A few notes that may or may not add some color:
In terms of the yield curve in general, for the past three or four months we have resorted to increasing the sampling timeframes in our charts in an effort to filter out some of the 'false' breakouts. We are really stepping back and trying to get a bigger picture of things, and we are also trading smaller size and hanging on to positions for days instead of hours without adversely impacting P&L metrics. Another note; we think that the Eurex curve has been "better behaved" in general than the CBOT curve as of late.
All of these spreads go through periods where they seem to 'behave', and periods where the price action is downright mystifying. This is where we will also be looking at grain or metals or energy spreads for some action - and in the end, in a twisted and convoluted way, it ends up making us better curve traders in the sense that we end up being more selective about the curve trades we do take.