1.9999 seconds would show as 2 seconds on the clock. So the most that the could have been on the clock was 1.000000. And I mean, this most.
Actually, this is true and I stand corrected on my previous hypothesis that the max could have been 1.999.
There are exactly 24 1-second intervals between the values of 0 and 24 that the shot clock can take. So the shot clock starts at 24.0, and changes values every time 1 full second has elapsed. When the value changes to 23, that means at least 1 full second has elapsed (but not necessarily two full seconds, or the clock would have changed to 22). By extension of this logic, when the shot clock shows 1, then at least 23 (but possibly more) of the shot clock has elapsed.
In other words, while it's still true that the actual remaining time is not necessarily exactly equal to 1 second to the zero, it actually could be anywhere between 0.001 and .999 seconds.
The reverse counting threw me off this one a little bit. But the answer should be absolutely obvious if one recalls that the signal goes off exactly when the clock hits 0, rather than 1 second after that.
I actually timed it after the game and I came up with 1.45 seconds for that one second to come of the clock. Make no mistake about it, we were jobbed on that play. But more than likely, it was the clock operator giving the hometown aid there.
The clock operator is supposed to start the timing when the referee put his hand down. So if you're timing the play, it may be instructive to see when this happened.