This discussion is only getting more interesting.
Here's another wrinkle on this: there appears to be penalty in the DS if you point too high, a place where you might think you are still sailing close-hauled but effectively you are almost or already "pinching". The result would be a higher leeway angle, so that while your heading looks optimistic, the course isn't that great and the speed falls off below the optimum as well.
The difference is not 10 degrees, but much less. I'm not sure I've seen differences of a 1/4 of a degree, but 10 degrees is huge. I'm not sure I know how to reliably tell when I am in that condition, the lack of opportunity for comparison with a like boat leaves the GPS as a poor substitute.
For keel boats it's common to have a so called "polar diagram" where you can read off the theoretical best boat speed for each wind angle (at a constant true wind speed). The maximum boat speed upwind is usually at some close hauled angle, with a steep fall off as you "pinch" but before you stall the sails. This curve would be different for each wind speed.
Something like Talbot's calculation could then be used to translate the curve from boat speed to VMG (or effective speed to an upwind destination). The conventional wisdom is that this curve still has its maximum at the close-hauled angle (the same angle where speed through the water is maximized). However, again, you need to draw one curve for each wind speed.
The question is then whether in light or very light winds, for example, it remains the case, that the maximum boat speed and maximum VMG continue to coincide for a displacement hull.
Part of the reason they should is that, especially in light airs, the boat speed contributes so much to the apparent wind speed, and more if you point higher. At the same time, at the relatively low boat speeds you might have the best chance at driving the boat noticeably faster with a correspondingly smaller increase in power (as you get near hull speed, fat chance to get faster at all). Light air racing is definitely different, but there are many other aspects besides pure tacking angle that drive success there. (Where you are and which wind you have trimmed your sail for matters a lot when a puff finally comes -- I remember once being surrounded on all four sides by boats just a few boatlengths away to find myself accelerating and sailing away from them, while they appeared to stand still -- you'd sworn I was using a motor...)
One interesting effect is cited in the article linked by K.C. - the transition between laminar and turbulent flow over the sails around 5 kn of apparent wind, with its corresponding strong increase in driving force (and ability of the air flow to stay attached to fuller sail):
The technique is to sail the lulls with "Light Air" settings. At
the onset of a puff, accelerate by sailing a little full, so that the leeward jib tuft is
just on the point of agitating. ... As the boat accelerates, the sails will suddenly begin to
"pull" -when you feel this, sheet jib and main in smoothly towards their "Moderate Air"
settings, ... and point as high as possible while retaining speed through that puff.
As soon as the puff dies, return to Light Air" settings... This is a "flat water
only" game: it won't work in waves.
Here the "win" is not in getting a higher speed over the total leg by sailing fuller, but just to get a higher initial boat speed, which can then be "converted" to higher wind speed by pointing higher, for a final win providing both higher speed and high pointing angle -- as long as the puff lasts. Playing this game requires control of sail shape with vang etc. that not every DS is set up for. Also, whether this works as quickly as needed to take advantage of a light air 'puff' is a question, because the DS is heavier and accelerates less quickly.
The "sailing a little full" that is mentioned here, I would guess, is something on the order of at most 5° probably much less; I've never measured the angle needed to get the jib telltale to rise, perhaps someone has a more accurate answer.