The kpi list above appears to be the included angle. The kpi plus the caster equals the included angle. For example, a older miata has a included angle of 11.75. Subtracting the recommended caster setting of 4.44 will give you the kpi, which is 7.31 degrees. I haven’t looked up any others.
The angles I've quoted are KPI, plain and simple... the angle a line through the ball joints makes relative to vertical viewed in the YZ plane (i.e. viewed from the front of the car).
Caster is the angle a line through the ball joints makes relaive to vertical viewed in the ZX plane (i.e. viewed from the side of them car).
Subtracting the caster from the KPI makes no sense.
You seem to be trying hard to fit theory to an anecdotal experience. Fitting 35mm offset wheels to your Miata should give you the same general situation as the Elise, F1 and C5 shown. If you have problems with it, look elsewhere for answers; it's not likely the +10mm scrub radius at fault. Gordon Murray ain't stupid.
A former Lotus chassis engineer explained that keeping the (static design) scrub radius a little bit positive is better than zero because it keeps the state the same as the tire squirms from normal road irregularities and minor steering corrections... you don't want the scrub radius constantly flopping from pos to neg as you drive, alternately loading the joints in different directions when you just want to go straight. Makes sense to me.
I have a stack of Carroll Smith's books. He doesn't dwell on kpi and scrub radius, but he does state that kpi is normally 6 to 8 degrees. He goes on to say that increasing scrub radius is unlikely to have any beneficial effects on the steering.
Scrub radius is analogous to bumpsteer. Like bumpsteer, scrub is feedback through the steering wheel. Less feel from scrub is a good thing. If you want more feel, use a numerically lower ratio rack.
Herb Adams states that scrub radius should be reduced to the minimum and by minimizing it, there will be an improvement in handling, control, and steering effort.
Norman Garett states that scrub radius is not dangerous, but it is "tiring and unnerving". "Mazda spent millions of dollars developing the Miata suspension to avoid exactly this kind of bad behavior."
I am inclined to agree with these fellows. I don't feel there is any worthwhile reason to add to the scrub radius on the front of a rwd car. On my car, I've decided not to modify my spindles and see how it goes. I have not determined what the tbird's scrub radius is, so at this point, I don't know if I am making it worse or better. I will be checking it.
Changing rack ratio can't magnify a signal that you've eliminated. In any case, scrub radius is dynamic. The Miata only has zero INITIAL scrub radius .
King pin inclination causes the contact patch to move outward when you steer off-center; to a positive scrub radius situation. So it's not zero scrub radius vs. non-zero... it's zero migrating to positive, positive migrating to more positive, or negative migrating to less negative and possibly crossing over to positive. Which of these three situations is constant in sign?
Always keep in mind, when reading textbooks, not just the authors' credibility, but also the era, environment, and practices at the time of writing. I've had, and frequently re-read, a (first draft, hand corrected) copy of the works of Maurice Olley ("father of vehicle dynamics"... brilliant guy) since 1974, twenty-five years before the Millikens were given permission to publish them. I've worn out Carroll Smith's books and gone back for seconds. I worship these guys, but temper their words with the realization that they were only the best understanding at the time. For example when someone rails against scrub radius it's helpful to know that it's in the context of many production cars with 4" (made parking easy with manual steering... roll the tire around the steering axis rather than twisting a patch) and race cars with 6" at the time
. That 6" is very bad doesn't necessarily make 0.6" worse than zero. Be aware that what applies to a race car with massive downforce and front tires a foot wide may not make sense for a car with neither... different situations call for different tradeoffs.
I'd point out that none of the engineers involved with the four examples I gave were hindered by using donor car parts, nor were they limited to existing wheels available from Tire Rack; they could manufacture any geometry, KPI, scrub radius, caster they want... they don't lack performance knowledge; they are closely linked to the (factory effort) experience from multiple F1 WCs, multiple LeMans overall and class victories, etc. They have test tracks, instrumentation, contracted test drivers, and the means to test anything they think up. When they build their highly visible "halo" high-performance road cars they all settled on KPI around 10°, caster around 5°, and three out of four have significant positive scrub radius. I hate to discourage thinking for oneself, but I can't ignore the conclusions of the aforementioned either.
Don't assume everything you read is right... I've seen chassis engineers and tire engineers ridicule Herb Adams' book (quote: "I view it as the absolute worst book on suspension design ever written. I figure the fewer people read it, the fewer will get hurt.
My copy is hidden away somewhere in storage, so I haven't looked at it for years (not sure what it is that upsets them), but I'd be a little more kind... he was racing at a time of rapid change 40 years ago when no one really quite understood what they were doing (compared to now, when the best halfway understand almost half of what they're doing)... and trying to boil down all the complex interactions to simple explanations does tend to produce rubbish.