocabulary, reading comprehension under strict time limits, mathematical problem solving under time pressure, and mastery of math skills covered from algebra I, geometry, and algebra II. Now while there is still some contribution of intelligence to test scores it is doubtful that the correlation is as high as .86 (corrected for non-linearity) for the 2005 revised test. It still takes someone intelligent to get a perfect score of course, but just because someone gets an average score doesn’t mean they’re not intelligent. There’s a good number of “gifted” students who have slower processing speeds that can hurt them on these tests. Clinically administered IQ tests are a lot more flexible on time in a lot of subtests they administer because IQ or intelligence or however you want to define it is not all about speed. Also the ASVAB correlates with other measures of IQ around .7 which is substantial but not enough to establish cross-validity!
To become eminent in a field like physics you need intelligence, competence, creativity, and a passion!