Basically there is no said amount that is being lost from the drivetrain, all cars vary, even cars of the same make and model will not necessarily loose the same amount.
There are also several areas where descipancy occurs.
1. Ford rate the engine in the GT at 260bhp SAE Net. However there are many types of bhp. But they don't all carry the same value, the same as $1 US dollar has a different moneytary value to $1 Australian dollar, yet they are both still dollars.
-PS
-HP
-BHP
-Imperial
-Metric
These all cause different numbers to be produced. So depending on WHICH scale of bhp a rolling road dyno is measuring will vary what the numbers are.
2. The standard of how bhp is rated SAE stands for the Society of Automotive Engineers and have many standards which concern temp, humidity and which ancillaries are on the engine (there are many other standards used across the world, which means cars form other countries with the same bhp claim may be producing more or less than an American car rated under SAE standards). As a rolling road dyno will be done at ambiant temps in a work shop or outside there is no control over other influencial factors which could cause an engine to produce more/less power. There are correction calculations which can be used, but are not always and they are not always used correctly when implemented.
All in all this will cause a variance in recorded numbers.
3. Accuracy of the rolling road dyno. As most of these will be used all day every day it is safe to assume that they are not 100% fully calibrated all of the time, because it would be expensive and time consuming to do so every day. If you except that any rolling road dyno pull could be +- 5-10% it makes it a lot easier to understand why cars produce such a wide variance of numbers.
4. How the power of the engine is measured. A rolling road dyno measure the rotational force at the rear wheels, so even tyre pressure and tread pattern can and do affect outputs. More importantly is how a measurment is acheived, torque is usually measured as a rotaional force against a counter weight this is then calculated to give a HP reading according to what engine rpms are. But there are other ways of measuring HP, originally I believe it was how fast a Horse could pull a weight of X amount vertically via a pulley over a set distance. Hence the name Horse Power.
In addition to these factors there are different claims on how power is lost via the drivetrain. If you increase the HP of car via a power adder like NOS you are not changig anything in the g/box, axels and so on. The drivetrain remains constant so should it suddenly be using more power to rotate the same components? Well yes to an extent because if there is a greater force acting on it then the amount of friction will also have incrased, but the 15% rule does not cater for this, e.g.
A car dyno's at 220rwhp if we then / .85 (for 15% loss) = 259bhp.
Drivetrain loss = 39bhp
If that same car now dyno'd 420rwhp. / .85 = 494bhp
Drivetrain loss = 74bhp
Is the drivetrain really taking almost double the HP the rotate it even with out any changes? I doubt it.
It should be requiring more power but the % loss should be less, for argumants sake say the 420rwhp car only lost 10% (420 / .90 = 467bhp), this would mean a 47bhp loss from the dirvetrain, much more realistic. Although it does mean your crank BHP is lower than you may are wished for. However wishing will not defeat physics any more than hope can.
And because of this there are quite a few engine builders out there that will claim any Mustang GT 4.6 with a manual gearbox will loose about 30-40bhp thru drive train loss, except it exceptional circumstances. So a rolling road dyno of 230rwhp corrected to SAE Net standards plus 30bhp for the drivetrain loss would funnily enough equal the magic figure of 260BHP.
I doubt that the GT's are underatted, there will always be a few that produce that bit more power, but this will be due to manufacturing tolorancies. All of the should be producing 260bhp if they where put on an engine dyno and rated as per the standards.
This also applies to the 03/4 Cobra's.
The Cobra's are one hell of a car and rank high on my list in my fantasy garage, but if they ALL really produced over 400BHP SAE Net I'm sure Ford would have advertised the fact, for two reasons:
1. The Cobra is Fords premier performance model in the US (forget the GT, but that didn't exist then and is a totally different league), so there was no higher end car to steal sales away from in the Ford group, unlike GM they deliberatly underated the Fbody's so as not to over shadow the Corvette with the same engine, it was of course a lot cheaper to purly'Advertise' a lower BHP than actually go to the expense of detuning the units or building specific spec units just for the Fbody. Very clever in all reality.
2. The Corvette Z06 had 405BHP SAE Net, and although slightly more expensive could well be considered as an alternative to the Cobra, so it would have been beneficial for Ford to advertise the highest BHP but they didn't they advertised less than the Vette which again would indicate that the engine did only produce 390BHP SAE Net.
Personally I dont' believe the Cobra's are underatted, or not by much anyhow because they just arn't fast enough to carry bigger statistics.
If tested on a like for like basis by the same driver under the same conditions an 03 Cobra will put down almost identical 0-60mph and 0-100mph times as a manual C5 Corvette.
The Vette has 345bhp SAE Net compared to the Cobra's 390bhp SAE Net but the Cobra weighs more.
The Vette weighs i at about 1456kg so that = 345 / 1.456 = 234bhp per metric tonne power to weight.
A 03 Cobra weighs in at about 1650kg, 390 / 1.650 = 236bhp per metric tonne power to weight ratio.
Almost identical, only 2hp in it!
The big thig the Cobra has is HOW it produces the power, as it is FI it produces more power and specifically torque than the Corvette for more of the time. LS1 is a great motor but doesn't have the low and even mid range grunt of engines of old. The blown 4.6 will be producing more power from low down.
So remeber PEAK power outputs are pretty meaningless, its the profile of the graphs that are important.
And only ever take dyno numbers with a pinch of salt because they are anything but accurate.