Is a #1 seed really more preferable than a 2 seed?

Submitted by Piston Blue on February 23rd, 2019 at 8:27 PM

With so many impactful games going on this weekend in college hoops, it got me wondering about the race for the #1 seed, and how matchup wise it compares to other seeds in the NCAA tournament. Considering the current lay of the land after today's results, Michigan is likely the top 2 seed. Here are some interesting facts about the last 10 NCAA tournaments:

 

Traditional path of the 2 seed:

3 seeds advance to the sweet sixteen 55% of the time

7 seeds advance to the second round 60% of the time

Average seed of sweet sixteen opponent: 5.725

 

Traditional path of the 1 seed:

4 or 5 seeds advance to the sweet sixteen 85% of the time

Second Round matchups must include either an 8 or 9 seed

Average seed of sweet sixteen opponent: 5.525

 

Since top seeds are most advantageous for matchups from rounds 1-3, do you think it's more advantageous to be a 1 seed or 2 seed? The best argument for the 2 seed seems to be that 3 seeds only make the sweet sixteen 55% of the time.

Comments

Piston Blue

February 23rd, 2019 at 8:37 PM ^

Definitely agree with your first point, but the main benefit from being a 1 seed is playing worse opponents. However, over the last 10 years, this hasn't really been the case. 2 seeds more often than not have a lower-seeded sweet sixteen opponent than 1 seeds, making the path to the elite eight a little easier. 11s often outperform their seeding, and make the sweet sixteen at a surprising rate. Since a 1 is almost guaranteed to see a 4 or 5 seed in the sweet sixteen, is it worth it to just get the 2 seed and have a 45% chance of seeing a 6 seed or worse?

Indy Pete - Go Blue

February 23rd, 2019 at 9:35 PM ^

This is an excellent commonsense question. Although, you could argue that one seeds advance way more often because they are better teams, not because their preferential seeding placed them in a much better position compared to the two seeds (per all the academic arguments provided earlier here). 

 I don’t usually consider Mineral King the seed of seers, but his comment was pretty solid. 

BlockM

February 23rd, 2019 at 9:35 PM ^

Because 1 seeds are usually better teams? If you give a worse team an easier schedule, that doesn't mean they'll win more than a better team.

Not saying I'd rather have Michigan be a 2 than a 1, but presenting that stat that way implies the teams are the same quality.

I think the bigger question would be what's the percentage of final fours for the fourth 1 seed vs the 1st 2 seed.

Indy Pete - Go Blue

February 23rd, 2019 at 9:47 PM ^

Block M,  I think you are absolutely framing the question the way the OP intended. If the fourth and fifth best teams in the tournament are completely equal, is it more favorable to be the fourth one seed or the best 2 seed?  I have not seen a compelling answer either way, but once again, I circle back to mineral king’s excellent response on this one. 

NotADuck

February 24th, 2019 at 12:18 AM ^

I think that says a lot more about the usual disparity between the top 2 or 3 teams in college basketball every year and everyone else, not necessarily that 1's are more likely to make it to the Final Four than 2's.  Every year there are 1-3 teams that are clearly better than everyone else and those teams typically make it to the Final Four barring a MASSIVE upset.  Examples:  2018 Villanova, 2017 UNC, 2015 Duke & Wisconsin, 2013 Louisville, 2012 Kentucky, 2010 Duke, 2008 Kansas & Memphis, that Florida team that returned all their starters from the previous national champ season... etc.

The great teams are very consistent.  Duke probably makes it to the Final Four this year.  Everyone else is a coin flip or worse to me.

wesq

February 24th, 2019 at 10:30 AM ^

Sure 1 seeds being better is a large part of that stat but that’s such a huge gap it makes it very hard to make the case that the 2 seed road is easier, which was the hypothesis of the OP.  I also think with the lack of depth of this team seeding and luck does matter.  I think getting the starters rest on Thursday/Friday games will be big on those quick turnarounds. 

stephenrjking

February 23rd, 2019 at 8:34 PM ^

Yes. You might have a better chance at getting a "cinderella" opponent that's an easier out in the 2 bracket, but the potential caliber of opponent is higher. 

The goal is to win and not lose. You want to avoid land mines. Land mines tend to have more talent and/or quality. 

It's also preferable to face a 2-seed caliber team than a 1-seed caliber team when regional final comes along, should it come to that. 

 

Squad16

February 24th, 2019 at 1:58 PM ^

This is just wrong.

Michigan's best post season home field Tourney advantage in modern history was in Los Angeles last year at Staples.

We have massive amounts of alumni in large coastal metropolitan areas, which tend to not have much competition for great teams.

DC this year is a FAR better regional for us than Louisville or Kansas City. And I'd argue Anaheim is second best for us.  Louisville might be the worst regional for us even though it's theoretically closest; we have relatively no alumni in Kentucky and it will have tons of competition from schools that are closer (Tenn, Kentucky, etc.). 

stephenrjking

February 24th, 2019 at 2:57 PM ^

The competition issue could be an issue in Louisville. TN or KY would be a serious ticket problem, and given their records, at least one is probably going there. 

Otherwise the distance makes it a good choice, because lots of Michigan residents can make it. As many as might show on the coasts. 

So we’ll see. 

My main concern is getting to Minneapolis. I’ll go if Michigan does. 

Goggles Paisano

February 23rd, 2019 at 8:35 PM ^

I never thought it was a big deal or a big advantage to be one over the other.  The first round is easier and sometimes quite a bit easier, but other than that...meh.  Nice job with the stats.  

Mr Miggle

February 23rd, 2019 at 8:43 PM ^

If you go one round further I bet the numbers swing back to favoring the 1 seeds. There's really no reason to stop at the sweet sixteen. The biggest advantage of being a 1-seed is that you never have to face one in your region.

Piston Blue

February 23rd, 2019 at 9:05 PM ^

Right, but the 2 seed is also the top seed in its mini region. You definitely have a point, being a 1 seed does give you an advantage if you make it to the elite eight because there's 3 chances each for both the 2 and 3 seeds to fall, but 1 seeds are more vulnerable to losses during that period because they're playing a higher seeded team on average. Since seeding is so random at that point in the tournament, I don't think seeding really matters at that point. For reference, here are the stats for typical elite eight matchups:

Average seed of elite eight team from 1's side: 2.775

Average seed of elite eight team from 2's side: 4.075

Michigan Arrogance

February 23rd, 2019 at 9:00 PM ^

at the 1-2-3 seed line, it's about what upsets down the road you'll have - IOW, what top teams/matchups you can avoid.

This year, I'd avoid Duke assumine ZW is playing.

I wouldn't want UNC again either - they can score like almost no other team. Basically the top ACC teams are legit. can we avoid those teams until the final? That's bracket luck IMO

Mike Damone

February 23rd, 2019 at 9:01 PM ^

This year - May not make much of a difference.  I think the top 7 teams in the country are Michigan, NC, Kentucky, Duke, Virginia, Tennessee, Gonzaga.  Assuming these 7 were split as #1 and #2 seeds, the best bracket would be the #1 seed who doesn't get any of these 7 as the #2 seed.

Of the seven, would prefer to draw Tennessee on our side of the bracket...

mfan_in_ohio

February 23rd, 2019 at 9:22 PM ^

The goal is to get to the Final Four and win, so you should also look at the average seed of the elite 8 opponent.  I'll bet that's a much tougher road for the 2 seeds than the 1 seeds.  You can also look at how rare it has been for 2 seeds to win the championship  (1 seeds are 4 times as likely), or the fact that 2 seeds have made the sweet 16 63% of the time vs. 86% for 1 seeds.  

But if you just want to look at sweet 16 opponent difficulty, let's use the actual numbers (through the 2017 tournament).   According to that chart, the average seed of the sweet 16 opponent for a 2 seed is (0.508*3 + 0.326*6 + 0.152*11 + 0.015*14) = 5.352.  The average opponent seed for a 1 seed in the sweet 16 is (0.470*4 + 0.326*5 + 0.152*12 + 0.045*13) = 5.919.  So the sweet 16 opponent for a 1 seed is over half a seed line worse than for a 2 seed.  

J.

February 24th, 2019 at 3:19 AM ^

No.  If you are a 1 seed, variance is not your friend.  You would rather play a team that is no better than a 4, rather than playing a team that has a 50% chance to be a 3.

Furthermore, that's assuming a perfect S curve, but the real situation is actually more dire.  The committee has shown that they are perfectly willing to move teams around within the same seed, but less willing to bump them up or down a seed.  You're assuming that the best #2 will be matched up with the worst #1, but that isn't necessarily the case.  The best #2 might get matched with the best #1, because they may need to move other teams around to make the bracket work.

Your path to the championship is always better as a #1 seed than it is as a #2 seed.