So I'm driving home, alone, from Peru, and I'm thinking about the Town Hall
Meeting in Topeka this fall, and about how I'll probably wind up giving a
"State of the SM Union" address, and how I want to keep it *short*, and so
I have to drill down to the essentials of what SM has accomplished this
year, and what it needs to keep moving forward in 2001.
And I'm thinking about how a certain someone is running around telling
people that "SM is unfair to front drivers" - which is *totally* bogus, as
CSP (even the "slow" CSP cars which are SM eligible, not just the CRXen)
has no problem running competitive SM times is *SP* trim, never mind adding
extra power to them. But still, well prepared is well armed and all that,
so I'm running over data points in my mind.
The Peru Tour is instructive, for two reasons. 1) It's the first event this
year where Kent and I met up on a wide open, straight-up fight and 2) It's
the first National event where a small, light, but not particularly
powerful front-driver has shown up.
# 1 I find interesting, because I've noticed that as the courses get
shorter and tighter, I tend to do better against the Supra. As they open
up, the gap opens up, all else being equal.
# 2 is interesting too, as the same goes for this car, only backwards. As
the courses get shorter and tighter, this car (A VW Sirocco) does better
and better. In fact, he's savaged me at the last couple of local events, on
slow, tight, narrow courses. But the tables were turned in Peru. This car
weighs 1800lbs (we weighed it at the event) which is a little more than
_half_ of my car (and LESS than half of the Supra) but he was still
outgunned, power-wise.
So imagine that we had an index for courses. a "1.000" course is a 2 km
long straight stretch, 25m wide. A "0.001" course is your typical CCM
course - 10' wide, and snaking though every possible atom of pavement in a
200m^2 pad, with no turn less than 160 degrees, and no radius larger than
about 10m.
On the 1.000 course, you'd expect to see Supra, Talon, VW, with large
intervals between the cars because of the large power differential. On the
0.001 course, it's reversed: VW, Talon, Supra, because of weight, size, and
tire area.
Now if you had some sort of magic pavement, that you could redraw courses
by turning a knob, as you turned the knob from 1.000 to 0.001, you reduce
the power advantage differential until you reach the point of course
complexity where all three cars are equal (or at the very least, where the
differences between them fall into the noise from driver ability)
Now consider this - let's say the "equality point" is at 0.500. This means
the VW has an advantage roughly 1/4 of the time, the Supra has an advantage
1/4 of the time, and all three cars are even roughly half the time -
**assuming that the distribution of actuall course complexity is linear**.
If the actual distribution of course complexity is a bell curve centered on
.500, then the Supra gets an advantage rarely, the VW gets an advantage
rarely, and most of the time the class is fair.
Now, let's assume that the "equality point" is not .500, but .250, and that
the course distribution is a bell curve centered on .500. If this is the
case, then the Supra has a slight advantage more often than not, and a
large advantage rarely, while the poor VW has a strong advantage rarely,
but a slight disadvantage more often than not. This is an unfair class.
So it seems to me that there's a pair of important data points missing - at
what course complexity value is a given car's performance equal to
another's, and what is the frequency distribution of course complexity?
I would guess too, that different events have different average
complexities. Nationals and Tours may have a .700 complexity, Pros a .500
complexity, and local events a .300 complexity. If the spread is this wide,
then a given class's "fairness" may vary widely with the kind of course the
cars are running on.
It would be _very_ interesting to come up with some sort of measurable
course complexity value, and then plot results in classes against the
complexity value, and see what correlations one could discover.
Except for one thing - I haven't a clue of how to do that. Any ideas?
DG
|