Da Groj

Jeremy wrote:Seriously though, I don't understand how you guys think formulas are an easy way to class cars. Let's say I am a noob that shows up to an event in a 1991 Camaro V8. Who is going to help me class my car that morning, and how long is it going to take? How many people are going to need help the morning of the event with car classing?

Classes can be adjusted (and have been) following an event. If a noob shows up IN A STOCK CAR WITH STREET TIRES and no one knows how to class him, let him run with whatever class has the fewest people and later get the facts to class him.

Personally I think that SCCA PAX numbers don't work for our events for various reasons. I looked at this in some depth a few years ago when we were preparing to go to NASA.

I've run the numbers for most of the SCCA PAX cars from an earlier post along with some others we have seen in classes F and G against Loren's proposed formula. I've used the 2008 model year unless otherwise stated. Curb weight was from Edmunds.com (not the most reliable source, but ...) and OE tire size from tirerack.com

Corvette C5 Z06
3200/(505/325) = 2059

911 GT3
3075/(415/306) = 2260

'06 Pontiac GTO
3725/(400/245) = 2281

Elise
1984/(189/225) = 2361

BMW M3
3726/(414/265) = 2385

Subaru STi
3300/(300/225) = 2475

Mazdaspeed 3
3153/(263/215) = 2577

Mustang GT
3356/(300/235) = 2629

'01 Integra Type R
2639/(195/195) = 2639

350Z
3327/(306/245) = 2663

'05 Neon SRT4
2900/(220/205) = 2702

Infiniti G35
3532/(306/245) = 2828

Subara WRX
3142/(224/205) = 2875

S2000
2835/(240/245) = 2894

'04 Mazdaspeed MX5
2529/(178/205) = 2912

RX8
3053/(232/225) = 2960

'02 BMW 330Ci
3285/(197/215) = 2993

Miata 3rd gen
2478/(166/205) = 3060

Civic SI
2886/(197/215) = 3150

Mini Cooper S
2668/(172/205) = 3179

'07 Malibu 3.5L V6
3315/(217/215) = 3284

'05 Celica GT
2425/(140/195) = 3377

Pontiac G6
3415/(219/225) = 3508

'98 240SX
2800/(155/195) = 3522

VW GTI
3151/(200/225) = 3544

'07 Crossfire
3061/(215/255) = 3630

Miata 2nd gen
2300/(130/205) = 3637

'07 Ford Focus
2605/(136/195) = 3735

Mazda 3
2906/(156/205) = 3818

Miata 1st gen
2200/(110/195) = 3900

Mini Cooper
2546/(118/195) = 4207

It seems to me (based on how I remember different cars performing at our events) that there is a logical division point at 2500 and another at 3250, although you could perhaps argue that it should be at 3000. And another division point could possibly be inserted somewhere higher (3500? 3750?), but I wouldn't recommend it.

It looks like this formula could work to eliminate our current "Production street tire" classes F and G and also the forced induction multiplier. In fact, it could conceivably account for aftermarket turbos or superchargers and other engine mods if people had some way to prove what the modded performance is (dyno results?).

I'll let someone else answer "why do our own formula instead of using SCCA".

911 GT3
3075/(415/306) = 2260

'06 Pontiac GTO
3725/(400/245) = 2281

Not enough emphasis on weight, clearly demonstrated above.
Also demostrated by the fact that Miata's and S2000's have a very favorable ranking.

Jeremy wrote:Also demostrated by the fact that Miata's and S2000's have a very favorable ranking.

In what way? Ignore the numbers except to put them in groups; where a car falls within its group is irrelevant. If we call the groups SS1 (<2500), SS2 (2500-3250) and SS3 (>3250) (SS for Street Stock) then it breaks down like this:

SS1
===
Corvette C5 Z06
911 GT3
'06 Pontiac GTO
Elise
BMW M3
Subaru STi

SS2
===
Mazdaspeed 3
Mustang GT
'01 Integra Type R
350Z
'05 Neon SRT4
Infiniti G35
Subara WRX
S2000
'04 Mazdaspeed MX5
RX8
'02 BMW 330Ci
Miata 3rd gen
Civic SI
Mini Cooper S

SS3
===
'07 Malibu 3.5L V6
'05 Celica GT
Pontiac G6
'98 240SX
VW GTI
'07 Crossfire
Miata 2nd gen
'07 Ford Focus
Mazda 3
Miata 1st gen
Mini Cooper

So that keeps the 1st and 2nd gen Miatas (NA and NB) together along with the Mini Cooper (not S) and others that all seem to have similar performance specs to me in SS3.

And the S2000 is with the WRX (not STi), Cooper S, MazdaSpeed Miata and 3rd gen Miata (NC) in SS2. Is this wrong?

The division between SS1 and SS2 might want to be shifted to 2600 to keep the '04 SVT Cobra (2584) in SS1 instead of head-to-head with the current Mustang GT. Or not.

As I noted in a previous post the formula isn't perfect. But I think I've shown to my own satisfaction that it ain't bad, either.

Taking a guess at power, a guess at weight, and looking at the drive wheels for a tire size isn't all that difficult or time consuming. I think any of us could class a newb about as quickly as we could look their car up in a book... and probably a lot quicker than the time it would take us to FIND the book. And then we have to keep at least one current copy of the book on-hand, yadda yadda. All to make use of a classing system that we already know has pretty serious imperfections. Nevermind the fact that SCCA has EIGHT stock classes, and we want no more than four.

My opinion is that classing using SCCA's classing structure is just not for us. I'm not making the decision, but that's my personal opinion.

I think Bill's examples show that this simple calculation provides a MUCH better parity between cars than our current classing system does. The remaining question from me is: Does it give us a good spread of cars between classes given our current crop of competitors? Does it solve the "one class is far too big" issue?

Anybody wanna run those numbers?

Jeremy wrote:Not enough emphasis on weight, clearly demonstrated above.

Feel free to rerun all the numbers (weight, hp, tire) with the following formula:

(weight * weight / 3000) / (hp / tire)

This biases the results so that cars above 3000 lbs score "slower" and those much less will score much "faster". (I picked 3000 because it looked approx. correct).

Sort the results and suggest some divisions. Maybe it'll work and maybe not.

Quick glance it says the NC (3rd gen) Miata is faster than the S2000 which is just a bit faster than the NB Miata, followed by the NA Miata. It also says the GTO is between the NA and NB Miata. and the STi is between the NC and S2000; that can't be right!

Maybe 3000 is the wrong number, or maybe this is a bad formula. It is someone else's turn to crunch numbers!

Eventually we will have a formula that is complicated enough to tell us the same thing that the PAX index already does.

Jeremy wrote:Eventually we will have a formula that is complicated enough to tell us the same thing that the PAX index already does.

This is why I liked the formula I proposed. It's very simple and seems to do a good enough job. Again... far better than a simple displacement limit.

Jeremy wrote:Eventually we will have a formula that is complicated enough to tell us the same thing that the PAX index already does.

The extra complication is why I didn't choose to run all those numbers again myself. However, if you feel strongly that Loren's proposed formula doesn't give enough bias to weight, you are welcome to try.

Because I felt strongly that the PAX index does not reflect the conditions under which we compete I put the effort into doing much of the needed research and all those calculations and sorting.

Jeremy, I think the ball is now in your court to prove that the PAX index does a better job of matching our conditions and expected results than the numbers I posted, and then suggesting how to break up the index into FAST classes.

I have no doubt that it works for SCCA, but we aren't SCCA and don't compete under the same conditions, IMO.

Well you could split up the pax index in as many divisions as you want. I originally suggested splitting it in half.

CLASS A
100
SS
Elise
Corvette C5 ZO6
911 GT3

99
AS
S2000
Corvette C4
Subaru STi

98
BS
Rx-8
350Z
Porsche 944S

97.9
CS
Miata (2nd/3rd Gen)
Solstice N/A
MR2 Spyder


CLASS B
96.9
FS
Mustang
Infiniti G35
Camaro z28

96.5
ES
Miata (1st Gen)
MR2 (2nd Gen)
Porsche 944

96.3
GS
Toyota Celica GT
Mini Cooper S
Civic Si (8th gen)

96
DS
Subaru WRX
Integra Type R
Mazdaspeed 3

94
HS
Mini N/A
Civic Si (all except 8th gen)
Mazda 3

These two classes would simply replace the over/under 3L division in the current ruleset. All you would need to do is download a few pages of the rule book dealing with stock car classifications, which is available online for free.

No complex formulas, no penalties for heavy cars, no guessing weight and HP.

Specific problems from the examples that Bill posted...
Pontiac GTO, Mazdaspeed3, and Mustang GT are too high.
S2000, RX-8, Miatas are too low.

I guess something like that could work. I think we need more than two classes, though. Back when we were talking about 40 car events and less than half the cars were stock... two classes was good, about 10 cars per class.

Now, we're talking 60+ car events with 30 stock cars. I think 3-4 classes would be appropriate.

But, I will concede that if we can keep the classing list on-hand, it is a pretty easy method of classing.

I like a split based on PAX numbers, simply because I don't think a formula method would be ideal, since you are grouping some cars that are similar in different classes (STi and S2000 - in reality these are competitive to each other), but puts the Integra Type-R (a fast autocross car yes, but not competitive with the S2000) with the S2000, and the WRX with those, when an ITR and S2000 will beat the crap out of a WRX.

There's so many variables that a simple formula I don't think will really show what cars are really competitive with each other..

If you want 4 stock classes try:
1. SS, AS, BS
2. CS, FS
3. ES, DS
4. GS, HS

Yes I put GS and DS out of order, but in my experience DS cars are faster than GS, maybe not.

I thought we were just splitting G class not reorganizing the entire thing. Take the two group pax model and just list the cars that would be in G to show how it would relate to FAST.

impalanut wrote:I thought we were just splitting G class not reorganizing the entire thing. Take the two group pax model and just list the cars that would be in G to show how it would relate to FAST.

G split into 2 using PAX, from the last event....

Class A
1T 49 Dave Balogh 06 Honda S2000
2 113 Mark Lasota 05 Mazda Mx-5
4 55 Yang Gao 04 Honda S2000
5 5 Jack Neely AP1 S2000
7 1 Jeremy Damico 05 Honda S2000
8 84 Justin Brew 03 Honda S2000
9 65 Mark Jones 04 Mazda Mazdaspeed MX-5
13 58 Christina Edwards 05 Mazda RX8
22 27 Ed Courtemanche 99 Mazda Miata
23 20 Mike Kennedy 04 Mazda Miata

Class B
3 37 Steven Frank 94 Mazda Miata
6 30 Brian Kollar 03 Honda Civic Si
10 22 Nicholas Kriebel 02 Ford SVT Focus
11 100 Thomas Russo 06 Honda Civic SI
14 77 Carl Nyman 93 Mazda MX5
15 256 Jordan Young 91 Nissan 240sx
16 10 Mario Kikic 95 Mazda Miata
17 83 Carl McKim 96 Mazda Miata
18 711 Andrew Coeulton Chevy Malibu
19 44 Pamela Corbin 93 Mazda Miata
20 306 Larry Kollar Honda Civic Si
21 2 Rudy Kriebel 94 Mazda Miata
24 717 Tom Mann Chevy Malibu
25 109 Andrew Baraban 07 Scion tC
26 48 Timothy Taylor 05 Dodge Neon
27 183 Rick McKim 96 Mazda Miata
28 88 Erik Cordinas Mazda MX-3

thats what i thought . g is the problem child. thats the glass we need to deal with. just split off a miata only class

Look at all the other cars that are present in Jeremy's "B" sample. Is it fair to them to have a "Miata only" class? While FAST has roots with a Miata club, we're NOT a Miata club any more, and thus we can't be elitist.

Jeremy, you keep talking about wanting Mustangs to compete in the same class with cars like Miatas... how does your solution address that?

miazda man wrote:thats what i thought . g is the problem child. thats the glass we need to deal with. just split off a miata only class

I already said why a miata-only class wouldn't work. It could work, but you would need to reorganize all the classes, and at that point it's useless.

Loren wrote: Jeremy, you keep talking about wanting Mustangs to compete in the same class with cars like Miatas... how does your solution address that?

See Dave's example...

If you want 4 stock classes try:
1. SS, AS, BS
2. CS, FS
3. ES, DS
4. GS, HS

He grouped Mustangs (FS) against 2nd and 3rd Gen Miatas (CS).
Really I was leaving how to split up the pax to someone else. I haven't even looked at that yet. I was assuming that splitting the pax into 2 would be good enough to keep the class numbers down, but apparently it may not be.

I think Dave's classing map has merit if we want to use SCCA classing.

Two class model based on PAX, from last event results.
People highlighted in Blue were brought in from Class F.

Class A
1T 49 Dave Balogh 06 Honda S2000
2 113 Mark Lasota 05 Mazda Mx-5
4 55 Yang Gao 04 Honda S2000
3 3 Clinton Wilson 03 BMW M3
5 5 Jack Neely AP1 S2000
7 1 Jeremy Damico 05 Honda S2000
8 84 Justin Brew 03 Honda S2000
9 65 Mark Jones 04 Mazda Mazdaspeed MX-5
4 124 Joe Marchica 93 Toyota MR2 turbo
13 58 Christina Edwards 05 Mazda RX8
6 115 Chris Rousch 06 Nissan 350z
22 27 Ed Courtemanche 99 Mazda Miata
23 20 Mike Kennedy 04 Mazda Miata

Class B
1T 38 John Edwards 08 Mazda Mazdaspeed 3
3 37 Steven Frank 94 Mazda Miata
2 24 Aaron Parke 07 Mazda Mazdaspeed3
6 30 Brian Kollar 03 Honda Civic Si
10 22 Nicholas Kriebel 02 Ford SVT Focus
11 100 Thomas Russo 06 Honda Civic SI
14 77 Carl Nyman 93 Mazda MX5
15 256 Jordan Young 91 Nissan 240sx
16 10 Mario Kikic 95 Mazda Miata
17 83 Carl McKim 96 Mazda Miata
5 45 Travis Floyd 08 Pontiac G6
18 711 Andrew Coeulton Chevy Malibu
19 44 Pamela Corbin 93 Mazda Miata
20 306 Larry Kollar Honda Civic Si
21 2 Rudy Kriebel 94 Mazda Miata
24 717 Tom Mann Chevy Malibu
99 Mike Chawk 06 vw gti
25 109 Andrew Baraban 07 Scion tC
8 117 Harold Heimmer 05 Chrysler Crossfire
26 48 Timothy Taylor 05 Dodge Neon
27 183 Rick McKim 96 Mazda Miata
9 310 Steve Woodbury Ford Mustang(ASSUMING V6*)
28 88 Erik Cordinas Mazda MX-3