The Rolling Start, A Better EV Performance Metric


When comparing vehicles, many people give credence to 0-60 mph times as a surrogate for overall performance.  While there are other metrics that also quantify car performance; such as braking distance, skidpad g’s and track lap times; the standard many people use is the time it takes to accelerate from a dead stop to 60 mph.

This is a useful comparison for stoplight racers and for the drag strip, but it is not representative of how cars are normally driven nor how they normally perform.  The rolling start test is a better metric, especially when comparing electric vehicle performance against traditional cars.

The Classic 0-60mph Test


0-60 mph Launches Get An Asterisk* Over Reality In Most Cases

The technique to minimize a 0-60 run in a car with an internal combustion engine (ICE) involves brake torqueing (for an automatic) and/or a raised engine RPM (for a manual).  This puts the engine into a more energetic state before releasing the clutch and brakes.

So before the car starts moving, the drivetrain is in an optimal configuration to maximize the power available once the light turns green.  The turbos are spun up and ready to provide boost.  The flywheel and crankshaft are rotating and ready to drop that energy into the tires.  Performance cars even have “launch control” to assist the driver with getting the engine into the proper state and transferring the power to the ground.  But we don’t normally drive like this.

The Rolling Start (5-60mph) Test

Car & Driver magazine recognized the limitations of the 0-60 test and also collects “rolling start” test data.  They describe their testing in this article:

smart ed raceSince most owners will seldom subject their cars to brutal launch techniques, we also perform what we call a street-start acceleration test from 5 to 60 mph. While rolling with the car in gear, we floor the accelerator at 5 mph and shift quickly at the optimal shift point.” 

(Note that this is different and should not be confused with a “rollout”, which is the distance a vehicle can move before triggering the timing lights at a drag strip.  More info is available on this here)

In a rolling start test, the engine is at idle, with slow turbos and an engaged clutch or locked torque converter.  The car’s engine is obviously in lower energy state than when staged at a drag strip and the data C&D collected demonstrates the difference in acceleration times.  Even with the 5 mph head start, most cars take LONGER to get to 60 mph during the rolling start test but, crucially, the difference between the tests varies widely from car to car. To understand this variability, review the C&D test results from a representative selection of cars.

Some 0-60mph and 5-60mph Rolling Start Times (data via Car and Driver)

Some 0-60mph and 5-60mph Rolling Start Times (data via Car and Driver)

The differences between the cars can be attributed to characteristics such as turbo lag, available tire grip or engine torque curve.  The cars with turbo lag or a peaky torque curve tend to have a wider difference between their 0-60 and 5-60 acceleration times (ex: Macan S, 30% or Subaru BRZ, 27%).

Cars with larger displacement engines and no turbos perform similarly on the two tests because they produce higher power at lower RPM and don’t suffer from turbo lag (ex: Camaro SS, 3% or Camry V6, 5%).  This difference in performance between the two tests, which ranges from almost nothing to over 30%, indicates that the 0-60 test is not a consistent way to judge how responsive the car will feel in everyday driving.

Since the 5-60 rolling start test better represents real-world driving conditions and since 0-60 performance is not directly correlated to 5-60 performance, the 5-60 rolling start test is a more realistic and consistent performance metric than the classic 0-60 test.

BEV Acceleration

Now let’s look at battery electric vehicle (BEV) acceleration data, also from Car & Driver.

Some 0-60mph and 5-60mph Rolling Start Times (data via Car and Driver)

Some 0-60mph and 5-60mph Rolling Start Times For All-Electric Cars (data via Car and Driver)

Interesting, isn’t it?  The BEVs 5-60 rolling start times are MUCH closer to their respective 0-60 times; a number even have lower times.  This shouldn’t be a surprise because the tricks that ICE cars have to propel themselves off the line faster are not available in cars with only electric motors.  This means that if you only use 0-60 times to compare BEV vs ICE vehicle performance, the ICE performance results overestimate the real-world driving experience of the ICE car as compared to the BEV.

BMW i3 - 0-60mph, or Rolling Start, The Result Is Identical

BMW i3 – 0-60mph, or Rolling Start, The Result Is Identical

Two examples demonstrate this: 

*- Compare the Model S 60 results (5.5 sec/5.2 sec; 0-60 and 5-60, respectively) with the VW GTI (5.6/6.3).  They are similar from a standstill but the S 60 is a full second faster in the rolling start test.  The S 60 will feel more responsive than the GTI in everyday driving.

*- Compare the BMW i3 (6.5/6.5) with the M-B C300 (6.1/7.1), M-B CLA250 (6.3/7.0), Durango R/T (6.4/7.0) and Subaru BRZ (6.3/8.0).  All these cars have faster 0-60 times than the i3, but the i3 has a faster rolling start and likely feels punchier around town.

So What?

We should congratulate and appreciate Car and Driver for collecting performance data that is relevant and encourage other reviewers to do the same.  More so, rolling start times are easier to collect consistently because driver skill and road friction have a reduced influence on the results.

We should also be careful when comparing the 0-60 performance of the forthcoming 200-mi BEVs against ICE cars, especially those that are more “sporty” than the current 100-mi BEVs.  The ICE cars could be as much 30% slower in real-world driving.  Rolling start times, when available, are a more realistic and consistent metric.

The 2017 Chevrolet Bolt EV With A Stated 0-60 Time Of Less Than 7 Seconds Will Be More Than Competitive With

The 2017 Chevrolet Bolt EV With A Stated 0-60 Time Of Less Than 7 Seconds, Should Be More Than Competitive With Other Hot Hatches On The Market When It Arrives At Year’s End

The Chevrolet Bolt EV, a car GM claims to be faster than 7 seconds to 60 mph, may have similar acceleration performance to the i3; but with greater than double the range at a lower price, it will be uniquely positioned in the market.  It will compete directly with traditional cars.

Assuming the Bolt EV’s rolling start time is around 6.7 seconds, it will be marginally slower than performance-oriented hatchbacks, such as the VW GTI or Ford Focus ST.  On the other hand, it could easily be more responsive in real-world driving conditions than many entry-level luxury cars, such as the Audi A3 1.8T, BMW 320i, BMW X3 xDrive28i, M-B C300, M-B CLA250 or M-B GLA250.

Ultimately, we should use many diverse criteria to compare cars, and acceleration performance is just one.  If you value everyday drivability and responsiveness, look for 5-60 mph rolling start test data to compare EV performance against traditional cars.

Categories: BMW, Chevrolet, General, Nissan, Tesla


Leave a Reply

28 Comments on "The Rolling Start, A Better EV Performance Metric"

newest oldest most voted

From the article: “Ultimately, we should use many diverse criteria to compare cars, and acceleration performance is just one.”

True and when one type of vehicle consistently performs better when measuring all types of diverse criteria, whatever that means, then we must conclude that evs are simply better than ices.

Thanks Car and Driver for ‘spailnin that.

Great article!

This is a really good review. EV’s are so different from ICE cars that people can’t compare them.
This measurement will catch most of the differences between EVs and ICEs.

Just one thing is missing, if im suddenly in a hurry in my ICE car I have to spend seconds on changing gears down.
In an EV I will just press the pedal and without any delay increase the speed.
EVs are never in the wrong gear, this is a really big advantage!

I try to explain this to ICE drivers and they usually don’t get it, when you suddenly accelerate in an ICE all these things happen, the engine revs up, the torque converter slips, the tranny downshifts, the turbo spools up. In an EV the power is instant, no shifting no lag! It’s so different…

.. and that’s one reason why I do so much ‘splainin with my right foot, one stoplight at a time. 🙂

Great article! This is what I had in my blog in trying to examine why Vette with better power-weight ratio than P90DL is slower in 0-60. Ultimately, I came to the conclusion that ICE would have to destroy itself in order to be competitive.

When I harped on Bolt’s seemingly low performance in my blog, I did similar comparison as this article for 0-60 and 5-60 from car and driver for comparably priced compact cars like GTI, WRX, etc. Bolt is about middle in 5-60, but bragging rights not so much (ie, not the top like SparkEV is among sub $20K cars). General public won’t know that it’s 5-60 that really matter, instead only looking at 0-60.

In my mind, the only important acceleration metric for a regular street car is the 20-70 mph time. That typical max accel scenario for the real driver is accelerating as quickly as possible from a low surface-street speed to freeway speed on a short freeway on-ramp.

Better check Nürburgring Lap times if you are looking for something that has relationship to car performance. Street racing is for idiots.

Read the article, please.

Buying street cars outside of Germany based on Nürburgring Lap times is for pretentious idiots.

Haha couldn’t have said it better!

I had an great, great, uncle who went to one of their conventions once:

Robb Stark said:

“Buying street cars outside of Germany based on Nürburgring Lap times is for pretentious idiots.”

Well said, sir. Or perhaps even inside Germany.

I don’t want a car designed to perform well on the infamous “Green Hell” racetrack. I want one that will perform well for real-world driving on public roads, without costing me an arm and a leg.

Buying street cars based on real car performance is bullshit, noone drives races at the normal street. Measuring real car performace with green hell laptimes is usefull.

People who buy cars based on 0-60 numbers or lap times are all idiots. Autonomous driving will make them go the way of the dodo.

True, but most people drive their cars on the street. That makes the numbers more relevant for estimating around town performance and feel.

(Just as the article said.)

I am not sure the performance of something like the P90 is all that relevant to any practical driving situation. It is just FUN!

Impressive and persuasive analysis. Thanks !

OTOH, I will be taking my very recently upgraded to “Ludicrous” P85D(L) to the local drag strip tomorrow afternoon to see what it can do all the way out to around 11.4 seconds.

8 – )

Excellent point… one that I make to anyone who asks me about our i3s performance. The instant power, lack of downshifting, etc., makes the i3 significantly quicker around town than the ~7 second 0-60 indicates. Even 5-60 isn’t the i3’s wheelhouse, IMO, because it’s power limited at 0-15/20mph, but 20-50 is very quick.

Yes, 20-50 acceleration in the i3 is very addictive. If only they had a dual motor M-Sport i3, I would be all over it.

I hope you’re still feeling well after few days. Do you have torque curve for i3 BEV?

Interesting how even rolling times don’t reflect driving experience. We have both Fiat and Spark. While Sparkie’s #s look better, our off the line driving experience with the Fiat is much better–especially in the 5-30 range. Spark really performs best in the 40-70 range in straight line merging into freeway traffic, otherwise Fiat gets our nod for all around driving performance

0-30 MPH: 500e=2.8 sec, SparkEV=3.1 sec. In fact, I think 500e might be quicker than Bolt in 0-30. Only if Fiat came with DCFC and as low priced as SparkEV, it would’ve been a great car.

Isn’t a 0-30mph time more indicative of urban driving performance than 0-60mph times or even 5-60mph? A bit like the NFL combine looking at the 40-yd dash metric and wondering if they should change it to better reflect actual football playing potential.

Most people, if not all people, couldn’t tell you which car was quicker, if you sat them in the passenger seat and drove to 60mph, when the differences are a second or so. What they can tell is if they can feel the g-force of being pressed into their seats as the car launches. Perhaps, someone can just start using max g-force readings. Wouldn’t that be a better indicator of urban driving performance?