Not sure what you're getting at here. Are you suggesting that the
(theoretical) 3dB increase occurs only when the total system power is
doubled, but not when the cone area is doubled?

NEITHER!!!

You get the 3db for the simple reason that two like sounds combine.
it has nothing to do with doubling the power or doubling the cone area!
(other than the power or cone area make the sound in the first place.)

This is the same thing though. Two like sounds combining is the same thing
as doubling the cone area (note: all else being equal, which as you pointed
out before, is a difficult proposition in the real world). Oftentimes,
modeling loudspeaker output consists of treating the cone as an infinite
amount of point sources. Whether you double cone area or double the number
of cones, you're doubling the number of "point sources". In effect, you've
got the same thing going on. So to say it has nothing to do with doubling
the power or cone area is not correct. It has everything to do with it.

I can have a speaker with an 8 inch cone and a speaker with a 15 inch
cone, the 15 may not be louder... so cone area alone is not a factor...

That's correct. There are other variables at play. But there's an
interdependence between all of them.


I can have a speaker being driven with 100 watts and when I increase it
to 200 watts I may not get 3dB more ...

You probably won't. Power compression is a key reason.

BUT, if I have one sound and add another sound the same SPL then I
WILL get 3db!!!

Assuming the wavelength is long enough.

See your way of doubling the cone or doubling the power MIGHT NOT
GIVE ME 3DB.... But my way ALWAYS DOES!!!!!

Not sure what you mean here. Doubling the power or cone area WILL ALWAYS
yield a 3dB increase, assuming everything else is held the same. That is,
same power output, same enclosure size/tuning characteristics, etc.

And, if I were to add two like sounds that were correlated
(in phase perfectly, very likely for bass in a car) then we would
increase
6db!!

the 3db for cone doubling and another 3db for power doubling is a MYTH.

It most certainly is not a myth. You're simply failing to keep the other
variables constant.