[morzh]

I am not fully sure about the explanation on Chuck's site, but I think the representation of the equivalent circuit for the displaced lead is off.
It has no practical ramifications or Kirk, so, Kirk, you may disregard this post.

However, speaking of the circuit, by the same token (1/3 and 2/3 inductances in parallel serialized with the full cap value) if we keep the cap with the lead barely displaced (say 1/100 and 99/100 length division), then we will have reduced the inductance by 99%, creating practically no ESL cap. Better yet, no displacement at all. Leave the pin at the beginning.

I think, a still crude equivalent should have each L (1/3 and 2/3) inuctor connected with the respective partial cap in series and then two of them paralleled. This will take some calcs to figure the resonances and impednces, but will be closer to the truth.