Injuries is the 2014-15 excuse. Other times it has been coaching, or apathy/conditioning of players.
If there is a stat somewhere tht shows the last 10 seasons and jackets records for the first 41 games and last 41 games of the season I suspect you see it more times than not it is true. Not every year, but more so than not. I tried to find it one day and couldn't (so not being passive aggressive but short of manually counting pts and games I couldn't find it tracked).
I will stand corrected if proven false - but I've been down this road several times (at least in my mind lol)
JD - Feels that way because of last 3 seasons. A mixed bag prior to that, slightly favoring better 2nd halves of seasons. The "differential" I calculated is based on points, not wins.
Year 1st half 2nd half point differential
00-01 13-22-6 15-17-9 +7
01-02 11-23-7 11-24-6 -1
02-03 15-20-6 14-22-5 -3
03-04 11-23-7 14-22-5 +4
04-05 lockout entire season
05-06 12-27-2 23-16-2
+22
06-07 16-21-4 17-21-3 +1
07-08 19-16-6 15-20-6
-8
08-09 19-18-4 22-13-6 +8
09-10 15-18-8 17-17-7 +3 (started year 12-6-2, then cratered)
10-11 20-18-3 14-17-10 -5
11-12 11-25-5 18-21-2 +11
12-13 8-12-4 16-5-3 +15 (lockout shortened season)
13-14 18-19-4 25-13-3 +13
14-15 18-20-3 16-15-1
(20-19-2) (est) +3
First 10 seasons, 6 seasons had better 2nd half, 4 seasons with better 1st half. The average difference is 2.8 points more in 2nd half than first half for the same 10 years; if you take out the outlier 05-06 season, the average differential is measurably less than 1 point differential between halves.
09-10 was the schizophrenic year- leading the division at 12-6-2 at Thanksgiving time, immediately in the tank thereafter and then a small recovery - made it feel like a huge 2nd half improvement when in fact the two halves weren't that much different in points earned.
However, these immediate past 4 seasons support JD's theory; 3 of the 4 seasons show double-digit improvement in points in the second half (with this year tbd, but extrapolated to be a +3 - and this year is slanted because the December run was actually in the first half). I think this is a perfect example of the "recency bias."