ii) Test number two:

Consider a parallel 4-bar, with upper and lower links parallel to the ground.  As noted in the “The Natural Mirror Bike” section, if we imagine moving the IC back, by moving the two forward linkage pivots together, until they are coaxial and at the same height as the rear axle, producing a pp-coaxial 4-bar, we will have an identical situation to the parallel 4-bar.  The parallel 4-bar IC is located as far out in front of the bike as you can get it.  The pp-coaxial 4-bar has an IC located quite far back.  (If one wants to object that the parallel 4-bar has no IC, then just imagine moving the two forward linkage pivots together a distance of one angstrom.  One will then have a bike with a remote, forwardly located IC, that does not differ significantly from the pp-coaxial 4-bar.)

PA correctly states that the two situations are identical.

ICT claims that an IC moving far out in front of the bike provides greater efficiency in a wider range of gears.  Clearly this is not the case for the two 4-bars under consideration.

In addition, IC location and movement have no direct correlation to the performance of a wp-coaxial 4-bar (beyond suspension rate), including efficiency over a range of gears.

Since the Ellsworth bikes are essentially mono-pivots under pedaling, we see that they are no more efficient then typical mono-pivots.

PA thus passes this second test, while ICT shows a third flaw: arbitrary nature.

This second, and most disturbing, Ellsworth failure is the fundamental misconception of what an IC is and what it does.

As we have noted, an IC moves relative to the front triangle as the suspension moves.  A pivot does not.  As a result, the IC does not control frame motion in quite the same way that a pivot would.

But Ellsworth and Ellsworth's consulting engineer Mike Kojima view the IC as a pivot.  This is clear from the following statements Kojima makes in critiquing the early PA (the PA statements are in black, while Kojima's statements are in red:

Within any small segment of any non-URT suspension's travel, that suspension will behave like a mono-pivot, with pivot located along the line perpendicular to the tangent of the wheel path relative to the main triangle.  That is, pedaling a multi-link at any particular position in the travel, at equilibrium for example, will be like pedaling a particular mono-pivot.

This is not true at all and is the point where the author errs.  By multi-link he has to mean a true multi-link with the pivot below the axle.  A pivot above the axle makes a multi-link a single pivot bike.  A true multilink is actually a single pivot also, the single pivot being the instant center.  The beauty of an IC bike is that the pivot can be placed in a less compromising point due to that point not being controlled as much by the frame packaging, because it is a virtual pivot point in space.

Accounting for friction and suspension rate, the reactivity of all non-URT suspension types will increase by practically the same magnitude as the gearing varies from ideal.  That is, no geometry will be significantly more efficient in a wider range of gears then any other (though if the ideal is in the middle of the gear range, such a design will have a better average performance).

This is wrong, the non-reactive point can be made to include a bigger range of gears when a very long virtual swingarm can be made by placing the instant center well forward of the bike.

The first Kojima statement is somewhat amusing, since he first states that PA is wrong in saying that any non-URT (4-bars in particular) will behave like a mono-pivot through any small segment of the rear axle path, but then goes on to claim that “A true multilink is actually a single pivot...”

But more to the point, Kojima clearly believes that an IC far out in front of the bike acts like a pivot, producing “a very long virtual swingarm”.  The parallel 4-bar definitively demonstrates that this is false.

Test number three:

PA says that, at any point along the rear axle path about the main triangle, the tangent to the path is what determines the initial response of a suspension to pedaling.  This means that there are an infinite number of IC locations, along the line through the rear axle and perpendicular to the path tangent, that will produce the same initial results.

ICT is in direct conflict with PA, claiming that each IC location, along the line perpendicular to the path, gives a different result in balancing squat with anti-squat. 

Again, IC location and movement have no direct correlation to the performance of a wp-coaxial 4-bar. 

PA's claim that, for a given path tangent, IC location does not matter in a 4-bar thus properly characterizes the wp-coaxial 4-bar.

ICT's claim that one IC location is preferable to others clearly improperly characterizes the wp-coaxial 4-bar.