input power?

Quoting mrbarton-at-ix-dot-netcom-dot-com Sat Feb 17 09:56 MST 1996

<Tesla coil spark length is a weird thing that no one has really figured out 
<yet.  The length seems to depend on three things:  Input power, output 
<voltage, and rep rate.  The more of any of those, the longer the spark.  Using 
<length to figure output voltage is just plain wrong.  This can clearly be seen 
<because spark length will go up with rep rate, which has no effect on output 
<voltage.  Confused yet?  Me too.

I agree that coil length as a comparison to spark is wrong too.Since I've
always thought of the output in terms of input power (read efficiency at
processing input power) maybe we should be thinking about what amount of
power is REALLY reaching the secondary and relate this to spark length.
Has anyone considered if the K factor should be dialed into this?The way 
I understand it the K factor describes how much power is being coupled
to the secondary  where .15 is really 15%. 
We all read the RMS meters feeding our systems.If we could figure out at
least the 'window' of opportunity for power to flow by way of breaks per
second and then use the K factor to adjust this number it may give us a real
input value to relate spark length.

Does this make any sense or am I just up too late?