Yeah, unfortunately it's really hard to get any really solid information on the subject for some reason. JVC's tend to have a lot of passionate fans that are quite vocal and provide "rigorous defense/promotion" of their favorite machines (let's be clear, I'm not saying JVC hasn't earned their fans, they most definitely have), while it seems like at the high end there's not necessarilly a lot of tolerance for "questioning" observations, or for doing detailed measurements.
What I mean is when the topic comes up there ends up being a lot more noise than signal in the discussion. That thread is a great example, the folks supporting the "Lumis is better" contention seem to get a little up tight that anyone would question their opinions. Of course at the same time they don't help themselves by posting confusing numbers either. Then there's the JVC side which seems to bias toward the "The JVCs numbers are higher so it must be better and it's cheaper to boot".
A couple things I'll say for my "interpretation" (using the term super loosely here):
1) Contrast performance is incredibly complicated, especially when dynamic features are factored in. It's far more complicated than 1, 2 or even 100 measurements can explain. Mark P's thread is awesome, but I think what it does best is illustrate how hard it would be to really, scientifically, explain contrast performance of a projector.
Let's just look at two of Mark's graphs to illustrate that point:
Here we have contrast measured with 12 measurements, that should be enough right? Obviously the JVC is better right?
Well, like I said, it's complicated...
2) When you throw dynamics into the picture native, "contrast" doesn't translate directly to black level.
In this case the "inferior" Sony has a lower ("better") black level at low APLs.
Then to further complicate things when you compare something with "poor" ANSI contrast and great on/off to something with the reverse (like an RS1 vs a Sharp 20K):
Now something like the Lumis just makes things all the more complicated, take a bit of everything above:
Best DI solution on the market
Best (or close to it) ANSI contrast on the market
Potentially very good Native contrast (haven't seen a number but it doesn't look way off the JVC somehow from the DB Off picture)
And you've got an outcome with way too many variables to predict based on only one specification (on/off contrast).
A thought excercise:
Supposedly the Lumis has an ANSI CR of 1000:1 (or better), and judging by the relative closeness of the DB Off picture, lets just say it's about 8000:1 native like the Sharp 20k in high contrast mode. I would expect that the Bond scene would be a high ADL to where the Lumis would probably win anyway. And even the Pioneer Logo, since the text is so bright, could be high enough APL to win there too.
Some interesting references: