"give you about 56-68% ethanol" !!
Now isn't that an interesting proposition! Under gov legislation 10% is the maximum allowable level and must be branded (exclude flex-fuel), with 5% un-branded, although waivers 'are' issued under certain circumstances. Trouble is vehicle manufacturers won't honor warranties where more than 10% is involved...hmmmm.
Are you sure he knows the 80% is 98 and not the other way around.

whenever the conv comes up about which 98oct fuel people use i always hear neg stories about caltex 98.

the pdf that Purple5ive was talking about is located here

My calculations came up with a rule of thumb of 10%E=~2.5-3 octane points, and this seems to coincide with what I've seen elswhere, but that reduces obviously as the % increases.
Other thing is, I've been doing a bit of reaserch to find why low level splash blends are so effective in many cases, and  found a couple of answers in a SAE white paper.

just_ace

whenever the conv comes up about which 98oct fuel people use i always hear neg stories about caltex 98.

Well I can attest that Vortex 98 runs quite well in my cars, but there is a significant reduction in fuel economy. Where I'd see 350km from 36L in my SW20 (with say Shell 98 V-Power), I'm suddently seeing 310-320km with identical (as close to as possible at least) driving patterns. This could theoretically be due to higher ethanol content. I do prefer how it drives compared to V-Power though, feels smoother/more stable. Although in all fairness, my SW20 never seemed to ping with either of them.

I worked it out as:
12% E-Flex (~6.5 litres) @ 105 RON (max) = 12.6
88% ULP98 (~48 litres) @ 98 RON = 86.24
86.24 + 12.6 =  98.84 RON (total)
 
Well I worked it out to about 28% E-Flex (E85) to 72% ULP98 should provide 99.96 RON so not quite the magical century but very close!! 
Using the 80/20 mix only provided 99.4 RON.
28% E-Flex is a lot of ethanol running through your car causing harm. 
 
Hence, you'd need > 28% E-Flex to reach that 'JDM 100'.

The increased RON isn't linear, go into that in a min.
Just going on real world results ie.
MHR claim: 95 RON unleaded fuel blended with 10% Ethanol will achieve an octane rating of between 97.5 and 98.5 RON.
Shell claim: 98 RON plus 5% = 100 RON.
Manildra claim: 91 RON + 10% = 95 RON. etc, etc. gives some initial indication, bearing in mind this is E100.
The white paper from SAE Int. is copywrited & internet post restricted, so I can't post the graphs. But basically here's some basic data that's relevant:
 
RON - Petrol (typical) 95 - Ethanol 109 - Methanol 106.
MON - Petrol (typical) 85 - Ethanol 98  -  Methanol 92
Specific Energy (MJ/kg mixture) - Petrol 2.905 - Ethanol 2.978 - Methanol 3.062
Oxygen Content by Weight (%) - Petrol 0 - Ethanol 34.8 - Methanol 50. 
 
a. If you look at RON levels, you see an obvious increase between petrol and ethanol. Now look at the MON rating and you can see a distinct advantage with ethanol, in a more 'real world' rating.
b. When you look at specific energy levels, there's negliable difference between the two at 100%E (at stochiometer), so really there's little gain in the ethanol itself at that level, and gains are coming from VE and increased Ign Advance to MBT.
c. SAE states: Direct injection of alcohols can also give significant increases in volumetric efficiency in spark ignition engines, and even a relatively small amount of ethanol or methanol blended in petroleum fuel* can have a significant effect [12]
An accompaning graph lists 'VE Relative to 95 Petroleum Fuel*' vs Volume Fraction of Ethanol in Petroleum Fuel*
The graph showing smaller quantities of ethanol maintaining a higher level of VE in relation to 95 petrol*, on a linear scale.
In other words, the first 1% will maintain a higher level of VE than will the last 1%.
d. The oxygen content by weight of 0 vs 34.8% is self explanatory as far as VE in concerned.
 
Anyhow, having said all that, the only way to determine how it will effect your engine is to put it to the test. 10% is safe, 20% is safe in a turbo engine (according to my logs).
As for fuel system damage, 10% is safe, any more is unknown but I personally dispute some of the claims for a number of reasons, but a few fills isn't going to hurt anything, unless it sits around for a long period. By far the biggest issue is water contamination as that is the cause of any tank corrosion. O2 + H2O = rust.
  
  
  
  
 

It's not based on any size 'tank', just ratios and percentages. It's all relative.
I'm pretty sure in Shells case (sounds like V-Power Racing) that if you had 100 litres of ULP98 and then added 5 litres of Ethanol to the 100 l mix, you'd end up net 100 RON. Correct me if I'm wrong MuMan....
 
I still don't understand how measuring in RON or MON makes any difference to the interpretation of the advantage of ethanol?