All cars driven - 10 miles of testing data.

[Car-less]

With GT5 release day just a couple of hours away, I've decided its time to reveal my latest large project in GT4.

It all started with Famines GT4 Complete Arcade Car Testing List. I thought, if B-spec Bob gets to drive every car, then why can't I?

So after a few test drives, I chose myself three tracks to test different aspects of each car.
1) Trial Mountain, as used by Famine is a good test of handling on uneven surfaces.
2) Seoul Central (yawn) is a good test for acceleration, braking and city driving.
3) Tsukuba is a good test of cornering on a flat racetrack.

I went into Arcade Mode, Time Trial and proceeded to drive every car* in the game for three laps at all three of these tracks, on default tyres (Racing:Hard for race cars, Sport:Medium for everything else) and without any driving aids.
Thats right. 726 cars*, 3 laps, 3 tracks, no aids, all in A-spec.

Thats not all. As I was nearing the end of this task, -Mr.M- revealed his similar project WHICH IS THE FASTEST? All cars tested over 6 tracks! where he had tested most of the cars in B-spec.

With both Famine and myself having used Trial Mountain, and -Mr.M- and myself having used Tsukuba, I just had to include his Tsukuba times. A little number crunching later and I discovered that the sum of one lap at each was 9.17 miles. Motorland just happens to be 0.83 miles in length. You can guess what I did next.

I sent my B-spec driver round Motorland at a gentle pace (you don't want to see him pushing it here) in every car*.

So now I have a spreadsheet showing every car* and their fastest laps in all 6 tests.
1) Trial Mtn-B as driven by Famines B-spec driver.
2) Trial Mtn as driven by myself in A-spec.
3) Seoul Central as driven by myself in A-spec.
4) Tsukuba as driven by myself in A-spec.
5) Tsukuba-B as driven by -Mr.M-s B-spec driver.
6) Motorland-B as driven by my B-spec driver.
There are then columns for;
7) the total time taken to drive all six laps (10.0 miles), and
8) ranking the cars according to their times.

What can you do with this data? Well you can compare the performance of different cars, matching them up to create fun races for head-to-head or LAN races.
You can compare my times to the B-spec times at two tracks to discover which cars Bob is good at driving, and which he is not so good in.
You can crunch the numbers and draw a graph showing the distribution of different times, which I have done, and included on a seperate sheet.

For those unable, or not wanting to download the .xls file, here's a snippet of the data showing the results for all 'special' cars that cannot normally be raced.

[U][B]Country	Make	  Model			    Type     Year  Drive  Power	Weight	Pwr/Wt	Trial Mtn-B Trial Mtn	Seoul	    Tsukuba	Tsukuba-B   Motorland-B	Total	    Ranking[/B][/U]
USA	Nike	  One			    Special  2022  4WD	  259	760	0.341	1'34.416    1'31.008	1'00.869    0'57.936	1'00.620    0'46.882	6'51.731    113
Japan	Toyota	  Motor Triathlon Race Car  Special  2004  ---	  266	1200	0.222	1'40.972    1'36.435	1'04.054    1'02.012	1'03.506    0'48.469	7'15.448    180
UK	Caterham  Seven Fire Blade	    Special  2002  FR	  160	369	0.434	1'46.136    1'38.839	1'07.200    1'02.570	1'05.180    0'48.731	7'28.656    265
USA	Chrysler  Prowler		    Special  2002  FR	  254	1287	0.197	1'48.200    1'43.077	1'09.609    1'04.774	1'08.119    0'50.585	7'44.364    384
USA	Jay Leno  Tank Car		    Special  2003  FR	  901	3500	0.257	1'57.037    1'51.848	1'09.292    1'08.204	1'14.312    0'54.141	8'14.834    572
USA	Ford	  Model T Tourer	    Special  1915  FR	   20	635	0.031	2'55.645    2'49.755	2'00.215    1'40.709	1'43.556    1'06.498   12'16.378    723

*The two 1886 Benz's and the Audi Auto Union have been excluded due to their extreme limitations preventing them from completing any laps.

I'm sure I have forgotten something, but it's getting late, so any questions, just ask.

 

[PF]

Wow, that's an impressive data set. Thanks very much for sharing this, it'll be extremely useful and also fun to play around with graphs and the sort function.

 

[AlexGTV]

I would have chosen the ultimate 12.93 miles of driving.
Anyhow, this is great stuff! 👍

 

[Blitz24]

Great list. I'm still sorting through it but its very good work.

 

[Austin343]

That's quite an achievement, Car-less. How did you decide in what order to proceed with testing? I imagine that it would have been quite difficult to summon the motivation if, for example, you had to drive all of the Skylines consecutively. Also, I assume that you tested each car at all tracks before moving onto the next car, rather than testing every car on one track at a time, in order to reduce any "learning" that might color the results (although it's probably also true that you didn't mean this to be at all scientific in the first place).

 

[granturismite]

i really really want to read this, but I can't un-zip the file. windows says the file is invalid, and that there is no content within the folder. Help??

 

[GT_Prologue5]

Wow. How long did that take?

 

[Mike1027]

Nice, how many century's did that take!👍 amazing good job!👍

 

[Blitz24]

i really really want to read this, but I can't un-zip the file. windows says the file is invalid, and that there is no content within the folder. Help??

I could repost the file unzipped if need be. Or just post it directly on here, but I need Car-less's permission.

 

[Famine]

Download 7zip instead.

Trust me - we're well experienced with this issue on GTP

 

[Car-less]

I would have chosen the ultimate 12.93 miles of driving.
Anyhow, this is great stuff! 👍

As much as I'd like to drive them all at the 'Ring, it would have taken so much longer, and the data would be less reliable.
With the shorter tracks it was easy enough to just do an extra lap to make up for any major mistakes.

That's quite an achievement, Car-less. How did you decide in what order to proceed with testing? I imagine that it would have been quite difficult to summon the motivation if, for example, you had to drive all of the Skylines consecutively. Also, I assume that you tested each car at all tracks before moving onto the next car, rather than testing every car on one track at a time, in order to reduce any "learning" that might color the results (although it's probably also true that you didn't mean this to be at all scientific in the first place).

The testing order at first glance would appear totally random, though there were in fact multiple patterns employed. At times I'd run a series of cars previously used at SFGTP or UKGTP events, usually beginning with the slowest. Other times I would choose one manufacturer and run through the cars in the order they appear. Sometimes I'd choose a bodystyle (i.e stationwagon) and run them alphabetically, etc. I'd go where the mood took me, though usually I'd run 5-10 similar cars in one session.

Once I had chosen a car, I'd run all three tracks before moving onto the next car, though some "learning" was definately observed. Sometimes this caused me to revisit earlier cars, or previous tracks in the same car just to check there was no major improvement in time. Most of the times are repeatable to within 0.5 seconds any time I wish to repeat a session.

Wow. How long did that take?

Nice, how many century's did that take!👍 amazing good job!👍

About 18 months to 2 years on and off. Somedays I'd run 30 cars, sometimes I'd go 30 days without running a single car.

i really really want to read this, but I can't un-zip the file. windows says the file is invalid, and that there is no content within the folder. Help??

Download 7zip instead.

Trust me - we're well experienced with this issue on GTP

What Famine said. I ran into issues with his datafile (and others I found on the forums) using Winzip, so now use 7zip without any problems.

 

[seanyb505]

I embarked on a similar task, but with severely less cars (no boring cars or duplicates). Somewhere around 360 cars, grouped into like classes. My goal was to test all the cars, then tune them to the max and compare. I used the test course and midfield. I drove each about 20-some miles. I never finished thought. What I found is that after playing the same course over two-three years is that my racing line changed drastically. I raced cars 1-3xx, then started over with tuned versions of 1-3xx. While I was racing a much faster 53 Vette, I was also much faster. It sure was fun.

 

[lan88]

Erm... whats the extreme limitation of the Audi V16?

PS nice job!

 

[Car-less]

It's extremely limited in that it can only be used in Free run or photo drive at Nurburgring, or test course or Las Vegas in power and speed section of GT mode. It cannot be driven in arcade mode at all.

Edit: Actually, it can be accessed by going into 2 player battle, then chosing it from your garage on your memory card. It's not quite the same, but I'll run some tests tommorrow and decide if it's worth adding those times.

Edit2: Unfortunately you cannot use B-spec in 2 player battle, so thats half the times unable to be recorded. On top of that, this car is very difficult to get a clean lap out of without tiptoeing around the track, especially when using only half the screen, so any time I set myself will be unrepresentative.

 

[granturismite]

thanks for the troubleshooting tips. works fine now. I'm working on getting all cars, but to test each one? that must have taken years

 

[granturismite]

just a quick question. what version are you running (NTSC, PAL, etc)?

 

[Car-less]

I'll give you a clue. Look at my signature.