> According to this, the tesla regenerative braking is just as efficient as the motor is at acceleration.

That's true but misleading. If the motor's combined electrical and mechanical losses cause only 64% of the battery's energy to propel the car forward, then by that reasoning, only 64% of the braking energy is made available for charging.

Let's look at all the factors that influence the outcome:

1. Both electrical and mechanical losses must be accounted for.

2. The electrical loss creates a factor of 0.8, the mechanical loss also create a factor of 0.8. Result 0.64.

3. On the Tesla, only the rear wheels have regenerative braking, the front wheels have disc brakes. Factor 0.5, so we are now at 0.32.

4. During a stop, the car pitches forward onto the front wheels and unloads the rear wheels, this needs to be taken into account. Factor 0.8 for a total of 0.2: 26%.

So I stand corrected -- this analysis concludes that the regenerative braking scheme recovers 26%, not 20%, of the original kinetic energy.

A literature search reveals that the best regenerative braking systems that involve all four wheels and in which every effort is made to maximize efficiency, can produce about a 40% energy recovery:

http://www2.hesston.edu/physics/201112/regenerativeenergy_cw...

A quote: "Since regenerative brakes can only recover about 40% of the energy we take the recoverable energy and multiply by .4 (Everett, Michael)."

The linked paper goes on to show that 40% is actually overly optimistic for various practical reasons. But notwithstanding that, we can take the original 40% figure and apply it to the Tesla. By taking into account that only two of four of the Tesla's wheels have regenerative braking, we arrive at a figure of 20%.

> And this page about the Volt suggests that the efficiency might get up to 70% in some situations.

No reputable source makes claims this high for a full round trip from battery stored energy to battery stored energy (the linked article doesn't try to do this). Also, the linked chart reports energy recovered in a four-wheel test and measured before being applied to the battery, not energy delivered to the battery as stored energy.

> This is confirmed by measurements which show pretty much all hybrid vehicles get better mileage in city driving than in highway driving.

That's true, but that doesn't result from regenerative braking. It results from lower average speeds and reduced wind and mechanical resistance.

Remember that regenerative braking cannot possibly improve on driving at a steady speed. That would violate some very basic physical principles.