Bond markets begin the day on the run (the bad kind of run), and if Thanksgivings past are any indication, things could get worse before they get better.  

Traders still trade on the day before Thanksgiving, but the trading environment is certainly different than an average Wednesday.  On an average Wednesday, traders know they'll have at least 2 more days in the week to do whatever they need to do before the weekend.

"The weekend" is an important wild card for bond markets because it allows the world 2 days to do "stuff" that could impact bond markets without anyone being around to trade it.  As such, traders do their best to calculate, divine, and guess at the best possible position to be in heading into the weekend.  

The Wednesday before Thanksgiving not only must serve the role of that "last day before the weekend," but it has to do even more heavy lifting, because it's dealing with 4 days off (yes, Friday is a half day, but no one really treats it that way) instead of 2.   Adding a layer of complexity is the fact that the following week is the first few days of "month-end."

All of the above occurs against the backdrop of lower liquidity.  A holiday week simply has fewer traders occupying a wider range of desired price points.  In a highly liquid market, there are eager buyers and sellers at any given price point.  In a 'Wednesday-before-Thanksgiving' market, there can be bigger gaps between eager buyers and sellers, and fewer buyers and sellers for any given price point.  (To be clear, volume and liquidity are two different things.  The simplest definition of liquidity is "volume at price")

This all adds up to the possibility of big volatility without much provocation.  After all, if there's a wide canyon between my price point and the next eager buyer/seller, one of us is going to have to jump that canyon if we need to get a trade done.  If a few trades jump those canyons in the same direction, snowball selling momentum can quickly increase, as is already the case today.