Think you're hardcore? Ha! Check out this guy!

[Zardoz]

The Ultimate Car Test:

http://boards.ign.com/Gran_Turismo/b5208/88887317/?20

 

[deadboy1975]

now that is HARDCORE

should be an interesting read once complete. sometime in 2007 !!

 

[vladimir]

so much work and then he will use laptimes that include spins or mistakes? well in my opinion that is just a huge waste of time. there is no point in comparing untidy laptimes...rubbish!


a test with a collection of 200, untuned, N-tyred and tidily lapped cars would make much more sense.

 

[Mr.Slow]

a test with a collection of 200, untuned, N-tyred and tidily lapped cars would make much more sense.

Get going then!

 

[RenesisEvo]

so much work and then he will use laptimes that include spins or mistakes? well in my opinion that is just a huge waste of time. there is no point in comparing untidy laptimes...rubbish!


a test with a collection of 200, untuned, N-tyred and tidily lapped cars would make much more sense.

And the fact that they are tuned as well.... I agree - untuned, N tyres (but either N1 or N2 or N3, no mixing etc). Best way to ensure consistent laps time? A B-spec driver with 10,000/100/100/100 - THAT will take some hardcore work to get! (there is a thread somewhere that tells you how to do it)

 

[DQuaN]

So he is gonna do the motor carriage as well?

 

[vladimir]

So he is gonna do the motor carriage as well?

no.

Get going then!

i am already, but don't expect it to be ready before 2010.

 

[bigjoe]

Its not a big deal, although doing them all at once is. I'm doing it as I play the game. Every car I win or buy gets a lap or two at the Ring in Arcade Time Trial. Adds less than 20 minutes to each acquisition, so far I've done 78 cars ranging from 17 minutes + for the 2CV down to 5:29 for the F1 car.

Only difference I do either 1 or 2 laps, 1 if I get a clean lap or 2 if I really mess up the first lap. If I get an almost clean lap I might adjust the time by a second or two depending on the severity of the screw up. If after 2 laps I still don't have a clean lap I chalk it up to the car being a bad handling car and assume it would always be that way.

All the cars are stock out of the box, with default tires. I also don't change transmissions ratios so Rally cars suffer.

BTW I didn't start this until I passed M34 and got gold in both Ring tests. That eliminates some of the learning curve and, hopefully, helps get more even laps.

 

[Dave A]

If you get a clean first lap then a do the second lap just as clean the second lap will be faster by maybe 3 or 4 seconds, you're going a lot slower when you go over the line from the rolling start as opposed to coming round the last corner and putting you're foot down.

 

[Dan Watkins]

I am doing something similar, I take a car, do some tests, i.e. standing 1/4, top speed, and then hotlaps around Tskuba and the ring, clear laps I hasten to add , then using a formula I get a final figure, this figure relates to how good the car is overall.

This is how it looks so far.

Dan.

 

[Dave A]

I'm presuming the lower the final score is, the better the car. And if you don't mind me asking, what formula do you use to calculate that score.

 

[KaffeinE 86]

Watkins, are you A or B-Specing that? Cuz the those cars can get better times.

 

[Dan Watkins]

Yeah, the lower the score is the better the car is, is "should" work out that the slowest car in GT4 should be around 1000, and the fastest should be around 1.

It may or may not work out right, but it seems to be giving a fair representation.

As for the formula, wouldnt make a lot of sense if I showed it to you, its in excel format and is just a great long equation.

Dan.

 

[Dan Watkins]

Watkins, are you A or B-Specing that? Cuz the those cars can get better times.

A-specing, and I dont doubt that people can better my times, I am not the greatest driver in the world, but thats the beauty of my system, regardless of my driving skill the ratings still reflect the overall rating of the car, as I am driving each of the cars in question in the same manner.

Thats not to say the ratings wouldnt be different for someone else as they may be better as RWD cars then me but worse at FWD cars, but my list will give you are reasonable idea how each car stacks up against another.

Dan.

 

[zonekill]

@ Watkins: How many laps did you drive to get those Tsukuba and Nordschleife times??

The Mitsu FTO and the Lancia Delta are in the wrong order and also the Isuzu bellet and VW golf, but that are details

 

[Bones Brigade]

im goign to do the same with my cars on the high speed ring.

 

[Skant]

You must really like the Vanquish. You're getting much better times with it than cars that are nominally a good bit faster.

As you say though, your list makes a very good point. It's truly difficult to compare the performance of cars because it's the combination of man and machine that makes lap times.. not the machine alone.

This is why nobody should take magazine reviews on cars too seriously. Often times, they will hate on a car because it just didn't click with them even though it's a very good car. The particular thing I always find the most amusing is when sport compact magazines do performance tests on sport compacts versus real sports cars... and usually decide that the sport compacts are way better. And yet, it's pretty clear just reading what they've written that these guys were very slow when they tried the sports cars because they don't know how to handle a RWD car on the track!

Making a truly fair test is basically impossible.

- Skant

 

[Mr.Slow]

You must really like the Vanquish. You're getting much better times with it than cars that are nominally a good bit faster.

As you say though, your list makes a very good point. It's truly difficult to compare the performance of cars because it's the combination of man and machine that makes lap times.. not the machine alone.

This is why nobody should take magazine reviews on cars too seriously. Often times, they will hate on a car because it just didn't click with them even though it's a very good car. The particular thing I always find the most amusing is when sport compact magazines do performance tests on sport compacts versus real sports cars... and usually decide that the sport compacts are way better. And yet, it's pretty clear just reading what they've written that these guys were very slow when they tried the sports cars because they don't know how to handle a RWD car on the track!

Making a truly fair test is basically impossible.

- Skant

Amen brother!

Out of dozens of car reviews on the Honda Insight I have only seen ONE where the guy could actually achieve the published MPG readings. Yet I owned one for four years and it was a the exception when I didn't get the published MPG readings or better...

 

[jdw]

Making a truly fair test is basically impossible.

For a human being, you're probably right.

But for the computer it should work. You could conceivably B-spec each car from the same B-spec level and get the kind of consistency necessary to consider it at least pseudo-scientific. This wouldn't help so much in picking a car YOU should drive (that's what Dan's list does) but could be used to compare your competition.

 

[Skant]

For a human being, you're probably right.

But for the computer it should work. You could conceivably B-spec each car from the same B-spec level and get the kind of consistency necessary to consider it at least pseudo-scientific. This wouldn't help so much in picking a car YOU should drive (that's what Dan's list does) but could be used to compare your competition.

While that may seem like it would work, in fact, B-spec Bob has a driving personality much like any human does. There are lots of really great cars that he really sucks at driving. Ultimately, his computerized nature just means that he'll produce identical performance from one run to the next given the same car and same track. But when you change either of those, his new performance becomes highly variable. In a lot of ways, even more so than a human because he doesn't learn and he doesn't adjust.

You can put Bob in a car, get a lap time, increase the horsepower on that car, get the lap time again... and sometimes the new lap time will actually be slower because Bob is consistently overcooking some corners with his new horsepower, and he just doesn't learn. And yet, clearly the car he's driving is faster now.

- Skant

 

[jdw]

While that may seem like it would work, in fact, B-spec Bob has a driving personality much like any human does. There are lots of really great cars that he really sucks at driving. Ultimately, his computerized nature just means that he'll produce identical performance from one run to the next given the same car and same track. But when you change either of those, his new performance becomes highly variable. In a lot of ways, even more so than a human because he doesn't learn and he doesn't adjust.

You can put Bob in a car, get a lap time, increase the horsepower on that car, get the lap time again... and sometimes the new lap time will actually be slower because Bob is consistently overcooking some corners with his new horsepower, and he just doesn't learn. And yet, clearly the car he's driving is faster now.

- Skant

Yes, but it's consistent. And that's the key factor.

That is exactly the point I was trying to make. B-Spec Bob will not help pick a car for YOU to drive but, by observing that same consistency, it's easy to determine the relative competition of a given lineup of cars.

 

[NoQuarter]

To really be subjective in A-Spec, one needs to a consistant driver (and I have no doubt you are Dan ) and attempt to drive each drivetrain catagory as similarly as possible. Also, to average out any driving inconsistancies, the test driver should drive a minimum of 5 laps, throw out the fastest and slowest laps, and take an average of the remaining 3 laps. It would also help if the test driver is intimately familiar with the test tracks used. That would mean having a lot of seat time at each test track prior to testing.

I would imagine any of the regular WRS or OLR drivers either here, or at any of the other Premium GT forums would be vastly superior test pilots to miniguy, whose testing methodology appears dubious at best, and whose only credentials seems to be nearly 2,000 posts at IGN.

 

[Zardoz]

...There are lots of really great cars that he really sucks at driving...

In my Le Mans car test there was one (the Gillet) he couldn't keep on the track even at the "1" speed setting, others that he could only handle at the "2" setting, one that he couldn't drive nearly as fast as he should have (the R92CP), one that he drove faster on "3" than "4" even though he never went off the road with it (the '03 Pescarolo), and only one that he could drive perfectly at the "4" setting (the Audi R8).

Bob was all over the map with just the Le Mans cars, so imagine where he would be with all 720 of them.


.

 

[Speedy Samurai]

I am doing the same thing. I have an excel spreadsheet listing the car, make, power, drivetrain, weight, 0-400 time, 0-1000 time, max speed. 0-100kmh time etc. I then take each car for a hotlap around montegi east. So far the Zonda has been the fastest. I have them in categories such as supercar, sportscar, etc. Every race car I test gets taking around Le Sarthe instead of Montegi.

 

[jdw]

Ya know... it occurs to me that, if we could come to terms with creating some relative kind of consistency (e.g. no settings tinkering, start with the base car and maybe one specific mod platform - get lap times on say three different courses with different characteristics + 0-# times) and create a format we could agree on. If we could get say... 50 people to do 10 cars, or 25 to do 20 cars we'd have the majority of the cars in the game all mapped out.

Put it in an Excel format similar Lone Ranger's Excel spreadsheet complete with picture, perhaps combine it with some of the excellent data provided by Famine in summary format similar to what Dan posted and we could, collectively create, THE ULTIMATE CAR DATABASE!

What better resource to guide people to choose cars to try out?

Obviously, there are severe limitations. The biggest is the variance in driver skill and dealing with handling on particular cars not to mention particular tracks. To help neutralize some of the variance, you'd ALSO perform B-Speccing to see how the computer handles it.

How does that help neutralize variance? Well, given a static B-spec level (yes, I know, also hard to achieve), the relative difference on a given track between the B-spec time (which should be consistent) and the user time gives an indication of how good that driver was. Now if you drive that track with the same car, note where in that B-spec to human driver range you fit, it will give you a *feeling* for how you are compared to the driver that recorded the time. You do a few of these, and it may help you to pick other cars.


Well it's a thought. 💡

 

[NSX-R]

Respectable, and at the same time, sad...

 

[MasterGT]

All the cars are stock out of the box, with default tires. I also don't change transmissions ratios so Rally cars suffer.

Don't all cars have a similar, but not always the same (depending on the car...), basic default setting?

Some cars would (and do) handle great with the default settings, while others wouldn't (and don't). The cars that don't handle great can nearly always be adjusted so that they do handle better, so a test like this means absolutely nothing, to me.

Cheers,

MasterGT