While they're extremely important to institutional investors and analysts, the role of trading positions in the bond market is one of the most easily overlooked factors in rate movement over shorter time horizons.  "Positions" is just a fancy way to refer to how many dollars are betting for or against bonds.  

This is particularly relevant at the moment because the only traders doing much betting in favor of Treasuries are those that are using Treasuries to hedge positions elsewhere in the market.  We can glean this from the CFTC's weekly position report, which breaks Treasury positions into commercial and non-commercial categories.  In other words, non-commercial positions tell us the most about how speculators are betting on Treasuries.  To say that they're short would be a massive understatement.

The side-effect of such a positional imbalance is the risk of "short-covering."  This occurs when a short position (someone who is betting on yields rising, in the case of the bond market) hits a certain target or time limit and exits that position.  A short position is exited by buying bonds.  This buying, in turn, can cause trading levels to hit a short-covering target for another trader, and the cycle can snowball.

In addition to those sorts of snowballs of stop-loss short-covering (i.e. traders buying bonds to close short positions because bonds have improved by a certain amount after running higher in yield for the most part), there are also simple calendar considerations when it comes to exiting positions.  One of the most common situations is the 3-day weekend (like the one beginning tomorrow!).  Traders are often more likely to exit positions ahead of a 3-day weekend, especially if those positions have had a good run and especially if other markets will be open on the associated holiday. 

This weekend meets both of the above criteria.  Thus, a Friday morning rally today is almost certainly getting a lot of support from the massive short base exiting some of those positions ahead of the long weekend.  The point is that the rally doesn't necessarily represent organic buying demand in bonds.  

All this talk of short-covering aside, keep the prime directive in mind.  This bond market needs to show us a certain amount of commitment before it makes any sense to be optimistic about a consolidation, reversal, or temporary bounce.  The following chart lays out a few lines in the sand.