I know that there are many excellent examples being given on this thread demonstrating in exquisite detail how the melt season is progressing.
At a far more "broad-brush" level, and in response to a suggestion from Tor Bejnar, here is an updated version of a stacked bar chart showing how many "lowest 3 values for the date" are currently logged against the various years.
Here is a quick overview as to how one interprets the chart, and changes from the previous version.
1) Only 2006, 2007, 2010, 2011, 2012, 2015, 2016 and 2017 have meaningful numbers in the "lowest 3" categories, so all other years have been lumped together as "misc".
2) With the (hopefully) obvious exceptions of 2016 and 2017, each of the years has a group of 5 columns.
3) The first column in each year group shows the status at the end of 2015 - for the particular year indicated.
4) The second column shows the status at the end of 2016. The difference between cols 1 and cols 2 for each year therefore equates to the change wrought during 2016.
5) The third column shows the status as at the day/date indicated on the bottom right of the chart. The difference between this and the second column therefore indicates the impact that 2017 has had thus far.
6) The fourth column (marked "locked") indicates the number of "lowest 3" positions that have already been confirmed at the date the chart was generated. (In this case, Day 106, or 16th April)
7) The fifth column (marked "vulnerable") indicates the additional numbers that could hypothetically still be clocked up by December 31st 2017 - as long as there are no changes to positions from the equivalent dates last year. However, every additional day that 2017 has in the "lowest 3" will adversely impact at least one of the previous years.
For example, the first group of 5 relates to 2006. It can be seen that 2017 has already seen slightly more overall losses from the "lowest 3" than experienced in all of 2016. Additionally, although approximately half of the current instances have been "locked" by Day 106, approximately the same number could still be lost during the remainder of the year - i.e. they are vulnerable.
On the other hand, although 2010 lost ground during 2016, it has seen absolutely no change yet during 2017.
I hope this might help with the big picture perspective.