It's nearly impossible to pick a perfect bracket for the Men's NCAA Basketball Tournament.  But, there are ways you can improve your chances of picking the best possible bracket.  

We looked deep into the history of March Madness to find our 15 top tips for picking a winning bracket.  All records referenced are from the modern era - after 1985 - unless otherwise noted.

Tip 1: Pick #1 & #2 Seeds to Win in the First Round

This is pretty standard.  No top seed has ever lost to a #16 seed and just 7 of 120 (5.8%) #15 seeds have pulled off an upset over a #2 seed.

Tip 2: Take One #13 or #14 Seed to Advance to the Second Round

These seeds have a combined winning percentage of 17.92% in the first round.  That's just over one per year and it has happened at least once in 25 of the last 30 tournaments.

Tip 3: In the First Round, 12 is a Magic Number

Last year three of the four #12 seeds upset their 5th-seeded opponents!  While not typical, at least one #12 seed has defeated a #5 seed in the first round in 24 of the last 26 years.  If that's not convincing enough, #12 seeds have won 36.7% of their first round games since 1985.

Tip 4: #9 Seeds Win More Often Than #8 in the Opening Round

Don't go picking all #8 seeds to win in the first round.  The #9 seed holds a two-game advantage all time (61-59) over the #8 seed.  You should half 8's and half 9's.

Tip 5: #1 Seeds Win in the Second Round, Too

Top seeds are almost always the right choice in the Round of 32.  Since 1985, #1 seeds are 104-16 (86.7%) in the second round.

Tip 6:  In the Second Round, 12 is Still Pretty Magical

#12 seeds that advance to the second round do pretty well.  They are just under .500 at 20-24.  Keep in mind that #12 seeds have reached the Sweet 16 the same number of times as #7 seeds since 1985.

Tip 7: The Sweet 16 Won't Be By-The-Numbers

At least one #2 or #3 seed is likely to be upset.  Only once in the last 30 years has the Sweet 16 been made up of only #1, #2, and #3 seeds.

Tip 8: Your Sweet 16 Picks Should be Seeded #12 or Better

An overwhelming number of teams that make the Sweet 16 have been seeded #12 or less (98%).  A decent number of #10-#12 seeds do advance.  Anything higher and you'd be looking at a serious long-shot.

Tip 9: Take Two #1 Seeds to the Elite 8

There was a pretty good chance you'd look at doing this anyway.  You can feel pretty good about it as 68.3% of all #1 seeds since 1985 have made it at least to the Elite 8.

Tip 10: The Elite 8 is Where the Cinderella Story Ends

Once you get to this round of the tournament, it's time to stop taking huge underdogs.  Just 1 out of 240 Elite 8 teams over the last 30 years has been seeded #11 or worse.

Tip 11: Take at One or Two #1 Seeds to the Final Four

In 24 of the last 30 years of the tournament, the Final Four has consisted of exactly one or two top seeds.

Tip 12: Keep Any Team Seeded #10 or Worse Out of Your Final Four

Out of 144 Final Four teams since seeding began in 1979, only three have been seeded worse than #9.

Tip 13: Take Seeds Under #6 to the Championship Game

Ok, so last year this would have been a huge bust. A #7 seed played a #8 seed in the championship game.  With that being said, if history is any indication of what to expect, that's not going to happen again.  Out of the last 29 tournaments just three teams ranked #6 or worse have played for the championship.  And that includes the two from last year!

Tip 14: Don't Take Two #1 Seeds to the Final

There have been 36 tournaments since seeding was introduced in 1979.  Only six times (16.7%) has the championship game included a pair of #1 seeds.

Tip 15: A Team Seeded #4 or Better Should Win the Tournament

Before last season it had been 25-straight years since a team seeded worst than #4 won the championship.  Yes, it's certainly possible it could happen again.  But, the smart money says you'll see a very low seed win this year's tourney.

Final Thoughts

You shouldn't ignore the majority of these tips.  You have to remember to take a few chances to set yourself apart from the rest of the brackets you are up against.  

If you are in a small contest, play pretty conservatively.  If you don't have much competition you want to play the historical numbers since it's unlikely many of the other entries will match your picks. 

If you are playing in some huge contest like ESPN or Yahoo your strategy changes.  You'll need to stray from some of these tips as there could be many entries that take your same approach. 

Good Luck!