The Klippel test contains a lot of data, and I think we often get hung up on some of the numbers, mis-interpret what some of the graphs mean, and fail to realize what Klippel does not measure.
For example, consider the BL graph. Clearly extending the graph in the negative direction would yield [-6,2.6] for the range, which is about 4.5mm one way (if centered). However, we do see the protection bars creeping in for both the suspension and BL! Meaning that the Klippel did NOT collect enough data in that range to accurately calculate the values in that range - it was somewhat limited.
We also are reminded of this in that the Klippel did NOT find limits for Cms or Le during its measurements - an incomplete set of data was captured, and we know that by the listing of "> 4.0mm" for those limits. It's somewhere above that - we don't know where, though. That comes from not enough data captured.
A bit of an aside - a Klippel system does NOT measure BL, Cms, and Le! It measures voltage, current, and position (with SPL, optional) and back-calculates BL, Cms, and Le curves which would fit the measured data. (1)
Beyond that, though - the issue of centering. And this relates to what Klippel does NOT measure. Klippel cannot measure frequency versus inductance versus position. It shows you a BULK inductance, average over bandwidth for position - but not frequency-dependent (and further, that bandwidth is quite limited, usually just up to 500 Hz or 1 kHz)! Inductance changes with position, power, and frequency - it is possible to have increasing inductance at low frequencies with rearward motion, and decreasing inductance at high frequencies with the same rearward motion! This is some of the constraints of what a Klippel measures.
In the case of the VWR126X, the emphasis is on the midrange/wideband nature of the driver. Yes, it actually does have a good amount of mechanical - and BL based - throw, and can credibly reach low enough to combine with larger midranges, and low crossover points.
For by offsetting the voice coil back, we actually linearize inductance in the 250-8 kHz range! That is part of the trick in the very low THD and IMD values (especially odd order) for this speaker.
So the tradeoff: center the BL curve, give it a few mm "more" stroke for bottom end use, or purposefully set the voice coil off center and make the inductance more linear over frequency. For a woofer, the choice is obvious - more stroke please! For a wideband midrange (which is what this is), the choice is a bit less clear, but was made to prefer inductance over raw stroke.
Many times you'll see "off-center" speakers from some very high end brands. Lots of times people are concerned or curious about what they see - why would such prestigious names allow so much off-center production of their drivers? Well, sometimes the off-center nature is a compromise to improve other aspects of the speaker. It's pretty simple to build speakers in volume with centering as tight as 0.3mm (I'm doing it now with a high stroke woofer, with productions of 15,000 pieces a month). But what IS the correct center? It's not what Klippel says it is - it is what the driver designer decides is needed for the application - the tradeoffs to be made.
