Once you've done the mechanical stability checks that csorrows defined above, then you need to calibrate it.
Heartless' Tutorials (in order of operation):
My personal method:
I found this useful, but didn't follow it exactly because it has you crashing the machine like a fool.
I dislike testing extrusion steps/mm with the Hot End detached because there is no Back Pressure, which means the actual steps/mm will be different by an incalculable number. Total garbage.
Keep the Hot End on, heat it to the temp you print at, then calibrate the 100mm of filament travel.
This keeps back pressure closest to normal. Do this for every roll of filament. Don't mix up the numbers.
Mathematically calculate how many steps/mm of filament travel must exist.
Actual extrusion steps/mm will likely at least 10% more than this because of Hot End Back Pressure.
The filament diameter thing will invariably change through the roll.
Filament Diameter Setting is there for 1.75mm or 3mm selection, and doesn't help with fine tuning.
A bitter pill is that it is practically impossible to calibrate for filament roll variability.
Everyone who says to measure the filament with calipers is usually either ignorant or trolling.
Unless you've just opened a roll and are checking if it is too oval or damaged, but that's not a printer setting.
Standard retraction is 4.5mm @ 30mm/sec.
I prefer 1.5mm for most things, but most of my prints don't need that skill in the tool.
I haven't found perfect retraction when also having perfect part fill-out.
Personally I would love it if I could find one setting and just leave it, but no luck for me yet.
Ultimately fluid dynamics comes into play because the stuff isn't non-viscous.
The faster you try to pump out plastic, the more viscous friction is generated, which is harder to pump, which means less gets pumped with the same amount of power. This is what I've understood Flow Modifier is for!
The Proper Flow Percentage can only be obtained for a specific speed, for a specific printer, for a specific filament spool, and must be obtained experimentally.
If you change the nozzle diameter, you need a new measurement.
It is extremely sensitive. 0.1% change should produce dramatic results.
This is why people think using a caliper will do something useful when measuring filament diameter.
It doesn't do anything useful because what's really happening is the quantity of extrusion is different at different extrusion (thus printing) speeds, but most people don't make that connection and assume that the filament itself is the problem when it's the extrusion speed affecting extrusion quantity.