Personal note: My family began celebrating (if only through cheeky verbal acknowledgement) the winter solstice around the time I became old enough to protest the vague attempts my parents briefly and occasionally made at actual religious observance.

But in 2004, my mother thoughtfully sent me a bouquet of flowers from a Brooklyn florist to amp up the holiday spirit a notch. The card said:
Dear Laurin,
Enjoy the beuty of the winners solstice.
Love, Mom, Dad, and Mickel
Since then, every December 21st has been an opportunity to bask in the beutiful glow of the winners solstice instead.
Last year on the 21st I brought you a wikipedia-summarizing description of lesser-known winter festivals, from HumanLight to Soyal to Karachun.
This year, enjoy the above-provided peak into my family's endless capacity to derive humor from an unintentionally hilarious florist's card, and a brief explanation of the movement of the "official" date to celebrate this fine holiday.

This graph, courtesy of, well, you know, apparently shows how the Gregorian calendar gradually falls behind the seasons, despite making a notable correction (the frequent spikes on the chart) each leap year. But biggest corrections (marking the beginning and end of each chunk of the chart) happen on the centennial years, which are never leap years unless they're divisible by 400—like the year 2000, in which there was no realignment between the Gregorian calendar and the seasons.
Why these seismic shifts happen in years that aren't correcting for the disparity between our 24 hour day and the variable amount of time the planet actually takes to rotate on its axis 365-day calendar and the 365.2424 days over which we experience the change in seasons [thanks, Michael!] has now joined the fact that the universe is expanding (despite being infinite! how is that possible?), in the ranks astronomical/physical/mathematical facts that just hurt to think about. Hopefully my brother will offer up some help in the comments.
Happy winners solstice!
update: Michael sez:
It takes the Earth about 365.26 days to revolve around the Sun, but […] Earth's rotational axis is not perpendicular with the radius of Earth's orbit, […] which is why we have seasons in the first place. This axis, however, does not stay still. Polaris, a well-known nautical guide, is currently directly above the Earth's northern axis, but in a few thousand years this will not be so. This wobble accounts for the difference between a sidereal year [the time it takes for the sun to return to the same position (as viewed from earth) of alignment with the stars of the celestial sphere] and a tropical year [the time it takes to pass through a full year's worth of seasons], but in this case it subtracts from the time because the axis is wobbling in a clockwise direction. Finally, we arrive with the fact that a tropical year is about 365.2424 days.
Because the tropical year is longer than the calendar year, the Earth will be 0.9696 days behind the calendar every four years. That means that once the leap year comes, the Earth will actually be 0.0304 days ahead of the calendar. Multiply this number by 25 to arrive [at] 0.76, the number of days ahead that this error will [accumulate] to every 100 years. (There are 25 sets of 4 years in 100 years.) To account for THIS error, we have common years on centurial years, but, oops, now the Earth is 0.24 days behind. Multiply this number by 4 to arrive with 0.96, the number of days behind that this error will [accumulate] to every 400 years. To account for THIS error, we have a leap year on centurial years divisible by 400 (remember that normally centurial years WOULDN'T be leap years), reducing the error down to 0.04 days every 400 years. You'll recall that we had a leap year in 2000 despite the fact that it was a centurial year. We still have some error, but at this point it's so small that we'll let the people deal with it 10,000 years from now when the equinoxes are finally off by a whole day.