Does wine really improve in the bottle? To a large extent this depends on how the wine was handled before it ever got to the bottle. Determining whether a wine is alive or dead, much like Schrodinger's famous cat, can often be a tricky undertaking. The brutal fact is that all wine needs to be murdered, but how slow or fast this is accomplished determines whether you have a wine that will improve with time or something that is already as immutable as a piece of static artwork exhibited under glass, which accordingly may be representatuve a variety of well codified stylistic characteristics, whether they be of museum quality, (a Botticelli or an Andy Warhol) or, something you might rather tape up on the refrigerator.
The winemaker goes to great deal of trouble to 'stabilize' his product and this process is always to some extent a Holbeinian Dance of Death. Living yeast is introduced to convert sugars to alcohol but then the yeast must be killed off. This can be accomplished by any number of means from natural attrition to the introduction of anti-organics like meta-bisulfites, or what is called cold-stabilization, (literally freezing the little buggers to death). The wine itself contains some naturally stabilizing elements like tannins and acids. In some cases additional tannins are introduced, this also by a variety of means; from exposure to woody components of the grape itself like the rachis or the seeds, or through storage in a receptacle that leeches them slowly over time like barrels, or quicker methods like oak wood chips. These tannins, when 'young', tend to be harsh, like fresh idealogues without the tempering of time and experience to round and soften them. In other words, when it comes to tannins you don't want them to present themselves in black in white like the sharp edged Sean Hannity or Joe Scarborough but in nuanced variation, more like a fuzzy, elbow patched avuncular Tom Friedman or Chris Matthews who can evince everything in shades of gray.
The unfortunate fact is that wines may be overstabilized; they can smother, in other words, be filtered, fined and saturated with SO2 to the point where they will never live to improve, (of course, neither will they get any worse).
The dangers of understabiliziation stem from the fact that you will get volatile components organizing chemical revolutions from the dregs of unemployed proteins, rabblerousers, unantipated outcomes, guillotined aromatics, foul and inflammatory finishes, and if the most radical elements are given sufficient breathing room, that bane of polite society; vinegar.
In other words, as in any healthy society, in the bottle there must be a large stable constituency but one that does not completely stultify or overpower the smaller creative elements of discontent and ferment that move things forward, so the wine improves.