Every so often, amid the multitude of relatively vague and more complex technical indicators, bond markets just 'give us the answer.'  Granted, the answer in question isn't necessarily some epic truth that will solve all the mysteries of where rates are headed, but it does provide pretty clear commentary on the levels that are important in the short term. 

Identifying this sort of gimme is as easy as looking at the chart and noticing a level at the highs or lows of the day that has been frequently revisited and that continues to offer bounces at exactly the same spot.  If this happens to be the same high or low from the previous session(s), it just adds to the case.  Keep in mind though, when we talk about "the case" in terms of technical analysis, it's all about building a case for the significance of the levels when and if they're broken as opposed to building a case that the levels won't be broken.  It's all about knowing what's important rather than knowing the future.  Less glamorous than fortune-telling, but more realistic...

In other words, if we can see a level where rates have bounced incessantly, and if this level happens to be where rates bounced yesterday as well, we know that it will be important if rate break that level.  As you may have already guessed, one such level has revealed itself in overnight trading (i.e. you won't see it on a 2-day chart of US trading hours).  I've included equities as an overlay on this chart, simply to show the extreme levels of ongoing correlation.

So there you have it.  We begin the day in the mid-to-high 2.06 range and just barely overhead is the ominous ceiling at 2.077.  Breaking above would be bad.  The bigger and more sustained the break, the worse the implication for near term momentum.  Normally, on the Wednesday of an auction week, I'd be inclined to chalk up moderate weakness to the auction cycle as bond markets 'make room' for new debt supply.  But in this case, the aforementioned extreme correlation between stocks and bonds suggest we view this as organic, negative momentum.  That conclusion is further supported by the fact that stocks are bouncing along an almost perfectly-correlated ceiling.

Bringing this all back to MBS, the analogous level in Fannie 3.0s is roughly 101-15, though we should expect MBS to outperform on an auction week with monthly settlement ("the roll") less than a week away (prices tend to hold firmer against general bond market weakness heading into the roll).