We have a dedicated primer on pivot points here.  But let's look at a timely example with the pivot points we've been tracking in recent weeks.  Actually, only one of the two has been actively discussed, the other is both newer and older (it was in play several months ago, took a break, and is now back in play due to last week's gains).  

The pivot points in question are 10yr yield levels of 2.215 and 2.182% (read the primer on why we look at 10yr yields on an MBS site here, if you like).  Taking these levels out to the thousandths place isn't really necessary, so you're welcome to think about them as 2.21-2.22 and 2.18-ish.  Yes, I would love it if things were more concrete and tidy, but they never will be when it comes to markets.

There's nothing too mysterious about pivot point selection.  Most of the work can be done simply by observing where yields have bounced in the past.  A few trade secrets though...  Analysts give more weight to bounces that occurred in line with big events, higher volume, and higher liquidity.  Of equal importance is the timeliness of the bounce with more recent examples being more relevant.  When recent examples line up with the high volume examples, the case for a pivot point is strong.  2.215% and 2.218% tick both those boxes.

After the big move and big bounces in mid-May, yields have returned to these pivots several times, first as a ceiling in June and now as a floor in July and August. 4 out of the past 8 sessions have seen bounces near 2.182 with today being the latest example (overnight low of 2.187).  

Here's the thing about pivot points and technical levels in general: they're most informative AFTER they're broken.  In and of themselves, they don't predict the future.  The only exception is that we can look to major pivot points from the past and reasonably assume they'll act as a speed bump for subsequent moves at the very least.  That makes recent resistance around 2.182 fairly logical.  

From here, a break back above 2.215% would signal broader resistance (think "floor bounce") at recent lows.  If we happen to break below 2.182%, it would keep the rally (yellow diagonal lines) alive.  In either case, a "break" refers to yields CLOSING on the other side of a pivot point and maintaining that breakout through the following day.  Significance is decreased by lighter volume, so we'd be hard pressed to read too much into any movement early this week.