FanDuel's Bracket Pick'em: Which Seeds Historically Perform Best?

With FanDuel unveiling a new way to fill out brackets for the NCAA tournament with the Bracket Pick'em, we're starting from scratch. We've never gotten to fill these puppies out before, so clearly, it's going to be a bit of an adventure.

But that doesn't mean we can go into this with some DGAF-y attitude. There's $25,000 on the line with $5,000 going to the top entry. I don't know about you all, but I wanna get a piece of that meaty pie. That means it's time to dig in and do some research.

In a traditional NCAA tournament pool, it's pretty simple: pick the team you think is most likely to win. Here, we get bonuses for picking higher-seeded teams that do well. Let's run through all of the rules here quickly.

1. You pick five total teams you think will do well.

2. The seeds of those teams must add up to at least 20. For example, you can't just pick all high seeds; you can pick a 16 seed and four 1 seeds if you want, but that's probably not something desirable, as we'll see later on.

3. Scoring works as such:

So, to give this real-life relevance, a 2 seed that wins in the first round gets 20 points. They would get 50 points for a second-round one, 100 for a win in the third, etc. Every lineup that gets at least 400 points will win a prize. Those to score at least 2,500 points will split $20,000 with the others who hit that mark. The table above gives us a better idea of what we need to do to reach those benchmarks.

Clearly, finding a team that goes deep into the tournament is essential, incentivizing picking the top teams. But we also want that multiplier bonus for identifying lower-seeded teams. What, exactly, is the proper mix?

Let's try to answer that by looking back at how seeds have performed in previous tournaments. If teams ranked at a certain seed tend to perform well, then we're going to want to look harder at the teams in that area. If there are seeds that just do nothing, then we'd be wise to look elsewhere.

Which seeds should we be focusing on for 2018's Bracket Pick'em? Let's check it out.

Historical Performance By Seed

To check this out, let's go back through every NCAA tournament from 2000 on. This gives us 18 tournaments of data, amounting to 72 teams at each seed. That's a pretty decent track record to show which seeds have the ability to beast out.

Below is a table showing the average Bracket Pick'em score for each seed, the median score, and the maximum score. Each of these numbers will play a role in determining how we want to play this thing.

There's a lot to dissect here, so let's dive in.

That score of 6,545 for the 7 seeds was out of the Connecticut Huskies in 2014 when they won the whole thing. No other score has been higher than 3,480, so we've got ourselves a clear outlier. Unless you think a 7 seed can go all the way this year, you shouldn't be banking on that type of performance.

If you omit that one outlier, 7 seeds have an average score of 135, which would rank ninth among all seeds. It's unfair to omit the top score for any seed because that is still possible, but 7 seeds aren't quite the gems they appear to be based on their average and max.

When you're looking for a good mix of both floor and upside, it appears that 3 seeds are the magic spot. They have the highest median score, second-highest average score, and third-highest max score. The fourth-, fifth-, and sixth-highest individual scores all belong to 3 seeds that went on to win the national title, best only by 7 and 8 seeds that advanced to the championship game. It's not a bad idea to take a stab at multiple 3 seeds if you believe they've got a shot to run the table.

By looking at the table, you can see that 1 seeds have a considerable lack of upside. The maximum points they can score for winning the championship is 935. There have been 22 scores higher than that in the 18 years of data included. That's an issue.

As a result, it's fine to pick a 1 seed, but you must have a good amount of confidence that they'll win the whole championship, and you had better not have more than one on your roster. Doing so caps your team's upside, lowering the odds you'll be able to snag that top prize of $5,000.

The same is largely true with 2 seeds, though it's obviously a bit more relaxed. Their max score is 1,870, which is twice as high as the 1 seeds. But a 2 seed has won the tournament just twice in the past 18 years whereas 1 seeds have done so 12 times. As such, it's likely best to stick to one here, as well, if you do decide to roll with a 2 seed.

The luxury of picking higher-seeded teams is that they don't necessarily need to go all the way to pay off. The 10th- and 11th-best scores in the past 18 years have come from 11 seeds that advanced to the final four before losing. A 12 seed that merely advanced to the elite eight posted a score higher than the max score for a one seed. While you may not need to target seeds this high due to the lack of upside in the smaller seeds, it's not a bad idea to do so, even if it means your team's seed sum goes beyond the minimum of 20.

Conclusions

Every tournament is going to be different, and we'll have to modify our strategy once the brackets come out on Sunday. But based on this historical data, it seems like we do have some actionable takeaways.

First, we shouldn't go too heavy on the 1 and 2 seeds. They do have a better shot at winning the championship, but they also carry lower upside than the lower seeds. The best way to play this may be to pick a team in the top two seeds you expect to do well, plant your flag, and then jump down the totem pole.

Second, 3 seeds are a pretty sweet outlet due to both their floor and upside. This may be the seed where it's more okay to start using more than one team and hope that they're able to make a push deep into the tourney.

Third, there's nothing better than a high-upside lower seed. If you believe that a 10 or 11 seed has a chance to win two or three games, you need them on your team even if it means your seed total is above the minimum. These are the types of teams that can push your lineup over the top. Additionally, putting your seed total above 20 can allow you to differentiate from the field, increasing the odds that you'll get the top prize all to yourself rather than having to share it with others.

You're going to have some work to do in identifying which teams in these seed ranges have the best odds of making noise. But with this baseline knowledge, we at least have a template of what to expect. In any new game format, that's going to make a world of difference when we're trying to win that prize.