The common wisdom with any team going into the playoffs is that you don't want to peak too early. Many factors can contribute into this: injury, chemistry, strength of schedule and just plan luck. If players were robots (or at least cyborgs, like the Terminator), then the mental aspect of momentum wouldn't come into play -- have a bad game, shake it off and treat the next game as brand new.
However, we're dealing with real people here with varying degrees of mental toughness. Team history and personal history (and personal demons) can creep into one's psyche, causing doubt. We've seen it all too many times -- when a player or team's confidence drops, everything falls apart, but one good break and that belief comes back. The legs pump harder, the shots seem crisper, the timing gets better.
Thus, when you're looking at the last ten games of the season, there are two schools of thought. The first is that you want to be playing at your best with the highest levels of confidence so that successful play is instinctive rather than a work-in-progress. On the other hand, some people think that the playoffs are truly a second season and whatever happened prior is inconsequential.
I tend to believe more of the former than the latter. Of course there's no absolute in pro sports, that's why the games are played. But I'm a big believer in habits on the ice, both good and bad. During losing streaks, teams get into bad habits -- defensemen chase the puck, forwards try to stick-handle through everyone and goalies get overwhelmed. Even though those are all correctable issues, it all comes from individuals trying too hard and not sticking with the team concept.
To test this belief, I looked at the last three years of Stanley Cup playoff action. More specifically, I broke down winners and losers in the first round and divided it into teams that streaked and slumped into the playoffs.
The dividing line for that was 12 points; 13 or more counted as a streaking team while 12 and under counted as a slumping team. (Of course, 12 out of 20 points isn't exactly a slump, but consider the high level of play it takes to make the playoffs and this seemed like a reasonable cut-off point.)
Here are the results:
10 teams over, 6 wins
6 teams under, 2 wins
8 teams over, 7 wins
8 teams under, 1 win
8 teams over, 4 wins
8 teams under, 4 wins
26 teams over, 17 wins (65.4%)
22 teams under, 7 wins (31.8%)
When a team is playing better during the last ten games of the season, this data shows they win about 2/3 of the time, while you get the inverse for the 22 teams that didn't play as well. Here is the data broken down by season (tables have both Western and Eastern teams in there broken down by conference rank).
What's interesting here is that it doesn't necessarily depend on point totals. You have some very high seeds sliding into the playoffs before getting knocked out (see: San Jose, 2008 or Washington, 2010) and vice versa (see: Philadelphia and New York, 2008).
Of course, there's no exact predictor for playoff results, but this snapshot of data does seem to support the belief that it really does matter if a team is playing with good or bad habits at the end of the season. Unfortunately, coaches don't have a magical button to wipe a player's memory. Those actions have to be trained out of him and sometimes it's too little, too late.