Copper resistivity as a function of temperature

Using temperatures can be tricky and usually you are best off using Kelvin throughout.

Your version1 works, but in general when you subtract temperatures you have to use Delta °C and not just °C.

But then you can't divide a temperature in degree not just by ! K or 1 °C or 1 Delta °C to get rid of the unit

To strip the unit you first have to subtract 0 °C and then divide by 1 Delta °C = 1 K (NOT by 1 °C).

Basically these conversions are the reason why your short and at first sight intuitively looking version works.

Your version works because, while

this also means that

and if you divide that by (1)K you get the necessary value.

Using the correct(??) Delta °C would make the expression more bloated.

You may strip the temp unit from the result of the subtraction

or just from the parameter T and subtract the simple scalar 20

If you really understand whats going on you sure can stay with something like your version 1

as its the shortest and the subtraction of 20 °C  looks correct at first sight to the innocent reader 😉

Of course you can always do like ttokoro  did and omit the temp units altogether 😉

According the plot difference I agree with ttokoro 

The author seems to have used 0.000393 instead of 0.00393 when creating the plot.