You're calculating the fluidic drag caused by the hydrodynamic main bearings. You have not factored in the reduced friction in the valvetrain from the better/thicker oil film on the cams, lifters, etc. Also, better ring sealing around the pistons.
It was all, on average, factored in. Also, "ring sealing" is not the issue here.
I think the difference isn't there in the real world.
What you think differs from many real-world tests and engineering studies. Thicker oil has been proven many, many times to raise fuel consumption, and for most engines its about a ~1.5% increase per visc grade you go up.
This test got 2% increase in trucks:
http://papers.sae.org/2017-01-0888/
This one says about 1.5% fuel savings:
http://www.machinerylubrication.com/Read/30208/fuel-economy-oils
This one says 1.6%:
http://fleetowner.com/management/news/pouring-in-fuel-economy-1101
This one averages about 1.5% improvement:
https://priuschat.com/attachments/fuel-efficient-motor-oil-technical-article-pdf.11772/
Here is a 0w20 to 0w30 comparison that was about 1%:
http://s05.static-shell.com/content/dam/...-efficiency.pdf
Lots more....
The 1.5% is an average benefit, typical, and it does depend on a variety of factors. The spread is usually 0.5% to 3% better MPG.
A power increase also occurs as you decrease visc, same as the MPG benefit. Racers know this and often qualify on low-visc oil.
I've reviewed all of them. Nobody is recommending jumping
two grades, and the percentage increases in FE you've stated are virtually all based on going from a non friction modified Xw-20 to an Xw-40 oil, or from a dino 10/15w-40 to a 5w-30 syn oil. Going from an HTHS of close to 4 to one at 2.9 or less is more than a single grade jump. I mean, comparign the KV100 specs for 15w-40 (often exceeding 15), 10w-40 (in the mid/high 14's), and then a 5w-30 (in the 10's) is more than a single grade jump. Moreover, when friction modifiers come into play, the difference between adjacent grades (especially 5w-30 and 0w-40) shrinks to
well under a single percent. Not 1.5% as you suggest.
From the ML article:
"
Keep in mind that newer engines have tighter tolerances and better surface finishes. By using these lower viscosity oils with a cleaner engine design, you can optimize fuel economy. On older vehicles where the recommendation is a 5W30 or 10W30 oil, it would not be advised to switch to a lower viscosity oil. By decreasing the viscosity, you have the potential of added wear generation. With increased wear, you can expect engine life to decrease and performance to degrade. Once performance starts to fall off, your fuel mileage will likely decrease as well."
Lastly, it's quite clear that none of these tests for gasoline engines bear in consideration real world observations of engine longevity. One test is comparing running 5w-20 oil in a Porsche 911 (which specs at least 0w-40). How many
minutes does that 911 run in the real world on 5w-20 in the summer before it seizes?

Next, we have a BMW M5 (unknown vintage) that is compared using 5w-20, 0w-40, and 15w-40. Again, how many minutes does that BMW engine last in the real world during summer before it fails (consider that BMW has never spec'd Xw-20 oil for any M5 engine, EVER)?

Then we get into what causes friction in an engine:

So, in areas where viscosity is not necessarily the primary factor in engine friction, and where higher vis can actually reduce frictional losses (i.e. the valvetrain), the differences are not as stark. I'm just not seeing a meaningful statistical difference that even matches what different batches of gasoline can introduce, or with moisture instrusion into a gas station's storage tank, or with a mildly loaded air filter. I see little to lose, and much to gain by moving up in viscosity for the EJ. Many thousands of users have found the same.
Also, it seems that papers which advocate for the use of low viscosity oil don't comprehend the observed inadequacies of using thin oils in applications
which call for thick oil....so the act of substituting thin oil in engines which were not designed for it is a non sequitur. Most of the cited studies are also rather old, with much having changed in the last ten to fifteen years of oil additive and basestock technology. Further, these gasoline studies are speaking directly to altering engine designs to take advantage of thinner oils with newer (and many not-yet-developed additive packages), and not to substituting lower vis oils into the crankcase of vehicles which were designed for higher vis products.
Racers using low vis oil is a poor analogy, as they rebuild engines on the reg, with sponsorships to boot. What race engine goes over 200 hours between rebuilds? I digress. Nobody is sponsoring my car payment, or maintenance.