Theres no difference in fuel economy scientifically for running premium fuel unless you are tuned. The engine isn't knocking on 87 unless you are in the most extreme load conditions (like really hot weather). It was tuned for 87. It does advance timing with premium though. Advanced timing may help you drive more efficiently, but scientifically 87 and 93 octane contain the same amount of energy. It's more like the "placebo effect." Those who don't beat on their car will not notice a difference between 87 and 93 unless it is extremely hot outside.
I saw your posts on Cruzetalk, you've been fighting this 'myth' for a while

Don't be offended, I just don't agree with you on a few points.
- 87 and 93 do not have equal energy content, 93 actually has slightly LESS energy per gallon! But with more ignition advance, you can get more torque and efficiency.
I have been short shifting at 2000 rpm and below, keeping boost between -1 and 1 psi. So at low rpm and relatively high (for commuting) 40-50% throttle load, ignition could make a measurable difference.
- If you shift at 3000K+ rpm, at a lighter throttle input, then you may not notice any difference
On another thread I started:
http://www.sonicownersforum.com/forum/t1597/ laborsmith said:
Every time you fill up (per what I have read) the computer starts as if you used preium and backs off timing as it perceives knock. In other words, the computer resets itself
- So per what laborsmith says above, the ignition advance is set for 93, but the computer pulls the timing back if it senses knock. Maybe he could tell us more about where he read that. The computer will pull back the ignition in fractions of a second, you would not be aware of any knock.
- I understand the placebo effect, I also mentioned it in the thread above. To measure small changes, like <10%, you have be very careful about keeping all the conditions consistent.
I just ran 4 tankfuls of 87 (10% eth), now I'm gonna run some tanks of 93 (10% eth). I'm not claiming the results will be 100% accurate, the driving mix may be different, traffic, etc but averaging over more tanks will smooth some of that out. A big factor will be the warming weather, less warm-up time, less highway aero drag, all these are just ~1% effects, but they can skew the results.
Again, as I said earlier, the only way you could get a difference with identical driving styles is the amount of ethanol in each. Where I am, 87 is 10% and 91 is 0%.
Eric, I agree, ethanol screws up your fuel mileage. But again I don't agree that's the only way, like I said above.
As I wrote in the thread:
http://www.sonicownersforum.com/forum/t1030/ I can't get 0% eth in my state

I spoke to a fuel distributor and someone at my state's energy department. They said during the winter, all the fuel has 10%. Our pumps are marked "Up to 10% ethanol".
A few days ago, I finally ordered the ethanol test kit,
Fuel-Testers E10 Gas Precautions & Tips + Ethanol Fuel Test Kits to protect your engines from alcohol-caused damage.
so I'll be able to check what it really is, supposedly some distributors over-blend eth