Gran Turismo 4 A-spec Calculator

[shotamagee]

I'm very slowly and meticulously chasing GT4 A-spec points. I had been using Car-less and Wild Cobra Z28's information separately to try to identify which cars are the best for this. So I thought it might be a good idea to try to combine them somehow. I imported them into excel, and began playing around with the graphs and formulas which eventually lead to what you see now, the;

Gran Turismo 4 A-spec Calculator.

It features 3 worksheets
1 - The A-spec Calculator
2 - The testing data, which shows all the cars and each ones A-spec advantage
3 - The mathematical curves, which show what the values are based on.

While the mathematical work is my own, the data upon which it based has used the following GTPer's information;
1 Car-less's All cars driven 10miles of testing spreadsheet (As the basis for calculating how fast each car is)
2 Wild Cobra Z28's, A-Spec Point Data; Values, Races, and Modifications, data in the GT4 200 A-spec point forum. (For gathering each car's A-spec points value)
3 Famine's GT4 Complete Arcade Car Testing List (For the calculator drop down, full car names list)

I have created this thread in the general GT4 forum, as the 200 point sub forum, says not to create any new threads, but it might be more relevant there.

To view it you'll need a program capable of opening MS Excel spreadsheets.

Anyway have a look, I'm sure it will be useful to anyone (if there are any) who are still on the GT4 A-spec Points chase, it certainly is for me (hence the upload). And for everyone else, you might find the info interesting if nothing else.

Cheers Shotamagee

EDIT: I forgot to add that there are comments at the top of all the excel columns which hopefully go some way towards explaining what they each mean. Also a further explanation of what the main Column (S) is, and what it represents is in my reply to Car-less in post no4.

 

[BobK]

Downloaded. I haven't opened it yet but this sounds like a great tool. Thanks!

 

[Car-less]

Interesting. I have downloaded and examined it and it kind of makes sense - though I have never quite understood Wild Cobra's data!

I have tried inputing various cars and opponents, getting reasonable results for sensible data, and crazy results from ludicrous data - all as expected.



Looking at the graphs, I am happy that my data fits so well with the power curve and can understand why those farthest from the curve are so far off;
1) Top Left (1) - Formula GT. Lightweight with high downforce. Exceptional performance.
2) Inside Centre (111) - TVR Speed 12. Crazy power with little grip. Hard to tame this beast.
3) Outside Centre (492) - Dodge Ram. Well known for rewarding disproportionately high points.
4) Inside, Far Right (712) - Subaru 360. Slowest of the slow, and I'm not the only one to think so. It lagged way behind in the "Grand Prix Classics" and just doesn't seem to have the power it should have.
Everything else seems to be a very good fit to the curve.



Finally I examined the data table. It's useful that you put explanations into each column heading but some of that information might be better off in this thread, especially for columns S and T.

As far as I can gather;
Column S shows how much easier a car is to drive than the points awarded would suggest.
Column T shows the A-spec points advantage you would have over similarly performing cars.

So in a family cup race with a difficulty of 0;
A high S value should make for an easy race and a very low S value would make for a difficult race.
A high T value would award above average A-spec points, and a low T value would offer fewer points.



I then adjusted the colouring of these two columns;
Bright Blue: Above 180
Bright Green: 120 to 180
Lime Green: 60 to 120
Light Yellow: 0 to 60
Light Orange: 0 to -60
Light Pink: -60 to -120
Bright Pink: -120 to -180
Bright Red: Below -180

This colour gradient goes from green (good to go) through to red (stop and choose something better) with brighter colours for the extreme ends, and paler ones for average figures. It also makes it easy to rate the cars between -180 and 180 for a full 360 degrees .

It immediately becomes apparent that very few cars fall outside the circle;
Dodge Ram - very easy to drive and very high points as we all know.
Mitsubishi Lancer 1600 GSR Rally Car - very high points.
Plymouth Cuda 440 Six Pack - very low points.
Gillet Vertigo / Toyota 7 / Nissan Option Stream Z / TVR Speed 12 - very difficult to drive.

The Cuda and Lancer surprise me, but the others do not.

There are then around 10 cars in each of the next high and low points and difficult to drive categories, but just two easy to drive. Highlights from these sections include:
Formula GT / Honda S2000 LM - Easilly beat comparable cars.
Cizeta V16T / Subaru 360 - difficult to win against similar spec cars.
Suzuki GSX-R/4 / Lotus Elise 111R - score highly amongst their peers.
Audi RS4 / Jensen Interceptor - reward you poorly for your efforts.

I think my examples fit my interpretation of these columns, but a few races are called for to fully convince me I've got it right.

Nice job nonetheless. I just hope you're not the only one to get some use out of it.

 

[shotamagee]

Looking at the graphs, I am happy that my data fits so well with the power curve and can understand why those farthest from the curve are so far off;
1) Top Left (1) - Formula GT. Lightweight with high downforce. Exceptional performance.
2) Inside Centre (111) - TVR Speed 12. Crazy power with little grip. Hard to tame this beast.
3) Outside Centre (492) - Dodge Ram. Well known for rewarding disproportionately high points.
4) Inside, Far Right (712) - Subaru 360. Slowest of the slow, and I'm not the only one to think so. It lagged way behind in the "Grand Prix Classics" and just doesn't seem to have the power it should have.
Everything else seems to be a very good fit to the curve.

It's not a case of your data fitting the curve, it more a case of finding a curve which best fits your data. And when something is off, like for example the Dodge Ram, that just means that PD's base value for that car isn't very accurate.
Also if every car fit the curve perfectly, all this would be redundant, it would mean that all the cars would be just as capable as one another of gaining A-spec points. But because they don't, it means there are outliers (you can actually see which dot on the graph's the Dodge Ram is, it sticks out so far), which are the cars which are useful for getting higher A-spec point race wins.

Finally I examined the data table. It's useful that you put explanations into each column heading but some of that information might be better off in this thread, especially for columns S and T.

As far as I can gather;
Column S shows how much easier a car is to drive than the points awarded would suggest.
Column T shows the A-spec points advantage you would have over similarly performing cars.

Fair point;
Column S is essentailly the main part of this calculator.
I had originally only done the column T calculations, but they only take into consideration a cars position relative to its A-spec rank, so while 2 cars might be next to each other on the speed rank there might be 10 seconds between how fast they actually are, (and opposingly, in the middle of the table where the times are very close for a lot of cars, which means some of those car's values can also be quite far off as well) which means the values in Column T can be less accurate.
So I decided to try to use the total time instead of rank as the basis for comparison, and thats when I started on developing Column S. (The Graphs should demonstrate how the values (due to the formula I found) in this Column were calculated.)

So Column S should represents how many "free" A-spec points a car has, but not necessarily how easy or difficult it is to drive. (That should be factored in to its speed/time) For example often twitchy MR cars like the Lancia Stratos are good for high A-spec points, but are also among the toughest to drive.

Keep in mind its based on 60 being the middle. So in theory an AI difficulty @ 0 race with a car thats approx = 0 (in Column S) for example the Lotus Esprit Sport 350, should get a 60 point a-spec race, and that will be a fair repesentation of how difficult that race will be. But the Amuse S2000 R1 which is at 60 (in Column S) when entered in a AI @ 0 race should be offered 120 points, but the difficulty should actually be similar to the previous (Lotus Esprit) 60 A-spec points race.

Column T has been made a little redundant by that info, but I've still included it for reference as much as anything. It just shows how many places the cars A-spec position rank is away from its actual speed/time rank. It can't really be used to calculate points.

So in a family cup race with a difficulty of 0;
A high S value should make for an easy race and a very low S value would make for a difficult race.
A high T value would award above average A-spec points, and a low T value would offer fewer points.

This depends on the other cars in the race. But essentially yes, for example based on its speed, my info shows in column S that the Amuse S2000 R1 is undervalued by 60 A-spec points, so if you use it, GT4 will give you opponents that are on average 60 A-spec points slower than you, but GT4 doesn't think they're slower, so you still get 60 A-spec points for the race, but it should be a pretty easy race to win. (The actual race difficulty should be pretty close to 1 I think, I'm not sure if race points between 0-60 are linear)
For column T the values don't mean anything in GT4, but higher values do indicate which cars are more likely to be useful for trying to win higher A-spec point races.

I then adjusted the colouring of these two columns;
Bright Blue: Above 180
Bright Green: 120 to 180
Lime Green: 60 to 120
Light Yellow: 0 to 60
Light Orange: 0 to -60
Light Pink: -60 to -120
Bright Pink: -120 to -180
Bright Red: Below -180

This colour gradient goes from green (good to go) through to red (stop and choose something better) with brighter colours for the extreme ends, and paler ones for average figures. It also makes it easy to rate the cars between -180 and 180 for a full 360 degrees .

I used conditional formatting, (so excel auto colours each box) but it only has 3 (plus blank makes 4) data colour ranges to use, so I just concentrated on highlighting cars which will be more, rather than less useful for collecting points.

It immediately becomes apparent that very few cars fall outside the circle;
Dodge Ram - very easy to drive and very high points as we all know.
Mitsubishi Lancer 1600 GSR Rally Car - very high points.
Plymouth Cuda 440 Six Pack - very low points.
Gillet Vertigo / Toyota 7 / Nissan Option Stream Z / TVR Speed 12 - very difficult to drive.

This shows that PD was off in calculating their base points more than anything.
The race cars seem to do better than the road cars, but I think this is partly due to the fact that the road cars are mainly on S2's, while the race cars have R1's as their primary tyres, and higher grip tyres are undervalued in the GT4 A-spec points calculation.
Also the track selection you used, has a lot of tighter tracks, with not so many long straights, which exaggerates how "handicapped" cars like the TVR Speed 12 actually are.
For lower value cars, again it doesn't neccesarily mean they're hard to drive, its just that their A-spec points indicate that they should be quicker than they are. (Admittedly this is often due to them being difficult to drive) But my case in point is most FWD's, they generally are not difficult to drive, but for A-specing their not often useful becasue PD didn't give enough allowance for their drivetrain handicap when calculating what their base A-spec point value should be.

I hope that explains things a little better.

Thats my first multi quote, jeez it took a while to do properly.

 

[Wild Cobra Z28]

Glad someone put my data to use. Hard to believe I spent so much time compiling it, and you spent more time using it.

Find any faults in my data? After all, I'm not perfect. Probably some errors even though I tried and tried to eliminate all.

 

[Wild Cobra Z28]

The Honda Element is also a sweet vehicle to get 200 points with, especially on the dirt and snow.

 

[shotamagee]

Glad someone put my data to use. Hard to believe I spent so much time compiling it, and you spent more time using it.

Find any faults in my data? After all, I'm not perfect. Probably some errors even though I tried and tried to eliminate all.

Short of PD giving out the actual values, I'd say what you came up with was about as accurate as possible. It's definitely accurate enough to be the basis for this calculator. The only way to check the values is to enter an arcade race (that way the cars are stock and you know what tyres they're on) and see if there's a difference between what the calculator says you'll get, and what GT4 actually gives you on the screen.
Race's that say they will be worth 60 points seem to be pretty accurate, with the accuracy decreasing the further the estimation is from the base 60.

The Honda Element is also a sweet vehicle to get 200 points with, especially on the dirt and snow.

Yeah I know all about the Honda Element on dirt and snow. (EDIT Dirt only actually) It needs a Full custom Transmission, to really be useful though IMO.

 

[Wild Cobra Z28]

Yeah I know all about the Honda Element on dirt and snow. It needs a Full custom Transmission, to really be useful though IMO.

I found the full custom transmission is the best way to optimize the power curve for acceleration and top speed. Almost all 200 point races need it in my view.

I haven't raced the game for years. Maybe I should hook the steering wheel back up to the PS2.

 

[GTsail]

Glad someone put my data to use. Hard to believe I spent so much time compiling it, and you spent more time using it....

Great to hear from you again Wild Cobra!

Thanks for all your work on the A-spec car values. I know I found them extremely useful!

Respectfully,
GTsail

 

[iceboy214]

I wonder if anyone ever made a tuning calculator for GT3 A-Spec. That'd be cool to have, I'd make one if I knew how to.

 

[shotamagee]

I wonder if anyone ever made a tuning calculator for GT3 A-Spec. That'd be cool to have, I'd make one if I knew how to.

Thanks for the bump, but this isn't a tuning calculator, it's an A-spec points calculator (have you even looked at the attachment?), which is a component I don't think GT3 even has.
I'm not even sure what info a tuning calculator would contain, but the GT3 tuning section is probably your best bet for something resembling what you're looking for.