As a real-life example, here is the VSOP87 algorithm's variable function implemented with a fold() method. The VSOP87 algorithm is used to find planets and moons in the sky with really good accuracy, useful for simulators and telescope star finders, for example:

fn calculate_var(t: f64, var: &[(f64, f64, f64)]) -> f64 {
    var.iter()
       .fold(0_f64, |term, &(a, b, c)| term + a * (b + c * t).cos())
}

This is equivalent to this other code:

fn calculate_var(t: f64, var: &[(f64, f64, f64)]) -> f64 {
    let mut term = 0_f64;
    for &(a, b, c) in var {
        term += a * (b + c * t).cos();
    }
    term
}

And in C/C++, this would probably require a structure to hold the tuple. Five lines reduced to one with the same native code. As we talked about, this has no extra cost and will be compiled to the same machine code.