Balancing Risk and Value in Your Bracket
To win your bracket pool, you need to consider both the upside of unpopular value plays and the downside risk of making each selection.
When taking a charge, you have to balance the value of getting a turnover with the risk of getting crushed( Photo by Keith Gillett/Icon Sportswire)
Coming up with an optimal strategy for a bracket pool requires balancing both risk and value. To win a bracket pool, you must score some points that your opponents do not get. That’s the value part, where you are getting an edge on your pool in situations where your opponents may be under-picking a particular team relative to their chances.
Conversely, though, you must not miss out on too many points that your opponents do get if you hope to win. And therein comes the risk part of the evaluation. If you get too risky, you could be giving away points to your competition. Few things are more painful in a March Madness pool than sitting with no chance to get points, because an upset you picked to happen did not advance, while your competitors are about to add to their total in the standings.
If you do more of the former (gain points on the field) than the latter (take zero points when others are scoring), then you have a pretty good chance to cash out. But how do you balance the two competing considerations of risk and value? In this section, we will go through some of the things we have learned over the years.
First, let’s define what we mean by value. At a simple level, value represents the difference in what objective measures of probability say could happen, and how the your opponents (“the public”) are picking outcomes. In our case, we use our internal predictive power ratings, make manual adjustments based on things like injuries and lineups, and also monitor the betting markets. We also gather public pick information on how games are being selected, to assess value.
Technically, then, any outcome where the public pick rate does not exactly match our game-advancement odds provides value one one side. If we project a No. 4 seed to have a 50% chance of reaching the Sweet 16, but the public is only picking them 35%, that team has value to make the Sweet 16.
Similarly, if a No. 16 seed has a 5% chance of winning against a No. 1 seed, but is only being picked by 2% of the public, that No. 16 seed technically has value. But there are limits to when you would want to rely on value over risk.
Risk, as we use it here, is the difference between the safest possible outcome you could select, and the alternative you are considering. As the tournament goes on, each successive round becomes riskier overall, in the sense that it is less likely that even the best teams advance. But we are looking at the relative differences when we evaluate risk.
When evaluating whether to pick a first-round game, assessing risk is as simple as looking at the difference in win odds between the two teams. When it comes to evaluating risk in a decision on who to take to the Elite Eight, though, the risk assessment becomes the relative difference between the team you are considering, and the team most likely to advance out of that section of the bracket.
Take 2017, when North Carolina was our most frequent national title recommendation. North Carolina was, in absolute terms, a risky pick, as we assessed them as having a 15% chance of winning the title. But you have to pick somebody to win the national title, and no other team had a better individual chance than those 15% odds. Thus, when evaluating risk on who to pick, other teams’ title odds were compared with North Carolina’s 15%. (And with the public picking the Tar Heels slightly less frequently than our odds put them to win, both their risk and value profiles were favorable.)
The Limits of Value-Driven Picking
While value-based picking makes a lot of sense as a general strategy, that doesn’t mean you should pick every undervalued team in your NCAA bracket. In fact, far from it.
Let’s turn back to that No. 16 seed versus No. 1 seed example. If all you cared about was value, maybe you pick the massive upset. But the relative risk of picking that upset is huge. In most years and in most cases, you lose ground to nearly all of your opponents. The reward is small, and your opponents would still be positioned to get more points as the No. 1 seed advanced, hurting you even more in most pools.
If you make an unpopular pick that doesn’t end up playing out, and a very popular pick ends up surviving in its place, then it’s bad news for your bracket. And because an NCAA bracket is a web of 67 interdependent decisions, you can’t consider each individual pick in a vacuum.
Value Only Gets You So Far
Often in our early bracket write-ups analyzing each year’s tournament, we highlight the “most under-valued” picks to get to the Final Four from each region. Rarely if ever is one of those teams a No. 1 seed, and they’ve included as high as a No. 10 seed (Wichita State in 2017).
And every year, we write about those value plays to the Final Four: “[t]hat’s certainly a fun bracket, but it’s not going to be the smartest one for the overwhelming majority of pools.” We also point out that the most common result by going that “all value” route would be to finish with zero or one Final Four teams correct.
As a group, the teams we’ve highlighted as value Final Four picks have played better than their seed expectation. Most of them were good plays to advance farther than the public expected, and provided value. But only a small percentage of them have reached the Final Four.
So simply going all “best value plays” based on the difference between advancement odds and public pick popularity is too risky for all but the largest of pools. You don’t win small pools making risky picks expected to net you significantly less than one Final Four team.
Employ a Risk-Based Strategy Appropriate For Your Pool
In most bracket pool scenarios, the ideal risk and value balance is somewhere in between the extreme of picking all value picks on the one hand, and all favorites on the other. It usually means having some value picks interspersed throughout the bracket. From year to year, your “value bets” may end up being more concentrated on a fewer number of key picks (e.g. a big gamble on your NCAA champion pick), or spread across a greater number of earlier round picks.
Now that you have a solid understanding of how to use predictions and public picking trends as a framework for evaluating individual games and upsets, you also need to craft a high level strategy to win your bracket pool. Your high level strategy should be driven by one simple question:
How much risk do I need to take to maximize my odds to win this pool, given the number of people I’m competing against, my scoring system, and other dynamics?
The answer to this question is critical because it influences the logic you should apply to many individual bracket decisions. More than any other variables, the number of people in your bracket pool and your scoring system should determine whether you apply an overall conservative (picking mostly favorites) versus aggressive (picking some significant upsets) strategy.
The laws of probability dictate that the more people in your pool, the lower your odds of winning. The larger your pool is, the more likely it is that somebody will pick some highly improbable outcomes and, by sheer luck, get a bunch of them right. To provide an extreme example, in ESPN’s 2011 bracket contest, which featured millions of players, two people actually picked the entire Final Four correctly, despite the fact that it featured an 8-seed, Butler, and also 11-seeded VCU, who was a play-in team!
But in most small pools in that same year, simply hitting on one Final Four team in a wacky year, while performing better than others in the early rounds, may have been enough to finish in the money.
Small Pool Strategy: Minimize Risk
In a bracket pool of 10 people, it absolutely does not pay to get all wild and crazy with your picks. All else being equal, you’re starting with 10% odds to win that pool, which isn’t bad at all. You have much more to lose making risky picks than you do in a 1,000-person pool, where your base odds to win are 0.1%.
In a small pool, don’t even think about making a lot of long-shot picks in a traditional scoring system. The primary strategic blunder in a small pool is picking too many upsets, especially in the late rounds.
With only a few entries to beat, the chances that any one them catch the actual longer shot picks is relatively smaller. Sure, it could happen, but upset results that have low public popularity are generally not going to hurt you. Outcomes like No. 11 seed Loyola-Chicago advancing to the Final Four in 2018, and South Carolina doing the same in 2017 as a No. 7 seed, had such low pick popularity that in most smaller pools, no one picked those outcomes.
More often, the most costly outcomes in a small pool are those where a popular team you did not select advances deep into the tournament. As a concrete example, in the 2018 tournament, Michigan reaching the final (as a No. 3 seed of decent popularity) was far costlier to those that did not have them, than Loyola-Chicago reaching the Final Four.
Thus, the more optimal strategy in small pools is to limit the amount of risk you take, and not pick extreme results with low odds that could sink your chances.
Mid-Size Pool Strategy: Balance Risk and Value
As pool size increases, so should the risk tolerance. The number of value plays should increase, but should not solely drive decisions. In many bracket pools, two key areas in determining who wins the pool are:
- Picking the right champion (often) and maybe the opponent in the title game; and
- Doing well in picking the mid-seeded or toss-up games in the first three rounds (sometimes)
Our brackets in mid-size pool may concentrate risk and value plays in one, but often not both, of the two bracket areas mentioned above, as a type of hedging strategy. The key is to have the appropriate overall risk profile given the scoring system and pool size, but that risk need not be spread across all parts of the bracket equally.
If you are very conservative in choosing the winners of your late stage games – let’s say you have the two most popular choices meeting in the final – then you’re not differentiating yourself from many of your competitors. In this case, a more aggressive value upset picking strategy in the earlier rounds should increase your relative odds to win your pool.
On the other end of the spectrum, you can go with a very conservative early bracket, and focus your risk on avoiding the most popular NCAA champion pick or putting a few very good, but undervalued, teams in the Final Four.
In small to mid-sized pools (between 25 to 100 people or so) especially, these types of risk hedging and balancing strategies have proven highly effective for us over the years. With that number of people, it usually takes some success at both ends of the bracket to win, but even just a highly effective early or late round strategy can secure a prize some years.
Large Pool Strategy: Value Comes to the Fore
If your goal is to come in first on ESPN, or in any relatively large pool (say several hundred people or more), picking all the most popular teams is generally a terrible strategy. In huge pools you generally have to take substantial risks in order to give yourself a better shot at finishing first. This is because, while any individual entry is unlikely to nail the right combination of teams to advance deeply, with a large group, some entry is bound to hit those long-shot plays.
Few people can get themselves to alter their strategy based on pool size, since it requires them to think objectively and, often times, to make unpopular picks. It’s emotionally difficult, but if you can muster the guts to do it, you gain a huge advantage.
In large pools, selections from the public don’t differ all that much from how most people pick smaller pools. That means the favorites are over-represented, and some underdogs can have “pot odds” (a poker term referring to when some underdog hands can still be the right play because of the relative payoff). For example, if you are in a very large pool and you can take an outcome that has a 1% chance of happening (like a particular No. 5 seed to win the title), but the public is picking it 0.1% of the time, you can drastically improve your chances by picking that unlikely outcome.