I bet he learned the basic principle, but never understood it, and screwed it up when he taught it to you.
At least that’s what I’d tell myself if I were you.
Anyway, is the tenths digit equally likely to be any digit? Can someone tell me if that one thingamajigger principle where the last digit is 1 a disproportionate percent of the time applies here?
You’re talking about Benford’s Law, and yes the effect will be there for some sets of numbers, although it should be very small for the rounded digit in most cases. The effect is fairly large for the first digit, much smaller for the 2nd, much much smaller for the 3rd, etc.
The question is why do you wanna round? My mother always rounds down when she does her check book. If I owe someone money, I will round up.
Rounding becomes necessary when you apply calculations that result in a greater number of decimal places than you can conveniently work with, for example, when calculating percentage discounts or interest on currency values; you end up with parts of a penny that you must deal with by rounding.
I think ouryL was trying to point out that how you round (i.e. up, down, off) depends on why you’re rounding. The rule depends on the context.
As for the 5 thing, 38.5 is exactly halfway between 38 and 39, but 38.5… with anything after the 5 is closer to 39 than to 38. If you’re rounding to the nearest whole number, in the latter case, you should definitely round to 39. In the former case, there are different conventions, and what to do depends on which convention you follow.