[hunter2_]

> If someone counts me down, I expect them to say "5, 4, 3, 2, 1, GO", and I expect a timer to do the exact same thing.

"Exact same thing" isn't really possible here because the timer needs to show each integer for a duration of 1 second, while the spoken words each have a duration of much less than 1 second. So much less that they're really identifying a moment (perhaps the initial consonant, or the peak amplitude) rather than a window of time. This is why "go" is able to be intuitively anticipated accurately.

If the timer app flashed each number briefly, with no display for the vast majority of each second, then it could do the "exact same thing" as spoken English. Considering the transition from displaying one number to displaying the next as analogous to a spoken word is about as good as we'll get, in which case "1" ought to be displayed for one second immediately prior to "go" (don't display "0" for the last second and certainly not for a half second).