This article was first published in 2003.
The use of a peak power figure as short-hand for measuring vehicle
performance has always had a dubious rationale - and it's getting worse. In
fact, it's nearly as bad to suggest that a car with 150kW is going to cane one
with 130kW as something else that used to be very common, at least in this
country. So what was that, then?
A long time ago, when I was a young lad, it was a playground norm to discuss
how powerful each family's car was solely on the basis of engine size. In my
primary school days it was still measured in cubic inches, so we'd scornfully
say something like: "Brett's dad has only got a one-eight-six,"
while Brett writhed in shame and tried to retrieve the loyalty of his friends by
giving away newly-bought lollies.
"My dad's got a two-oh-two," I'd say proudly, basking in the glow of
widespread approbation until someone fiercely trumpeted, "Well, we've got
At which we'd all know that we'd been beaten and decide that we had better go
off and raid the girls' toilets... or something.
(I well remember spying a long wheelbase Jaguar V12 and being widely excited
by it - I thought 'XJ12L' stood for 12 litres of engine capacity....)
If all engines revved only to the same rpm, and if they all had
similar volumetric and frictional efficiencies, then basing power estimates only
on swept capacity would make a lot of sense. And in the Seventies in Australia,
the Big Three - Holden, Ford and Chrysler - mostly did have engines that
conformed to these three factors. So yes, a three-five-one did have a
helluva lot more power than a puny one-eight-six.
But it wasn't that many years later that I bought a car - a 1977 BMW 3.0si -
that was only a one-eight-three when measured in cubic inches. But it had
similar power to my erstwhile mate's three-five-one....
Forget the Capacity
These days, apportioning engine power on the basis of engine size is known by
nearly everyone to be a waste of time: a 2-litre engine might have 85, 94, 96,
101, 105, 110, 121 - or 147kW... and yes, they are all naturally aspirated
figures. But just as we discarded engine swept volume as the vital parameter of
performance, we now also need to get used to largely discarding the peak power
Simply, it's become increasingly insignificant.
The logic behind quoting just peak power as an indicator of performance in
similar-weighted cars is predicated on the past, where we've were all making
some implicit assumptions. Just like it underpinned my primary school playground
discussions (who then had heard of a standard one-eight-six that revved with
torque to 6500 rpm?!), in the more recent past we've automatically assumed that
an engine with (say) 100kW peak power at 6000 rpm would develop something like
50kW at 3000 rpm, and 30kW at 2000 rpm.
In effect, without even thinking of it, we could judge the rest of the power
curve from just the peak figure.
But that is no longer the case. Especially in turbocharged engines - but also
in variably valve timed designs - it's just not enough to say that a 1300kg car
has about 140kW. Literally, two similarly weighted cars, both with around 140kW,
could have a major variation in their on-road performance times.
Shape of the Torque Curve
So what's the answer? - how do we gain effective shorthand descriptors of
engine power and tie that to performance? Basically, you can't do it on engine
power alone. These days, you simply have to describe at least some
aspects of the torque curve. (If I shortly lose you, go read "Power versus Torque - Part 1" and "Power versus Torque - Part 2" - they're still widely acknowledged as the best discussion on
the web of the importance of the two measurements.)
But if you read those stories and get confused as to why the shape of the
torque curve is so important to my forthcoming discussion, just remember the
simple concept: torque x revs = power. That is, if you have more torque
anywhere in the rev range, at those revs you have more power. And
hell, if you have more power when you put your foot down, you're going to go
That's also one reason why a 0-100 km/h time is one way is a bit deceptive -
after first gear, you're never down low again in the rev range. Most cars will
take two gear changes get to 100 km/h (about 60 mph), and shortly after each
gear-change, the engine revs will be right up the top end again... that is, near
peak power. So the 0-100 time shows nothing about immediate response when you're
caught in the wrong gear, or even the performance of a car with different
gearing - it indicates only what you can gain when the car is launched hard and
pushed pedal to the metal until the magic '100' number appears on the
The example is best shown by an extreme. These days European turbo engines
are simply mind-bogglingly torquey... that is, they develop an extraordinary
amount of power at low revs, considering their capacity. Let's look at some
concrete examples. The Holden (Opel) Astra Turbo has a 2-litre engine that
develops a peak power of 147kW at 5600 rpm. Its power and torque graph is shown to the left. (For the knowledgeable, the low rpm
at which nearly 150kW is being developed from only 2 litres gives an immediate
indication that this engine must be producing a lot of torque for its size!)
At just 2000 rpm, it has 52kW available. (We'll keep all the discussion in kW
and rpm because people can get sidetracked when Nm are brought in...)
So, is 52kW at 2000 rpm a lot? (It is indicative of how we never look at the
spread of power that even the enthusiasts amongst you won't know the answer to
that!) Well, let's have a look at the 2.2-litre naturally aspirated engine from
the same Opel engineering stable. At 2000 rpm the NA engine - and yep, it's 10
per cent bigger - has only 38kW available. That's a stunning 27 per cent less
power at 2000 rpm! So how hard do you think that the 2-litre turbo will be
accelerating at 2000 rpm compared with the 2.2-litre NA engine?
(Or take the current 155kW Saab 9-3 Aero. The four cylinder engine is also
only 2 litres in capacity, but at 2000 rpm there's 58kW available - and by 2500
Now let's compare that with another 2-litre engine of similar max power
output. The 2002 Honda Integra Type R develops a peak power of 147kW at 7500 rpm
- exactly the same power as the Astra Turbo. (The Honda's power and torque graph is shown to the right.) So, same engine size, same peak
power output. But remember how by 2000 rpm the Astra had 52kW available? Well,
at those same revs, the Integra's engine can develop only 35kW... that's 33 per
Sure, different gearing can take some of this into account (the final drive
ratios of both cars are enough to see this: 4.764 for the Integra and 3.63 for
the Astra - and the Honda has a 6-speed) but it's now true that saying
something like: "It's a 2-litre that makes just under 150kW" doesn't really tell
you much about how the car will drive, how responsive it will be when caught in
the wrong gear - or how much gear changing will need to be done, for that
In fact a car with a 0-100 time of 8 seconds - but with lots of power
available through the rev range - will always feel like a car that previously
would have been a flat 7-second car. By that I mean that the on-road grunt in
road situations will be higher - often much higher - than you'd expect
from the bare 0-100 time.
Forget the Shorthand
To make an assessment of the sort of on-road performance that you're likely
to gain, you really need to see a power/torque graph, or have available a lot
more performance figures than just a 0-100 time. For example, if you have the
in-gear acceleration times in different gears, you can quickly get an idea for
how fast the car will be when overtaking, accelerating to fill a hole in
traffic, and so on.
You can also measure 'average power' across a rev range - some engine and
chassis dynos can currently do this. But over what rev range do you take the
average? The Integra drivers will be suggesting that they never get below 3000
rpm (which of course can't be true unless they always do a big clutch dump
launch!) while many others would suggest that the average power in many engines
would be most importantly measured across a 2000-5000 rpm range.
But neither approach is shorthand.
Simply, these days it's impossible to accurately predict on-road performance
(and I mean a lot more than a bare 0-100 km/h time) from the peak power figure.
Just as we now no longer use engine size as shorthand for performance, using
just a peak power figure is also becoming increasingly invalid.
And, incidentally, that applies in spades for modified cars, too.