To expand on what I said to Oswaldo, my experience has always been that the best old bottle from the case was the one with the tightest seal. But the bolus of oxygen at bottling may also be important. I think corks give a nonlinear delivery of oxygen that hasn't really been replicated with other closures.
From what I've seen in the long-term studies, the best cork seals do not, in fact, allow any oxygen transfer from outside the bottle ("best" here being defined is multiple ways: smallest loss of wine, plus most positive sensory response). The necessary oxygen for aging is what's in the bottle at bottling plus what little passes out of the cork; that wine needs some mystical and too-variable flow of oxygen around/through the cork to age is a myth, and studies have shown this. For those not inclined to pore over the long-term studies the quick-and-dirty demonstrations are, 1) as you say, the generally superior performance of the lowest-ullage bottles, and 2) the quick failure of synthetic corks, which is due to too-rapid oxygen transfer. Corks do exactly the same thing as synthetics, in this way, they just do it a lot more slowly...and in the best cases, slowly enough for it to not matter to us.
Apparently, the little pocket of oxygen (plus dissolved oxygen) at bottling that's OK for cork-finished wines is not always enough for screwcaps with the stronger type of liner (which is, after all the actual closure in a screwcap or crown cap system), and in some cases those wines thus show a predilection for reduction. And there's a regular but too-quick oxygen transfer with the other type of standard liner, which few use. The liner can apparently be adjusted to allow just enough oxygen transfer to avoid the peril of reduction, but not enough to adversely affect the wine's aging curve; one would presume that this would also shorten that aging curve -- which is similar to the very best corks, and thus sometimes deemed "too long" by critics, though I don't hear them complain about the identical problem under corks -- by some measurable amount. Again, though, no one really knows what this rate of transfer is, and since it's unquestionably different due to each wine's differing potential to reduce under seal, a lot of research is required. It will probably take decades.
Anti-screwcap folk keep insisting that aging trials are necessary. They're not, really. They've already been done, and successfully, though some might want them expanded a bit for sufficient rigor. What's necessary now is an answer to the trickier problem of crafting wine-specific liners so that Paso Robles syrah ages like Paso Robles syrah, but demi-sec Vouvray ages like demi-sec Vouvray. That is, I think, a long way from being answered.
