The formulas are relevant for many reasons. The question such as "Why not buy the exact resistance" is a good one and very simple.
You can't buy any resistance you want, because it either does not exist as stock item, or you place a custom order for a resistor manufacturer to make you one, which will be both expensive and take time.
There are only standard resistance values available. For example, if after calculations you end up needing a 4.9 kilo-ohm resistor, the value does not exist in any standard. So you have to buy 4700 ohms and 200 ohms which do exist, or 4700 ohms and two 100 ohms, or any other combination ending up with 4900 ohms.
Same applies with resistors in parallel. If you need a 4545 ohm resistor, it does not exist, but if you put two resistors of 9090 ohms in parallel, you end up with the value you need. Same for any combination of resistances in parallel that end up with a value you need.
In the examples I used very precise values available in E196 series which has very precise tolerances and as per the name, there are 196 different values per decade. Usually you don't need such high precision and to ease up your inventory of resistors accessible to your students in a lab, you might only use E12 series resistors, which have 12 values per decade. So instead of stocking 196 resistors per a single decade, you can limit to 120 resistors of E12 resistors spanning 10 decades. Which should already be larger range than you need. You rarely need resistors smaller than 0.1 or 0.01 ohms, or larger than 10 or 100 megaohms.
So even if you need non-standard values, or even a standard value if the lab is out of stock for some popular resistor, you can calculate a substitute that is close enough for the required value. Say you need about 5k resistance, but both 4k7 and 5k6 are out of stock in the lab and you must finish your electronics project for a deadline and can't wait, you can just use two 10k resistors in parallel to get the 5k needed.
Also it allows to do some slight adjustments. For example if a voltage divider is used to set the output voltage of an adjustable regulator, it may be that with the resistor values you have available you cannot hit the target voltage exactly, but slightly below or above say 5V or 3.3V that is needed. You can always put a third resistor (in parallel or series) to fine-tune the value of one of the resistors to fine-tune the voltage to be closer or nominally exact 3.3V or 5V, if it matters.
So understanding the applications of what you can do with the theory is useful.
Another example is to calculate input and outptut impedances of voltage dividers. For example, if I am making a product with coaxial SPDIF digital audio output, it must have 1Vpp (unterminated) voltage output and 75 ohms output impedance driving the coaxial cable that has 75 ohm characteristic impedance. If the digital signal is on a 3.3V buffer, I can calculate the voltage divider values needed because I know the 3.3Vpp must be brought down to 1Vpp (unterminated) and 75 ohm impedance.
Another, maybe even obvious example, is that if you need to make a 2000W heater, but have only certain type of heating wire used to make many types and different power heaters, you can calculate how many parallel wires you need to make an European 240V model and American 120V model with the same heating wire, just arranging the heaitng wire pieces in parallel and series combination as required.