Not delusional.
My model is based on a 15-entry league.

In a 15-entry league (say an "office pool"), people aren't going to take 4-5 NJ/FLA/PHX/NAS players. People probably won't even load up on NYR or STL this year (despite their success).
The teams that will get taken are: PIT, BOS, PHI, VAN, DET, CHI... teams that have won in the past. This is how the common hockey fan is thinking. People don't consult with each other and say "Hey, nobody seems to be taking NYR players, so I'll take them". People think "I have to take Crosby to have a chance... and I'll need Malkin too... and Fleury too... and I'll definitely need some Canucks." By the time they have covered all of the popular players (oh, and they'll need Ovechkin and Nicklas Backstrom, of course), they've used up so many roster spots that they aren't going to have 7-8 Rangers or 7-8 STL... and THAT is why only 3-4 players on several teams will do.

If this is a 100-entry league... yes, you need to go balls in and pick from 4 teams max. When people talk about most playoff pools... they are talking about office pools - which are maybe 15 entrants.

The OP left the league size vague.
I picked a size and went with an assumption.

It's a strategy that will work (as I said), once every six years.
If you win 1/6 years in a 15 entrant pool... you are making money.

17/20 still alive.
And the two teams I picked playing against each other: FLA vs. NJ... I guess that was a good series to be unsure about.

I didn't see this until now, but I agree fully with Pengwin, specifically for the typical office format he mentioned. Especially when the entrants are generally savvy enough to know the players, and will stack the typical teams.

You also have to keep in mind that the play offs are much more of a crap shoot, so even if everything works out perfectly as predicted (ie for me last year, I had all the winners of each match up right), picking Lucic over Krecji lost me the pool (to the only other guy that had Tbay players).

Otherwise if you actually think you 'know' the winners, you basically have to make points and game predictions and rank guys one through 20, which, yes, generally means top 10 guys are your two predicted finalists.

I'd say there are 3 styles:

-Pick the favourites/typical play off teams to not look silly
-Pick your favourite teams/some other arbitrary favoritism (its a crapshoot anyways, so have fun rooting for who you want: favourite team, Ontario players if you are Cherry...)
-Load up on a mixed bag of teams, evenly distributed al la Pengwin.

This year I went with Nash, NY, Philly, Phx/LA. I am already in first, with the most guys left.. (In our league you can replace a few guys between rounds, so could still lose).

And don't underestimate upsets.

-This year the guys that picked upsets are almost assured a win.
-Last year people that picked 'upsets' in Tampa, or SJS won (edit, along with picking Van and Bos, you only need 3-4 guys from an upset to win)
-Year before the people that picked Chi and Philly won (Wash/Pits/SJS were heavy favourites as I recall).