The document you cite shows the 80% capacity level as low as 600 cycles. 80% is the level customarily cited in life cycle testing. It is true it does flatten out a bit after that. But 70% is pretty much a substantial decrease in range.

The second reference is for an entirely different 400 mAh cell they are studying for EV battery packs. A white paper on what could be done.

There are several graphs. It would be helpful if you pointed out which one you are referencing. The first reference, the Panasonic D&E Forum 2011 has mostly 100% DOD charts. The NCR18650 cycle performance chart for example is a 100% DOD chart and has the roughly 80% capacity at 600 cycle data point. The cycle durability test in that one also shows roughly the same, but it does show that there is usable capacity at 25 degrees C out to 2000 cycles at roughly 75% capacity. That alone was huge, as it meant that one could have far higher life span if one was ok with roughly 25% capacity loss. Basically, the curve flattens out dramatically.

Until we found the 2nd reference, it was hard to quantify the gains given when one didn't cycle to 100% DOD. We had charts showing the effect of temperature on capacity loss in both cold and hot, we had high discharge, but we didn't really have charts showing limited DOD with high cycles and high charge/discharge rates. This second reference shows that. It uses the same cathode chemistry as what goes into the Model S and is tested with 18650 cylindrical cells. Not only does it show that limiting DOD dramatically limits the capacity loss (less than 20%) out through 3,000 cycles, it shows that it can have such capacity retention even with high charge and discharge rates. Now, a real world Model S will still have higher discharge rates, but lower charge rates and the vast majority of vehicles will have far less DOD. This second chart shows why Tesla can offer unlimited mileage warranty on the 85kWh battery pack.

3,000 cycles at a very conservative 100 miles per cycle is 300,000 miles. It still isn't' clear how many very small DOD cycles relates to a full charge cycle. Let's just say for a moment that worst case scenario is that every day is a full charge cycle. For 3,000 cycles, that 8.22 years. Obviously, it doesn't just drop off a cliff at 3,000 cycles. But let's just say it's a hard stop at 3,000 cycles. If you drive 150 miles a day each and every day of the year in a 85kWh Model S, you won't be at 100% DOD. That's 450,000 miles in 8.22 years or roughly 55,000 miles a year. What I don't know is if you drive 75 miles a day each and every day, if that also counts as the same kind of cycle every other day. In other words, just because you can drive 450,000 miles in 8 years doesn't mean it will achieve 450,000 miles of life in 16 years. We don't have good data on the absolute longevity in terms of decades.