@Kameleon,  apologies if I have mixed my terms.  I meant to point out that excessive change in toe (in or out) will lead to wear in the steering linkages if/when both wheels bump at the same time. Not relevant for track use where change in toe with roll seems to be issue.
 I certainly will get the alignment done. Do you have any suggested values for road use? The car is lowered 1" by Koni/Eibach shocks. If not I will go back to Alex W's articles and try to work out some mid point. cheers.
 
@B24  - thanks for the photos. I suspect a different ballgame altogether with that much lowering. 
  

No need to apologise. 
 
If it is just for road use i would use:-
 
Slight toe in front, little to no toe in rear.
 
1.5 deg neg camber front, 1 deg neg camber rear.
 
This will handle good but not dart across the road for a street car. Alex's settings will be more track focussed and i do not think that suits your usage.

Thanks Kameleon,
 
I'm not sure what "slight toe" is ?   1/2 deg ?
 

1-2mm

Thanks @Kameleon

Steering geometry that causes bump steer isn't going to make an appreciable difference to the loads and subsequent wear rates of the steering components.  Bump steer is caused by steering angle changes throughout suspension travel, and the steering movement produced is only a small fraction of the suspension travel.  
 
At any time a car is on the ground there are loads on the steering components, adding the dynamics of motion changes these loads in one way or another.  The single biggest factor that will increase steering component loads when hitting bumps is the distance between the steering axis at road level and the edge of the tyre, and this is related to the scrub radius which is the distance from the same point to the tyre centre line.  This distance is essentially the lever that produces steering forces when a bump is hit.  Factors that change it significantly are wider tyres and increased negative wheel offsets.  Small changes here can profoundly affect the transmission of bump forces into the steering.  Larger diameter tyres, due to steering axis inclination, tend to bring scrub radius back towards better values when offset has been negatively increased.  Large scrub radii also cause instability under brakes on uneven surfaces.
 
Another significant factor that loads steering components is increased caster angles.  High caster angles are useful to increase camber change on steering input and maintain flatter tyre contact on cornering, particularly with strut type suspensions where other camber change methods are not available.  You can quickly see the effect of caster loads by unbolting the tie rod ends and dropping the car - it immediately goes pigeon toed!
 
Finally, load on the rack bushes increases when the ride height is not ideal.  Ideal means that all four tie rod ends should be in a straight line with the car at rest, ie. tie rods horizontal.  When lowered, the outer tie rod ends are higher than the inners causing steering forces to place vertical loads on the rack, and this decreases the bush life.
 
Take Kameleon's advice and service everything in the car's suspension and steering systems, this is always the best starting point when troubleshooting sloppy handling and dynamics issues.  Then pay a bit of attention to your wheel offsets and sizing, keeping in mind they work intrinsically with the suspension geometry in determining the car's behaviour.  If you need corrections, make them thoughtfully and try not to just guess.  Ultimately if you make the car a fun place to spend time it'll be immensely satisfying knowing that you made it that way.  At least that's how it works for me!

He's Back  He's Back !!!  Welcome back.  Where yer been Cogs ?
I've enjoyed your technical articles in the past and I thank you.
Hope we might see some more good explanations from you into the future.
Cheers.

Thanks Cogs, appreciate the advice. I am indeed going down the path that you and Kameleon have recommended.
My post was really to advise those who may not know and are driving other people's efforts. I picked up a car from a young guy who either had no idea or was dumping all his worn out parts in this car, possibly keeping the good stuff for his main car.
The wheels are wrong, and lacked hub centre rings, the shocks were shot and lowered (cut?) and the geometry therefore needs sorting. So yes the reason for the steering wheel jerking may be not be from the different arcs of  tierod & control arm, but my message is really - don't ignore it. If you have major steering wheel action happening on one wheel bump, imagine when both wheels hit some bump, the steering wheel doesn't jerk at all, but the forces didn't cancel, they were absorbed and not by the suspension!  Best to reduce those forces with the changes you have listed, which I am doing.
 

Falcon
Where yer been Cogs ?

Haha, thanks Falcon!  I've always been around but just busy with RealLife™!  My '2 still sees daily use but the poor girl is sadly neglected - so many plans, so little time.  I'm a little envious of guys like TwoDogs who are just discovering the fun of MR2s, but happy to see they're keeping the scene alive.

I find the suspension articles over on the Wilhelm site very informative. Especially since several people who have pretty good understanding are collaborating their thoughts, sometimes differing, but usually not.
Many of you experienced track guys will already know the following. 
The main thing I get from their findings is that the SW20 camber (lack of) adjustment is not good.
a) The camber changes a lot with body roll,
b) the roll is increased if the car is lowered.
Untreated it means that you have to compromise between straight line grip and wear vs cornering grip.
One contributor points out that the camber change can be reduced moving the top mount of the suspension inwards 
with adjustment plates and I have no doubt that is true.
Alex Wilhelm focuses on b), that the extra roll can be removed with his compensating blocks. I have no doubt about that theory, I am not sure about how much extra roll there is, but any, makes the fundamental issue with SW20 camber, worse. I am thinking that perhaps sorting a) and b) first, might make spring/swaybar decisions a bit easier.