March 14, 2017 - by David Hess
Note: This 2017 bracket writeup is included as part of our NCAA Bracket Picks product, and will give you a very specific taste of how we approach bracket strategy, and why our approaches have been so successful.
As with classified government documents, some team names have been redacted for their safety. To reveal all the perpetrators, you can sign up here.
If you’re a returning customer of our premium bracket picks, you should already be familiar with one of the key tenets of our NCAA bracket analysis:
In other words, there are no golden rules for picking brackets that always give you the best chance to win, if you simply apply them like clockwork every single year.
The classic examples are all the seed-based picking “rules” floating around, which are based on insufficient substantiating data and no consideration whatsoever of the current year’s bracket dynamics.
If someone tells you that you should always pick three No. 1 seeds to make the Final Four, for instance, you should run away from them as fast as you can.
If the No. 1 seeds in a given year end up being relatively weak by historical standards yet also very popular picks, putting three of them in your Final Four would almost certainly be a horrible strategy.
In some years, there’s a team that’s head and shoulders better than the rest of the NCAA tournament field, and it’s worth picking them as your bracket champion — even if they’re a popular pick.
Other years, the public’s most popular pick to win it all is highly overvalued, and you should avoid that team and pick a less popular alternative.
This year’s situation, as usual, is different than recent years. In fact, as far back as we can remember, we’ve never seen a nation so divided.
No, we’re not talking about politics. We’re talking about this year’s NCAA tournament champion picks.
Most NCAA tournaments feature one or a few teams that the public really likes. This was the case last year.
Among the remaining 65 teams in last year’s tournament, no more than 6% of the public picked any of them to win it all.
In comparison, here’s the situation this year.
As you can see, there’s a big cluster of teams that are all modestly popular champion picks. No less than eight teams are the NCAA champion pick in at least 6% of the nation’s brackets right now, yet no team has more than 17% public support.
As it turns out, the eight most popular tournament champion picks this year are also the eight most likely teams to win it all according to our Survival Odds. However, the order of the teams is different.
Looking at that list above, the first observation is that there’s no big tournament favorite in 2017.
If the odds are true, it’s going to be significantly harder to correctly pick the NCAA champion than in some other years, even if that was your only goal.
Comparing this year’s tournament win odds to public picking rates, as of Monday afternoon, REDACTED is the highest value champion pick of the 2017 NCAA tournament.
The REDACTED have roughly a 12% chance to win it all, yet on Monday they were being picked by only about 7% of the public.
Of course, we want to sanity check our numbers, and make sure our projections for REDACTED aren’t too far off the market. Here are estimates of REDACTED‘s odds to win the 2017 NCAA Tournament according to various objective sources:
We’re happy with our position in that list. It’s often a yellow flag if your system’s projections are very far off the implied projections of the betting markets, and our odds for REDACTED line up pretty closely with Pinnacle’s.
At the same time, being a little off the betting markets is OK too. For a variety of reasons, betting odds aren’t always 100% efficient. Sometimes there is value to be had, as evidenced by smart people who make a living betting on sports.
If REDACTED is indeed the champion pick with the most value this year, what should you do with that information?
In comparison to 10 or 15 years ago, there is now a good amount of bracket advice floating around the web that purports to be “advanced” and “smart.”
Most of it is centered around the same theme — that your best strategy to win a bracket pool is to look for undervalued teams that have flown under the radar of the general public, especially for your NCAA champion pick.
Follow that logic, and REDACTED is a pretty clear play.
However, it’s not quite that simple. Again, there are no golden rules, and other factors (such as your pool size) also play a role in determining your optimal picking strategy.
We love value picks here at TeamRankings. There’s nothing better than finding a team that has a great shot to win a game — or the entire tournament — that is also flying under the general public’s radar.
That’s because to win a bracket pool, you don’t need to get a certain number of picks (or a certain percentage of picks) in your bracket correct. You don’t have to be amazing, by absolute standards, at picking winners.
To win a pool, you simply need to score higher than the rest of your opponents.
And the only way to do that is to make correct picks that your opponents get wrong.
Put another way, if you correctly pick North Carolina to survive the first round this year, that really doesn’t mean squat. Almost all your opponents will make that same pick, and they will gain the same number of points as you do. Everyone’s score might as well still be 0.
The most important picks in any bracket are the ones you make that are different from a large chunk of your opponents. Those picks give you a chance to gain points that your opponents miss out on.
And the best times to diverge from the picks of your opponents are when a team’s chance to advance to a given round is significantly higher than the rate at which your opponents are selecting them to make that round.
With that said, a set of NCAA bracket picks is a complex web of interdependent decisions. So to formulate a winning bracket strategy, you can’t consider every single decision in a vacuum.
In other words, while value-based picking approaches make a lot of sense, that doesn’t mean you should pick every undervalued team in your 2017 NCAA bracket. In fact, far from it.
North Dakota: Probably Not The Smartest Value Pick
Here’s an extreme example that illustrates why. We project that No. 15 North Dakota has a 5% chance to knock off No. 2 Arizona. But only 4% of the public is picking North Dakota to win. Therefore, North Dakota is indeed an undervalued pick.
But unless you’re in a giant pool (and/or a pool with huge upset bonuses), Arizona is clearly the better pick from a risk vs. reward standpoint.
If you take a big gamble on North Dakota and lose — the very likely outcome — you’ve just torpedoed your chances at collecting even more points from likely future-round Arizona wins. Worse yet, many of your opponents will be scoring those points, and distancing themselves from you in the pool standings.
Across the entire 2017 NCAA bracket, there are a lot of undervalued picks you could make. In fact, unless the pick popularity of two opposing teams exactly matches their respective win odds for a game, by definition, one of those two teams is undervalued.
But in most cases, if you filled out your bracket chock full of value picks, you’d end up taking way too much risk, and actually lowering your odds to win your pool.
The First Team All-Value Final Four For 2017
For example, if you picked all the most undervalued teams to reach the Final Four this year, your picks would be:
That’s certainly a fun bracket, but it’s not a particularly smart one.
The most likely result — with a 65% chance — would be that you get zero Final Four picks correct, and there would be only a 0.01% chance (1 in 10,000) of getting all four picks correct.
Even in a huge pool, that’s too risky. In contrast, picking “chalk” (all favorites) would give you a much better 19% chance of completely whiffing on the Final Four, and a 1.3% chance of a perfect Final Four.
In most scenarios, the ideal risk/value balance is somewhere in between these two extremes.
Now that we’ve established the limitations of picking for value, let’s take a look at some of the value picks you are likely to see in customized Early Deadline Brackets for the popular 1-2-4-8-16-32 scoring system.
Then, we’ll circle back to the all-important (in that scoring system, at least) champion pick.
There are two competing goals when crafting a bracket: Maximize reward and minimize risk.
In very small pools with standard 1-2-4-8-16-32 scoring, it makes more sense to focus on minimizing risk. After all, you only need to beat a few opponents. So simply picking the most likely winners, and relying on your opponents to miss most of the upset picks they’re bound to make, is a good idea. A lot of people love to brag about the upset picks they “called” in their bracket, and end up getting way too risky with their picks in smaller pools.
In very large pools, value comes more to the forefront, as you need to differentiate your bracket from many competing entries in order to have any realistic shot of winning (and even then, your odds to win are still long). So if you’re in a very large pool, you’re usually going to see some riskier pick suggestions in our customized Early Deadline Brackets.
For bracket pool sizes in between, figuring out the optimal balance of risk vs. reward — based on your pool’s size, scoring system, and payout structure — is often a much more nuanced challenge.
For example, using data from our Data Grid:
Now, repeat this same question and analysis for every undervalued pick in the entire tournament, compare all the risks and rewards, and try to figure out how much risk overall makes the most sense for your specific pool.
The complexity piles up rapidly, and the answers are all interdependent.
Finding the best set of picks for an entire bracket, for a particular type of pool, with a particular total number of entries, simply can’t be done manually. The complexity involved in doing all the optimization is much too high.
That’s why we built our NCAA Bracket Picks product, which leverages millions of computer simulations to identify the combinations of decisions for different pool types that, as a whole, will deliver a big edge.
(P.S. Based on our initial simulation run of bracket pools with standard 1-2-4-8-16-32 scoring, if you’re only entering one bracket in a pool this year, picking REDACTED to make the Final Four only starts to make sense once you have around 2000 entries in your pool. Try figuring that out by hand.)
In pools with standard 1-2-4-8-16-32 scoring, here are some value picks that you may see in our customized Early Deadline Brackets.
Most years there are teams with 25% or better odds to make the Final Four that are also being underrated by the public. That’s not the case this year, so the value picks listed here are significantly riskier than in other years.
All of the Final Four value picks above, plus…
All of the Final Four and Elite Eight value picks above, plus…
Keep in mind that most of the Early Deadline Brackets published on Monday will only feature some of these picks, since using all of them would almost certainly introduce too much overall risk into your bracket.
In addition, some of these picks may only be used in our alternative brackets for people making multiple entries into a pool, but not in our Best Bracket for a given pool.
It’s imperative to understand that we are not saying, or expecting, that all of the above value picks will win or make it to the rounds specified. In fact, we expect teams like REDACTED and REDACTED to lose in the first round, since we have their win odds at below 50%. But a gamble on a team like those just might be worth the risk in certain pools.
Notice we didn’t list any undervalued NCAA champion picks in the lists above. That doesn’t mean that none exist this year.
In fact, a handful of teams appear to be decent value tournament champion picks when you look at their profile in a vacuum:
Undervalued picks to win the NCAA tournament (that have a realistic chance)
Three of those teams stand out as having compelling cases for being the best NCAA champion pick in at least some pool scenarios:
Given those profiles, we initially figured that REDACTED would be a good champion pick in small pools, REDACTED a good pick in medium sized pools, and REDACTED a good pick in huge pools.
However, one of the key findings from our initial simulation runs this year is that making some of the earlier-round value picks listed above, in concert with a REDACTED-as-champion pick, is a better gambit in big pools. It provides a better reward than you would get from picking REDACTED as champion with slightly less overall risk.
So, with our two title favorites both also being undervalued by the public, it’s not a big surprise that the Early Deadline Best Brackets for 1-2-4-8-16-32 scoring recommend:
REDACTED is a slightly riskier pick than REDACTED, and that risk/reward tradeoff doesn’t quite make sense until your pool size gets above 75 or so entries.
As mentioned above, a bit more surprising is the fact that despite REDACTED‘s better value ratio, our Early Deadline Brackets for 1-2-4-8-16-32 scoring are also recommending:
By the way, roughly 50% of the bracket pools that our customers set up in our system use the most popular 1-2-4-8-16-32 scoring system, or an equivalent. That’s why we focus this writeup on that scoring system.
The remaining 50% of customer pools are spread quite thinly over a wide range of other scoring systems. So just remember that if your pool doesn’t use 1-2-4-8-16-32 scoring, you may have a different team than REDACTED or REDACTED as your recommended champion pick.
The bracket pool simulations not picking REDACTED as the champion in very large pools surprised us at first.
After all, if REDACTED is being picked by 0.3% of the public, then probably only three other people would pick REDACTED as champion in a 1000-entry pool. And with REDACTED‘s odds to win it all at roughly 2%, you would effectively have a 2% chance to be in a “mini-pool” for first place against only three other opponents, if REDACTED indeed wins the tournament.
Assuming your bracket has an average shot (25%) to outscore those other three opponents if REDACTED wins it all, a REDACTED champion pick would therefore give you about a 0.5% chance to win a 1000-entry pool.
Compare that to a REDACTED champion pick, with a 12% chance to win it all, and 6.7% of opponents picking the REDACTED. You’d have a 12% chance of being in a “mini-pool” for first place with an average of 67 opponents if REDACTED did win the tournament. With a 1-in-68 (1.5%) chance to win that mini-pool, overall you’d have about a 0.2% chance to win the pool.
Just considering the risks and rewards of these two potential champion picks, it seems like REDACTED is the better play in very large pools.
However, once we dug into the details, we realized we were overlooking a key factor. It’s much more difficult to use value picks in the earlier rounds to improve your overall win odds when picking REDACTED, than it is when you pick REDACTED.
As an example, in very large pools our Early Deadline Brackets are suggesting picking REDACTED to make the Final Four, paired with the REDACTED champion pick.
That REDACTED pick has only about an 11% chance of being correct, compared to 29% for picking REDACTED instead. As it turns out, making a risky pick like REDACTED to the Final Four doesn’t make much sense if you already would have great odds to win your pool (25%) if you picked REDACTED to win.
Imagine your pain if you made a huge bet on a relative longshot like REDACTED, and then they won…but you took fourth place instead of first place in 1000-entry pool because of a risky and incorrect REDACTED in the Final Four pick.
But the situation is different when your baseline odds to win a “mini pool” of people who correctly picked the NCAA tournament champion are only 1.5%, as in the REDACTED case.
In that scenario, not only is an additional gamble like REDACTED to the Final Four mathematically justified, but it also ends up creating a better overall risk vs. reward profile for your bracket as a whole than the REDACTED-as-champion option.
In the end, then, our initial intuition was both right and wrong: If the only pick you had control over in your very large pool bracket was the NCAA champion pick, then REDACTED would indeed be a better champion pick than REDACTED.
However, when also you factor in value picking options elsewhere in the bracket, taking multiple smaller risks in the earlier rounds and picking REDACTED as champion ends up being a better tradeoff than making one huge bet on REDACTED.
In addition to an undervalued REDACTED champion pick, in larger pools, the other calculated risk the majority of Early Deadline Best Brackets are taking for 1-2-4-8-16-32 scoring is REDACTED making the Final Four.
To be clear, REDACTED isn’t technically undervalued to make the Final Four; 37% of the public is picking them to make it, despite them having only a 27% chance to do so. But REDACTED is still less overvalued than alternative REDACTED, who is being picked at a 45% rate despite only a 31% chance to make the Final Four.
So it’s a bit of increased risk (4% less likely), but the potential reward makes it worth it (8% less popular).
In addition, in some larger 1-2-4-8-16-32 pools, we’ll be suggesting a truly undervalued Final Four pick of REDACTED or REDACTED to emerge from the REDACTED region instead of REDACTED.
REDACTED is the most overvalued Final Four pick this year, with 54% of the public picking them, despite only having a 29% chance to make it. The drop from REDACTED‘s 29% survival odds to 14% (for REDACTED) or 11% (for REDACTED) is indeed a big one, but in large pools it’s a justified risk given REDACTED‘s popularity.
If you think a pick like this sounds a too risky, just remember, you have almost no chance to win a huge pool to start. Increasing your odds to win therefore requires identifying the choicest opportunities to differentiate your bracket from a sea of others.
Following over half of your opponents in picking a team that doesn’t even have a 30% chance to make the Final Four is not the best way to do that. Even if REDACTED makes it out of its region, it will likely be a pyrrhic victory for you. Most likely, you’ll still be fighting an uphill battle against many other opponents still in contention to win the pool.
Our initial round of millions of 2017 bracket pool simulations is now complete. Thank you, 18 servers at Amazon running tournament and bracket pool simulations all night!
We’ve factored in recent public picking trends, Vegas odds, team ratings from several of the world’s top predictive systems, and hands-on analysis of injuries and lineups using proprietary tools we’ve built.
We manually reviewed the Early Deadline Brackets for many different scoring systems and pool sizes before posting them, and did our own separate analyses to confirm that their simulation-driven picks passed the smell test.
And now, we’re confident that these brackets will give you a significant edge in your pool. Just remember: if your pool doesn’t have an early submission deadline, we strongly recommend that you wait to use our Official 2017 Brackets, scheduled for release late Wednesday night. Those will take into account the latest public picking trends, Vegas lines, and any random injuries or suspensions that may happen before the games start.
Want to reveal all the team names and get the best bracket to win your pool? Subscribe now
Printed from TeamRankings.com - © 2005-2018 Team Rankings, LLC. All Rights Reserved.