FILLING OUT TWO NCAA BASKETBALL BRACKETS: PART 1

I’m in two pools this year, the first is a standard 1-2-4-8-16-32 pool and the second is a 1-2-4-6-10-14 format. That changes the strategy a lot.

As always, winning this bracket is about trying to pick the best possible outlier. Unless you’re in a pool with only a handful of people, picking the best average score does you no good. You want to go for the outlier that gets you in the top few of your pool if you get lucky. I’ll start with the traditional ESPN format.

10-20-40-80-160-320:

The most important round to win in this one is the champion. If you don’t get the Champion you probably aren’t winning. It also helps if you don’t take the team that everyone else is taking. If Kansas, Duke and Ohio State all have about the same chance of winning, you’re better off taking the least popular and hoping you get lucky.

So, I made a chart of marginal points to win*

*If you care, marginal points is a team’s (Chance to win)*(Points for round)*(% of ESPN contestants that picked the team to win)+(Chance to lose)*-(Average of all other teams points for round)*(1-% of ESPN contestants that picked the team)

To win it all:

Rk 

Team 

Pct picked 

% to win 

Marg. Pts 

1 

Duke  

15.00% 

12.70% 

31.257 

2 

Kansas  

22.80% 

13.25% 

29.790 

3 

Ohio State

26.30% 

12.39% 

26.351 

4 

Pittsburgh  

6.20% 

8.47% 

21.424 

5 

North Carolina  

4.10% 

4.01% 

7.820 

6 

Texas  

1.80% 

3.81% 

7.359 

7 

San Diego State  

1.80% 

3.81% 

7.359 

8 

Kentucky  

1.90% 

3.81% 

7.351 

9 

Notre Dame  

3.80% 

3.46% 

6.110 

10 

Florida  

1.30% 

2.99%

4.721 

I’m guessing I’m in a pool of 50-100 people so I think Duke is the way to go. If you’re trying to win the ESPN tourney challenge, I wouldn’t even bother with the top seeds. Even if you get it right, the odds of the rest of your bracket being good enough to win are not very good. I think San Diego St. and Kentucky are the way to go in super large pools. For me, I’ll stick with Duke again.

The next most important round is not surprisingly, the National Runner up. here’s the same chart of that:

Rk 

Team

Pct picked 

% to win 

Marg. Pts 

1 

Duke

24.20% 

20.79% 

22.836 

2 

Pittsburgh  

17.50% 

17.81% 

20.722 

3 

Kansas  

47.40% 

25.25% 

19.783 

4 

Ohio State  

39.20% 

20.52% 

18.038 

5 

Notre Dame  

12.00% 

8.46% 

8.224 

6 

San Diego State  

4.00% 

7.80% 

7.894 

7 

North Carolina

7.40% 

7.80% 

7.615 

8 

Kentucky  

3.50% 

7.42% 

7.329 

9 

Texas  

3.70% 

7.42% 

7.314 

10 

Florida  

4.20% 

7.05% 

6.667 

It’s pretty much a toss-up between Kansas, who is better, but super-popular, and Pitt the No. 1 seed in by far the weakest region in the bracket. I’ve already taken Duke to win, so that leaves me with Pitt. Trying to win ESPN’s overall pool, I like the Aztecs if you haven’t already taken them, Notre Dame could be a good option, but overall I’d slide down to the 11 spot (unlisted) and go with Purdue. If Wisconsin didn’t have an absurdly hard pod, then I would recommend them, too, but because of that they come in at 16 on my list. Pitt seems to be the No. 1 seed most people are picking to lose early, so how about a National runner up?

On to the rest of the final four, see a pattern developing?

Rk 

Team 

Pct picked 

% to win 

Marg. Pts 

1 

Duke

48.50% 

37.41% 

14.337 

2 

Pittsburgh

49.70% 

32.73% 

11.958 

3 

Ohio State  

58.90% 

36.35% 

11.064 

4 

Kansas  

60.20% 

36.16% 

10.650 

5 

San Diego State  

11.90% 

17.74% 

9.324 

6 

Purdue  

7.40% 

16.03% 

8.397 

7 

Notre Dame  

20.00% 

17.34% 

8.185 

8 

Kentucky  

8.20% 

15.82% 

8.148 

9 

Florida  

15.10% 

15.90% 

7.596 

10 

Texas  

10.50% 

14.96% 

7.263 

11 

Syracuse  

12.50% 

14.86% 

7.017 

12 

Louisville  

8.50% 

12.68% 

5.564 

13 

North Carolina  

16.30% 

13.14%

5.432 

14 

Connecticut  

25.00% 

13.24% 

4.931 

15 

Wisconsin  

6.70% 

10.43% 

3.796 

16 

BYU  

11.00% 

9.27% 

2.696 

17 

Kansas State  

5.80% 

6.68% 

0.657 

18 

St. John’s  

6.10% 

5.96% 

0.050 

19 

Washington  

1.10% 

5.42% 

-0.427 

20 

Georgetown  

1.50% 

4.47% 

-1.269 

Duke and Pitt are already off my bracket, so I’m looking for someone out of the East and Southwest. Ohio State and Kansas would fit in nicely, but I’m going with Kentucky and Purdue. I’m not sure what size pool going away from the No. 1 seeds is the best idea, but I’m hoping mine is big enough. As far as going for the ESPN bracket, with SDSU or Kentucky as your winner, and Purdue as the runner up, I think Kentucky and SDSU should be in the final four. It pains me to say it, but Florida might be the pick as the 2 seed in the SE. Wisconsin would be a perfect option out of this region, but they really get screwed by having such a hard first two games for a 4-seed. I’d tend to lean towards the more dangerous Wisconsin because they are less popular and I like them better than Florida.

Moving on to the Elite 8: (I couldn’t find odds so I switched to KPom’s log5 projections)

Rk 

Team 

Pct picked 

% to win 

Marg. Pts 

1 

San Diego State  

32.00% 

47.10% 

11.724 

2 

Purdue

31.10% 

41.80% 

10.186 

3 

BYU  

27.00% 

35.40% 

8.596 

4 

Texas

19.40% 

32.00% 

8.187 

5 

Syracuse  

42.90% 

35.20% 

6.670 

6 

Duke

73.10% 

56.50% 

5.789 

7 

Louisville  

16.60% 

25.20% 

5.740 

8 

Florida  

39.50% 

29.40% 

5.392 

9 

Notre Dame  

55.80% 

36.20% 

5.372 

10 

Ohio State  

78.10% 

62.00% 

5.250 

11 

Pittsburgh

70.00% 

47.40% 

5.214

12 

Kentucky

16.50% 

23.10% 

4.894 

13 

North Carolina  

45.20% 

28.60% 

4.673 

14 

Kansas  

79.20% 

52.40% 

4.090 

15 

Wisconsin  

12.80% 

20.40% 

3.958 

16 

Washington  

5.70% 

19.70% 

3.956 

17 

Connecticut  

60.60% 

22.10% 

2.117 

18 

Utah State  

1.40% 

11.90% 

0.320 

19 

Cincinnati  

2.70% 

11.80% 

0.267 

20 

Gonzaga  

3.00% 

11.10% 

-0.074 

Now, you’ve got to use your head a bit. I want to throw out BYU and Syracuse because of potential regional biases and BYU is probably overrated slightly by this anyway, but Cuse is the only remotely good option in the bottom half of the East, unless you’re trying to win ESPN, then take the Huskies. I doubt San Diego St. is quite likely to advance to the Elite 8 as KPom overstated their chances in winning, getting to the Champ and making the Final four, but they are still the way to go. Again, KPom seems to be overstating the case for Texas, and I think you pick up more by taking Duke out of this region in more important rounds, so I’m going to ignore that. In the Southwest, it’s hard to go against Louisville. KPom has Kansas at only around a coinflip to make the Elite 8, and he was close on all the other odds, so with nearly 80% of the people taking them, they aren’t worth it if you don’t already have them winning the title. In the Southeast, Florida is really the only good option for my purposes, where Jimmer should be overrated. On ESPN, I think taking BYU to the elite 8 is probably the way to go.

Rk 

Team 

Pct picked 

% to win 

Marg. Pts 

1 

Brigham Young  

50.20% 

55.70% 

4.896 

2 

Texas  

65.50% 

67.90%

4.448 

3 

Washington  

15.60% 

37.70% 

4.180 

4 

Purdue  

69.70% 

68.00% 

3.914 

5 

Florida

71.50% 

57.80% 

2.956 

6 

Wisconsin  

44.60% 

38.50% 

2.869 

7 

Kentucky

74.70% 

61.00% 

2.830 

8 

Louisville

76.20% 

60.80% 

2.650 

9 

Syracuse

78.80% 

62.20% 

2.435 

10 

San Diego St.

82.50% 

70.10% 

2.349 

11 

Cincinnati  

7.70% 

28.70% 

2.170 

12 

Notre Dame  

86.70% 

66.60% 

1.673 

13 

North Carolina  

81.40% 

52.40% 

1.668 

14 

Utah St.  

10.40% 

26.90% 

1.629 

15 

Pittsburgh

91.50% 

76.50% 

1.269 

16 

Connecticut  

85.50% 

48.30% 

1.142 

17 

Duke

94.10%

84.50% 

0.988 

18 

Ohio St.  

94.50% 

83.20% 

0.905 

19 

Arizona  

26.00% 

24.10% 

0.725 

20 

Kansas  

96.00% 

72.60% 

0.561 

Let’start with the Easy ones. Texas losing to Duke, Washington to the Sweet 16, (though KPom probably overstates their case), Ohio State, Wisconsin, Notre Dame, Kansas. That leaves BYU’s pod, where I would guarantee that they are more than 50% picked in my pool. Even at a 75% clip, I still feel like they are better than the next option, Gonzaga, so BYU it is. Finally, incredibly unpopular on ESPN Cincinnati rounds out the sweet 16.

Finally, let’s go matchup by matchup for the rest of it. George Mason is only slightly better to the point that it basically doesn’t matter in terms of marginal points. Feel free to pick this one by uniform color or mascot if you like. With 82% of people taking West Virginia, there is no point in not picking Clemson for the upset. Xavier and Marquette is another complete toss-up, so I’ll say Xavier, the favorite. North Carolina is actually a –EV pick, but I don’t want to give up the 9.6 expected points you’re getting in that matchup. UCLA is an underdog as a seven seed, but Michigan State is a little too popular. Gonzaga is an easy upset in basically a toss-up game against much more popular St. John’s. Old Dominion is an easy pick as an unpopular favorite. I’ll take UNLV in the 8/9 game in the Southwest, as well as USC/VCU against No. 6 Georgetown though that is a complete guess. In the two final games in that region, I’m going with, Richmond and Texas A&M. Michigan, Arizona, Bucknell and Temple round out my bracket.

Part 2, with a different scoring system will be up later tonight/tomorrow morning.

Advertisements

4 Comments

Filed under Uncategorized

4 responses to “FILLING OUT TWO NCAA BASKETBALL BRACKETS: PART 1

  1. Yahoo’s pick distribution is a bit more lopsided, with OSU at 31%, Kansas at 25.1%, and Duke at 14.7%:

    http://tournament.fantasysports.yahoo.com/t1/group/all/pickdistribution?round=6

    Not sure how much the general public cares about Irving possibly returning, so my pick is Duke as well unless things change dramatically overnight.

  2. Matt

    I’m trying to figure out precisely what formula you’re using for the marginal points, in order to fill out a Sweet 16 bracket. Pick distribution : http://tournament.fantasysports.yahoo.com/t2/group/all/pickdistribution?round=6

    I think I’m miscalculating the “Average of all other teams points for round”. In your above example, would this mean 320 (since every team would get 320 for winning), or does the 320 get reduced by the % chance of winning?

  3. Have the spreadsheet I used on a different computer, but I think the formula I ended up using for this was:
    ((total points available)-(expected points by picking for example Duke))/(1-teams that could win)

    So for the standard ESPN Bracket, and giving duke a 10% chance of winning it would look like this:
    (320-32)/63*

    *All the play-ins were treated as combos and one option.

  4. Matt

    Great, thanks. I’m in a non-standard pool where scoring is based on historical seed performance, so I had to do some weighting to get the total points available to work out. If I did the calculations correctly, in a pool like this, the 4 top teams in marginal points are: tOSU, Duke, Kansas, Richmond.

    I ran the calcs for a standard flat 320 scoring structure and the only three marginal points that were positive belonged to tOSU, Kansas, and Duke, which I guess isn’t very surprising.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s