Notes on Maximising A-Spec Points

[Famine]

I've been running a few more tests and I've reached the following conclusions.

A-Spec Points Are Dependant Upon (in rank order):
1. Comparison of figures with opponent car which would score lowest A-spec points if entered into race.
2. Tyre choice (softest axle only).
3. Weight of player car.
4. Power-to-weight ratio of player car.

A-Spec points are Independant of:
Car power.
Remainder of AI grid.
Any non-power, non-weight, non-tyre part (including suspension and Nitrous).
Track.

The Jury is Out On:
Torque (bastard... )

 

[f1king]

I had a chance today to test and crack out a theory.

"How can I maximise my potential A-spec points reward for a given race", I thought. Some of the gonks may answer "its easi just get a wurser car than the ai lololol !11!!111!", but how DOES GT4 define a worse car?

Power is an obvious starting point. So, I went to the Lupo cup, which allows in a choice of four cars - the Lupo 1.4, Lupo GTi, Lupo Cup Car and Lupo GTi Cup Car. In all cases, the AI field was made up entirely of Cup Cars and GTi Cup Cars.

So, test 1:
Lupo 1.4i, 100PS, 193pt
Lupo GTi, 116PS, 182pt
Lupo Cup Car, 123PS, 115pt
Lupo GTi Cup Car, 133PS, 117pt

Oh fudge. Looking good until that pesky GTi Cup Car wrecks the pattern - despite being the most powerful car, it garners more A-spec points than the standard Lupo Cup Car.

So, how about a power-to-weight ratio-based system? Let's look at test 1 more closely...
Lupo 1.4i, 100PS, 933kg (107.2PS/tonne), 193pt
Lupo GTi, 116PS, 975kg (119.0PS/tonne), 182pt
Lupo Cup Car, 123PS, 840kg (146.4PS/tonne), 115pt
Lupo GTi Cup Car, 133PS, 1010kg (131.7PS/tonne) 117pt

Bingo! Or so it seems. As the power-to-weight ratio increases (and PLEASE note that power-to-weight ratio is exactly that. Power per unit weight. As it increases the car gets faster. NOT a Weight-to-power ratio, which is weight per unit power and the car gets faster as it decreases. I've had this argument before, and if you disagree with me you are wrong) so the A-spec points get less.

So, I ran a quick examination of this with the Lupo 1.4i and got the results of test 2:
Lupo 1.4i, 100PS, 933kg, 107.2, 193
Lupo 1.4i, 116PS, 933kg, 124.3, 154
Lupo 1.4i, 116PS, 905kg, 128.2, 148
Lupo 1.4i, 116PS, 877kg, 132.3, 144
Lupo 1.4i, 116PS, 849kg, 136.6, 139

Not bad. As PWR steadily increases, so A-spec points decrease. Nice. Only, I'd forgotten ONE thing. I'd accounted for a PWR increase by weight loss, but not compared the weight loss cars with power increase cars to roughly the same PWR level. They should get roughly similar A-spec points. Right? Right?

Test 3:
Lupo 1.4i, 116PS, 933kg, 124.3, 154
Lupo 1.4i, 116PS, 905kg, 128.2, 148
Lupo 1.4i, 122PS, 933kg, 130.8, 164
Lupo 1.4i, 116PS, 877kg, 132.3, 144

Well, I'd like to be the first person to say "****!". We have a Lupo with a higher PWR getting disproportionately higher A-spec points. This led to my final modification of the theory, with a twin test (two cars identical in every measureable respect, but different models - to eliminate any variables other than the ones being tested). I bunged a 1996 Mitsubishi GTO SR and a 1998 Mitsubishi GTO SR into the Beginner 4WD Challenge - making sure that the grid always contained one Audi RS6 (by far the most powerful and highest PWR car competing in this tournament).

Test 4:
214PS, 1610kg, 132.9, 200
214PS, 1336kg, 160.2, 89
261PS, 1336kg, 195.4, 70
317PS, 1610kg, 196.9, 75
317PS, 1336kg, 237.3, 8

We can see that the middle car, despite having a marginally worse PWR than the fourth car, has a significantly improved A-spec points tally.

From this I conclude:

The primary determining factor in A-spec points calculation for a given field of AI cars is the player's vehicle mass, followed by a power-to-weight ratio curve.

The PWR curve appears to be different on a car-to-car basis - the Lupo Cup Car, despite having identical stats to the best AI car in the field, racked 115 points. A Lupo GTi (NON-Cup Car), with a fairly close PWR match, got 122 points. The GTO SRs, despite being 3% worse off for PWR, got only EIGHT points.

As far as I'm aware, it's only the difference to the BEST, statistically, AI car in the field that counts. After all, you only get your A-spec points for winning, so if even one car - the best car - beats you, you don't get any.

I've attempted to draw curves based on PWR and % difference from the lead AI car, to no avail, so any further thoughts on this matter would be appreciated. Nevertheless, the above theory seems to handily explain why a twelve million horsepower pickup truck still gets hundreds of A-spec points each race - it's a big fat bugger and weight is the primary function...

Lots of time on your hands, But NICE!!

 

[JParker]

Great job Famine. And Duke I was just reading through the thread and was getting a little irritated at the BS that guy was posting and was wandering if I could send a message to a moderator for some cleaning of house, and came across your post, "We're watching him". Thanks man. I don't know why so many people attack Famine. Everything I've seen him post is extremely informative, and you really see the time he puts into his research.

Keep up the good work Famine!!

I saw a post where you said NOS didn't effect the points, but I found differently. I was running the Tuner Championship with my 350Z Roadster modded out, and got a little aggrevated when the RUF spun me so bad he managed to gain nearly 6 seconds on me on the last lap, I restarted and enable my NOS to find less points. I have also noticed that increasing your NOS value also changes the A-specs. Additionally, last night I was running the JGTC, playing with my HCS, again restarting with frustation, and found some gearing changes to effect points as well.

Again, I would personally like to thank Famine for his extensive work he does and the time he takes out of his day to day to post it here for us to see.

 

[Sven]

I'm kind of wondering why an Impreza Rally car in the turbo cup got 180 pts, while a CTR Yellow Bird in the European Racer cup (against a Cizeta and SLR) got only about 50...

 

[Speedy Samurai]

Take your Peugot Rally Car into the Peugot manufacturer race and win, on average, 160 Aspec points.

 

[Solid Lifters]

Then it rang a bell on me: no wonder I was being offered 200 A-spec points for the Dodge RAM 1500 LARAME Hemi Quad Cab '04 (that has to be the longest name ever for a vehicle ), with its whopping 2013 Kg (after Weight Reduction: Stage 1).

The Wizard.

Before getting all excited about them points, perhaps you should record how much you're actually getting. I've discovered that you never get what the game states you will while using this truck.

 

[Cuandoman]

It looks like you all need a "Design of Experiments" type system to solve this kind of thing. Some sort of multi level factorial experiment where everyone can help test and provide data.

Do any of you know what I'm talking about? (I'm an Industrial Engineer)

Dan Durkee

 

[TheWizard]

Before getting all excited about them points, perhaps you should record how much you're actually getting. I've discovered that you never get what the game states you will while using this truck.

Nothing to worry about, here. I quit running the thing as soon as I was done with the Beginner Events. I can't stand trucks at all, I just bought it for the 'Sports Truck Race'. Then I noticed the amount of points I was getting - I checked, pretty much after a few races, and the total number of A-spec points seemed to increase by reasonable amounts, as expected - and I decided to take it for a spin for another few races here and there. But then I got bored out of it quickly. I am more of a Sports Car person - I don't enjoy trucks, never owned one in real life, and never will.

By the way, for the on topic part, I run some tests about A-spec points as well, and I will contribute in the next few days, as soon as I have a chance of finishing to test at least 2 or 3 more vehicles - to see if (part of) my theory is correct.

The Wizard.

 

[Midget_I]

Anyway, it's a helpful post in general (although much more data is required; I saw someone saying a modded Suzuki Capuchino -- a very light car -- got a ton of A-spec points even in races it crushed). It's just sad to see a long-time member who's more interested in stroking his ego than being civil.

Bats is right. I got 200 points for the Beginner FR Seatle race in a Ginetta at only 440 kg...after entering 10 times.

 

[Awaji Islander]

I was doing the F-1 series last night, and A-spec points changed because of change in down-force.

 

[JP Taillon]

Manual transmission will get you more points than an automatic. I have also found that every upgrade on the car changes the amount of points you get. Try it for yourself. Start a race and then exit when you know how many pts you would've gotten, than exit and turn non-power, non-weight mods off, you'll see that the amount of pts awarded increase.

 

[Neurotik18]

I'm a Mech Engineer. Design of Experiments would be good. Use software like JMP.

 

[Azuremen]

Famine... not sure if this might help, but the game still looks at things kinda funny from the weight ratio, which you pointed is moderately important.

I'm using the Mazda Autozam AZ-1 at full tune, good for 123 HP and weighs in at 650kg I believe. Full tune, everything, even the spoiler. Now, its extremely light weight, and is on SS tires, and is getting me 200 A-spec points in things such as the Pro clubman's Cup... when the Elise 111R 04 or whatever its called is on the field.

Not sure if that could make any diff, but most people tend to avoid the ultra light weight cars it seems, so I'd give some input.

Also, extremely nice work here, well thought out and what not. Also cleared up if A-Spec points actually do anything. Bummer that they don't, since I have almost 10,000 before completeing 20% of the game...

 

[Hemmo]

thats just stupid i wanna drive fast cars not some slow and boring car just to get a-spec points..so i dont care..

 

[Natural Newbie]

I bet A-spec points were going to be used to find a "quickmatch" (in xbox terms) for online races. But then again it would make more sence to have online points, rather than group people together with their single player/simulation stats. Ah well. I've also noticed then when re-trying a race that I lost, without changing anything the points could drop 2-15 points. The equation for calculating A-spec points is probably HUMUNGOUS. Keep diggin famine :-)

 

[Locost7Rules]

Several people having been making posts along the lines of "Hey, LOL, I did the XX series last night and only changed downforce and it IMO, LOL 👍 , totally affected the A-Spec points." Look guys, think about this for a second. Useful empirical data can only be obtained if you isolate what you are trying to test and run a control. What Famine has been doing (or so it seems from the posts, or if not, should be doing) is pick a car and purchase two of them new. Get an oil change in one. Test the bone stock one on a particular course. Record the A-Spec Points. Then test the one with the oil change. Record the A-Spec Points. The change something else, and test it, on the same course. You can only change one thing each time. This is going to be really difficult because the fields are not going to be the same every time, so it will take much starting and quitting. Anyway, this is how I was taught to set up and run experiments in the 6 science classes I took in high school, and the 5 science and 6 engineering classes I've taken in college. But what do I know, I am only a sophomore (junior after this quarter though, too bad I going to be here forever), maybe in 500 level engineering classes they just keep changing things randomly until it works.

 

[PhatFat]

Several people having been making posts along the lines of "Hey, LOL, I did the XX series last night and only changed downforce and it IMO, LOL 👍 , totally affected the A-Spec points." Look guys, think about this for a second. Useful empirical data can only be obtained if you isolate what you are trying to test and run a control. What Famine has been doing (or so it seems from the posts, or if not, should be doing) is pick a car and purchase two of them new. Get an oil change in one. Test the bone stock one on a particular course. Record the A-Spec Points. Then test the one with the oil change. Record the A-Spec Points. The change something else, and test it, on the same course. You can only change one thing each time. This is going to be really difficult because the fields are not going to be the same every time, so it will take much starting and quitting. Anyway, this is how I was taught to set up and run experiments in the 6 science classes I took in high school, and the 5 science and 6 engineering classes I've taken in college. But what do I know, I am only a sophomore (junior after this quarter though, too bad I going to be here forever), maybe in 500 level engineering classes they just keep changing things randomly until it works.

I have entered a race and then exited it, hit retry, so all the cars are the same, and then adjusted the down force and the A-spec points do change.

 

[jlosordo]

I'm not real big on building a spec stats, but there is a glitch on a-spec points. You can keep doing missions over and over and get 250pts a pop. In my opinion mission 8 is a sure win. Off-topic,: Love my blue and white F1.

 

[dbartucci]

I'm not real big on building a spec stats, but there is a glitch on a-spec points. You can keep doing missions over and over and get 250pts a pop. In my opinion mission 8 is a sure win. Off-topic,: Love my blue and white F1.

I'm pretty sure you won't actually receive the points again and again, it'll tell you what you would have received, so if it's higher than your previous high score you can add to your total, but equal or lesser values won't add on.

 

[jlosordo]

I'm pretty sure you won't actually receive the points again and again, it'll tell you what you would have received, so if it's higher than your previous high score you can add to your total, but equal or lesser values won't add on.

I know, found out last night. thanks

 

[SFC_200SX]

It's funny how often arrogant, condescending people are wrong.

Ain't irony a beautiful thing?

If this thread is a competition between people with time on their hands to slag people off vs. people with time on their hands to do something vaguely interesting and useful it's pretty clear who's in front.

Smart work famine, and on your other threads, I only registered to download some of your files and have pointed the rest of the sxoc gt4 fans this way, so be prepared for an influx of riced up rwd fanatics.

But Bill Bailey? WTF is that all about eh?

Anyway, back to getting 2 A Spec points a race or B Speccing my way to victory, Who needs licenses anyway?

 

[Famine]

In that case, keep your eyes peeled for some PAL-specific files I intend to upload tonight - along with a PAL/NTSCNA update for the Race Database...

 

[SFC_200SX]

In that case, keep your eyes peeled for some PAL-specific files I intend to upload tonight - along with a PAL/NTSCNA update for the Race Database...

Nice one. The thing I'm most interested in at the moment is when my car will come back in stock cos I missed it first time around (200SX S13).

I don't think it's the same timing as the 240SX for the US version.

Got to have your own car innit.

 

[Famine]

It certainly isn't the same as the US version - OR the Japanese version (which I spent a month doing the first 20 cycles of. Bugger). I've completed the first 10 cycles and will post it up this evening.

 

[Warp3]

Another possibility I've considered is that perhaps, instead of calculating the car attributes, the game quickly calculates a "steady" b-spec-style lap for each car on the grid, then rates your a-spec points based on how fast it ran with your car vs the fastest of the cpu cars.

While this sounds like it would be more accurate than a formula and thus could not be true (as it is obviously very inaccurate in certain scenarios, like the aforementioned Dodge Ram issue), the CPU often can't drive certain cars as quickly as a human can as it is unwilling to use the slip angles or techniques necessary to do so. So perhaps the reason you get so many points for using a Dodge Ram is because the cpu simply does a crappy job when *it* tries to drive it?

A great example of this are those spots on the course where the "downshift" indictator tells you to do something that is completely unnecessary and (in some cases) even downright moronic. For example, in my fully tuned 700hp Impreza Super Touring car (with racing super hard tires), it recommends I slow down but stay in 5th gear in the "kink" in the long straight on Tokyo R246 (and even recommends 4th when my tires are orange/red), yet as long as you are even remotely close to the right side of the road, you can shift to 6th (as I am just then redlining 5th at that point of the straight) and keep your foot planted to the floor and, with little to no steering input, charge right through with barely even a hint of squeal from the tires (and, to be honest, I haven't driven a car yet, that can't flat foot that spot on Tokyo R246 without issue (even heavy, high-powered stock cars with street tires can do it without a problem)). But if this is an indication that a cpu driver would choose to slow down here, then you could see how the lap times could differ greatly between a human driver and a cpu driver, if it does indeed calculate a-spec points that way.

 

[jake11375]

The power to weight theory sounds like a good fit. I remember someone saying they received a ton of A-spec points using a 500hp SVT Lightning. Trucks are heavy indeed...

 

[KAuss]

Starting grid plays a role in A-Spec points

I'm wondering is driving aids might take a role in the points... I'm too lazy to do this and don't even care enough to check back for the answer...

A-Spec points don't do anything, I'd rather use lap times for bragging rights...

 

[TheWizard]

To my knowledge driving aids do not have any effect on A-spec points. I did a test not too long ago, where everything else was held constant, and I only changed the parameter that I wanted to test. Driving aids was the first thing I checked, and the A-spec points did not change at all.

This is how I performed my test, if anybody is interested. I made a backup save of my memory card, so I wouldn't end up messing up my current save game. Then I bought one vehicle, bought every possible parts for it, did not apply any, entered a championship series, exited it, saved the game, and started testing the parts/mods that I wanted to check. Then I would reset the PS2, and start all over.

Up to this point I can confirm a few points brought up by Famine:

1) The track does not affect the number of A-spec points
2) Tires (what I tested the most) definitely do
3) Different combinations of tires yield the same amount of A-spec points that applying the softer of the two types of tires to both axles would
4) Driving aids do not affect A-spec points

Maybe one of these days I will finish my tests, but for now this is all I have.

The Wizard.

 

[durden]

now these a-spec points, can you accumulate them by racing the same race over and over again, meaning there is no limit to the amount of points you can get, or is there a set limit you can get for each race, and once you get that max amount, that is all you can get from that race? hopefully that question isnt too confusing. ill find out soon enough, ive just havent got around to actually racing yet, too busy golding all the licenses and beating the mission hall.

 

[Famine]

No - you can only get A-spec points from each race once.

That is, if you get 80 A-Spec points from a race, you cannot go back immediately and get 80 more - or go with a better car and get 5 more.

BUT you can go back with a worse car and you'll be "refunded the difference". If you went back with a 200 point car and won, you'd get the 120 points credited to your A-spec point tally.