A challenge to the "measurement" camp

10 gague is too thick, disregard, not necessary, 12 THHN is all you need, But my guess if you A/B a  12ga strand vs a  14ga , you could not tell the dif,,I perfer 12 gague as cost is the same as for the 14 gague,,, and 12 ga fits nicely in the speaker post hole. 
Same with power cord DIY, 10ga is too thick, 14ga might be OK, but ideally 12gague seems absolutely perfect in my DIY powercord. 

What thickness is right?

Well, the only cable property that matters is it's resistance.
A cable resistance equals to the material conductivity constant (cooper),
multiplied by its length (in meters) and divided by its cross section (mm sq.).
This resistance shall fit so it will be way smaller than the Amp's DF. (Output resistance). 
Usually, the Amp and length required are a given. 
So what left to play with is the thickness.
Today SS good Amps have a DF of 200 and above...
PASS LAB = 200
EMOTIVA = 400
Luxman = 700
Tube Amp have a low DF (20 and less) so it's insignificant.
D-class tend to have a DF in the thousands: as high as 4000 and more.
Tell me what is your Amp's DF and what length you need, and I'll calculate for you the optimal thickness (in # awg).

A test been conducted two years ago, by myself, and those who pioneered it, ended up very happy. The old cable (no matter how expensive it was, and some were as high as $17,000 a pair, went into storage. The DIY offered ended up (mostly) under $100.-

...some interesting reading and maybe some clarity...
https://blogs.scientificamerican.com/cross-check/the-delusion-of-scientific-omniscience/
https://getpocket.com/explore/item/should-quantum-anomalies-make-us-rethink-reality?utm_source=pocke...
https://blogs.scientificamerican.com/observations/physics-needs-philosophy-philosophy-needs-physics/

It seems refreshing articles.... Thanks

Bernardo Kastrup is in itself a new sign of the changing tides....

 And Goethe will be understood tomorrow....

There is some hope perhaps in these creepy times....

...speaking of putting math, the primary tool thingee that allows measurements to be analyzed, into perspective....

https://www.quantamagazine.org/how-godels-incompleteness-theorems-work-20200714/

....which kinda implies that in the big picture things run out of precision pretty darn quick...to the point where most seemingly absolute proofs are reduced to wild-assed guesses in fairly short order ( assuming of course you are using analysis of any sort that relies on math...)...the problem that occurs when observations are reduced to theory which depends on the use of a logic type different from the observation...historically the difference between magic and scientific cosmologies ( and no, magic here doesn’t refer to hats and rabbits...one refers more to the name, and the other to the thing named...different logic types eh... )