I think some of these values are nuts.
What exactly are you doing? the manufacturers need to COVA their ass.
I am lookin at a curve, at 35C above ambient for a single strand you get 3.75 amps. For a cable of four you get 2.8 amps.
If you have 80C rated cable, and your in normal places, you are still 45 C short of getting to the actual cable rating.
Then you have deviations like color of the cable, how its routed (race way or hanging in the air?).
If you are selling something to people you wanna definitely obey this stuff. If you are doing it for yourself then you have some leeway, so long you follow some precautions, like setting fuses. Different loads are different, how well behaved is your load? Can it stall out and draw massive amounts of current? Do you have a hard fuse or a poly fuse or a circuit breaker that you might keep resetting? What is the run time of your device?
For instance welding machine work cable is so different then NEC!
If you followed NEC for welding machine work cable the results would be preposterous beyond belief.
I used a 20 AWG 4 strand cable with 80C rating to carry 12 amps!
What is the result on this black TPE cable? After an hour of run time the cable gets warm to the touch. Its connected to a machine that makes alot of noise and its hanging in the air in a spiral shape. IF I take this machine out in the hot sun and let it run indefinately on the hottest day of the year, we might get close to the rating. The cable rating guideline is 0.8 * 6.5 amps (4 strand on 6.5 amp rating for single strand), aka 5.2 amps per wire, doubled up lets call it 10 amps. I ran it on 12 amps.
The table does not factor in the color, which has an effect on dissipation. This cable is braided with bare copper, no foil. The table just says to ignore shielding. Is this correct? Who knows. I imagine a braid might insulate this less then a foil. Maybe it just assumes insulated foil.
So oh no, I am running at 120% overload!!!!
But what does my hand tell me ? in a room temperature enviroment this cable is slightly warm after 2 hours. If it was zip tied into a strain relief, or jammed super tight inside of a cable routing thing inside of a box I might think hmm can this cause a problem??
When its floating in the air my analysis tells me that its safe enough for a loud specialized tool that might run for a few hours of supervision. Am I going to redesign this for 18 AWG? NO! I am happy with it! Would I sell it ? No, because someone might get upset. But in my case this lab machine runs just dandy and there is IMO No hazard.
For a different example, a spot welder. It has 8 AWG silicone wire. I have to handle the wire and electrodes with my hands to use it. It welds at 1500 amps. If I use it alot, the wire gets hot. Hard to touch. If I was going to automate this machine I would up the gauge if its being used heavy because it gets seriously hot (but it goes into thick copper bus bars in the machine). is 8 AWG wire the wrong choice for 1500 amps? No, it makes sense for k-weld. Infact its what the manufacturer supplies you with! Would making thicker cables improve quality? Maybe, maybe not. The dexterity offered by the thinner wires might allow for better more repeatable probe position and more repeatable spot welds. If I am fighting heavy cables so the wires don't get hot, I might do it wrong and compromise the weld. Logically I think I should just leave it be!
How about my tig machine? If the wire gets hot on your leg then let it rest! No NEC here! Its not even a normal wire. They route argon through the weird super flex braid to cool it. So its a wire inside of a gas tube that cools when your passing tons of current through it. Leave it to manufacturer to decide.
How about my wall outlet? You bet your ass I won't deviate from the NEC recommendation when its powering welding machines and it deserves some extra copper! no fooling around in the wall