All you need to know to answer this question is that parenthesis override any operator precedence.
Parentheses are part of the operator precedence 😉
Warning: this is somewhat off topic, but I think that statement is sufficiently wrong that I can properly go off topic to challenge it.
Not unless bracket pairs are operators, in which case you are talking about operators that (a) are distributed as opposed to unitary (the brackets occupy two positions in a string, one for the opening and another for the close, and the matching rules have to be observed), (b) operate on the semantics of an expression rather than working on the values of arguments, and (c) have somewhat strange scope rules (I think - as I'm not aware of any attempt to define brackets that way it's possible that there's some way of avoiding the complicated scope rules that I haven't noticed. I'm not aware of any mathematical language that treats brackets that way. It's far easier to have a BNF definition of expressions that defines (using sematic marking additions to basic BNF) how the brackets override operator precedence and the left to right rule in an obvious manner. At least that's what I thought back in 1968 - I'm hopelessly out of date on this sort of thing, so maybe I should not still think it (although I do).
Actually, Polish notation (whether reverse or forward) makes life much easier than infix notation (the style used in T-SQL and most other computer languages) - there's no need at all for brackets or even for operator precedence. There are a few computer languages that use reverse Polish notation for expressions and some (a smaller few, I think) that use forward Polish, and when I look at the debate on this QotD I'm tempted to say that it's unfortunate that school arithmetic and algebra and all computer languages don't use (reverse?) Polish notation.
edit: I find it pretty easy to resist that temptation, because that would also require all computer languages to have a sufficiently strong type system to make Polish notation unambiguous in the small scale rather than requiring the whole of an expression to be parsed before one knows whether any tentative parsing so far is correct, and this screws up some forms of polymorphism (maybe undesirable forms, but maybe not).