Stay informed with the
NEW Casino City Times newsletter!
Gary Trask Archives
More Strategy Experts
9 March 2017
By Gary Trask
Selection Sunday is just days away, which means — like it or not — you are about to be inundated by "bracket talk."
Doesn't matter if you don't know Coach Cal from Coach K, or you have no idea if the Gonzaga Zags come from Washington state or Washington, D.C. Flip on the TV, walk into the lunchroom at work or listen to random conversations on the train early next week. Everyone is going to be talking about March Madness and who they have in "their" Final Four.
The American Gaming Association estimated that more than 70 million NCAA pool brackets were filled out by Americans last year. That's more than the entire population of France.
My point? There's a very good chance that if you are reading this column, you are going to have an active bracket next week when the NCAA men's basketball tournament tips off. And we're here to help.
Ed Feng is a data scientist and writer who has parlayed his love for sports and numbers into a career of predicting outcomes of games. Feng created an algorithm based on his Ph.D. research at Stanford that takes the margin of victory in college basketball and football games and "accurately adjusts" for strength of schedule, resulting in his team rankings, which are updated regularly and can be viewed on his website, ThePowerRank.com.
Feng's results have been noteworthy enough to earn the attention of Deadspin, FiveThirtyEight and Business Week, all of which have featured his numbers and predictions. He's also a regular columnist for Bleacher Report and the Detroit News, and he's the author of the appropriately titled book How To Win Your NCAA Tournament Pool.
"I don't get invited to play in too many pools," shrugs Feng, 40, with a hearty chuckle. "In fact, I don't think I'll be in any at all this year."
But that doesn't mean he can't help us swoop in and take home bragging rights among our friends and co-workers. After reading Feng's book and speaking with him at length over the phone, I've brought you 10 tips to keep in mind when you get your hands on that blank bracket. (For additional assistance, go the ThePowerRank.com, sign up for the free newsletter and after Selection Sunday Feng will send you a free NCAA Tournament Cheat Sheet).
10. Analytics over randomness
Feng is forthright in saying it's difficult to predict the outcome of tournament games, for a couple of reasons. First, college basketball provides a limited number of games to evaluate teams. Secondly, randomness plays a huge role.
But, he warns, don't fall into the trap of saying "anything can happen" in March Madness, or blindly follow the narrative you'll hear on TV about the slew of upsets that happen every year.
"It's not possible to overcome the randomness of college basketball," Feng writes in his book. "But analytics can help with the lack of data and differing strength in schedule."
"You can't predict everything," he adds. "But, with that said, everything is not a 50-50 proposition, either. You can actually predict tournament results with some pretty surprising regularity."
And for proof . . .
9. Higher seeds rule
People tend to remember the monster upsets that take place seemingly every year in the NCAA tournament, but, in reality, these games are the exception.
"There are always lower-seeded teams that defy the odds, and we all love those upsets," Feng says. "But in the big picture, the favorite usually wins."
According to Feng's research, from 2002 through 2016 there were 978 NCAA tournament games. Taking out games in which opponents were seeded the same (which can happen in the Final Four and the newly added set of "first four" games), the higher-seeded team has prevailed 71.6% of the time (668-265).
8. Don’t fixate on the early rounds
Yes, it's true that a double-digit seed will most likely make the Sweet 16 this year. Yes, there will be "upset alerts" all day long on Thursday and Friday, prompting the people next to you in the sports bar to start cheering raucously for a school they'd never even heard of that same morning.
But since most pools give minimal points to first- and second-round games and heavily award correct picks later in the bracket, Feng suggests you save some energy and buzz right through those early round picks.
"Don't waste your time and effort scouring the internet and trying to find sleepers," he said. "You're much better off just following my own rankings, or another notable source like KenPom.com, and using those picks blindly for the first two rounds."
One more related note: There's a popular notion out there that the 12 seeds have enjoyed great success over the 5 seeds every year, and that this is a perfect situation to be bold and call for an upset. But Feng points out that since 2002, the 5 seed has won those match-ups 66.9% of the time.
7. Picking the champion is most important
While Feng downplays the importance of the early rounds, he can't be more ardent about how vital it is to pick the sole survivor.
"The most important choice in your bracket is the champion," he writes. "In the most common scoring system, the correct choice for champion is worth 32 points, one sixth of the total possible points and the most for any one game. Make this choice count."
Utilizing Feng's numbers for your overall winner is not a bad strategy to consider. During the 15-year period between 2002 and 2016, the team with the highest or next highest win probability by The Power Rank has won the tournament eight times. The win probabilities ranged from 32.2% for Kansas in 2008 to 12.2% for Florida in 2007.
6. Avoid big pools
One of Feng's essential tips is to stay out of the pools with a massive number of entries.
"Your chance to win a pool depends strongly on the size of the pool," Feng writes. "Don't get in a large pool of over 100 people."
5. Contrarian is the key
Despite his advice to shun the big pools that offer huge paydays, Feng is realistic. He's aware that you have little control over how many people are in the contest you enter and that, quite frankly, "it's boring to enter a small pool."
So, if you dive into the deep end and take on hundreds of other entrants, the best strategy is to be contrarian.
"If you pick the favorite to win and it wins, everyone else is going to get those big points with you and it becomes pretty likely someone else is going to beat you," he says.
Feng advises picking a "contrarian champion" that has a "large win probability" based on his rankings, but is being overlooked by the general public, which you can decipher on ESPN.com where the predictions of more than 13 million entrants to the NCAA Tournament Challenge Bracket could be viewed last year.
For example, in 2010 Kansas was the overwhelming favorite and had a 31.1% chance to win the title, the best win probability by over 10%, according to ThePowerRank.com.
Not surprisingly, the public was fully on board with "Rock Chalk, Jayhawk," as a whopping 41.8% of the ESPN brackets had Kansas prevailing.
Feng's suggested "contrary champion" was Duke, since the Blue Devils were the second-ranked team on The Power Rank, but only had 6.5% of the public backing them.
And who was your 2010 NCAA champion? Yep, Coach K and the Blue Devils.
Kansas, meanwhile, blew up brackets from coast to coast when it failed to get out of the second round after getting stunned by Northern Iowa.
4. Favor favorites in small pools
"Smaller pools give you better odds," Feng says. "If winning is your only goal, enter a small pool."
When doing so, it’s OK if your bracket leaves a trail of chalk dust behind. Feng writes that by picking all favorites, you will have a 38% chance to win a 10-person pool.
"That means you'll win about every other year, which is pretty good," he adds.
By comparison, for a 100-person pool, the win probability drops to 5% using the same strategy.
So, going back to the 2010 example, Feng's bracket for a smaller pool would have been exactly the same as the one for a larger pool with one important exception: Kansas would have been the predicted champion.
3. Fill out multiple brackets
OK, we realize purists and traditionalists are going to be horrified by this, but think of it like the lottery. You are entering a contest and the more entries you have, the better chance you have a winning.
But when applying this strategy, you must vary your picks in the latter stages of the tournament.
"You want to increase your chances and the way to do that is to have some diversity," Feng says. "Change around the teams you have in the Final Four and in the championship game, and have different winners."
2. Know who is in your pool
While the ESPN numbers provide a great approximation for what the brackets will look like in your pool, Feng says that sometimes you have to make subjective adjustments depending on who the other entrants are in your pool.
For instance, if your pool is Los Angeles-based, there will probably be a lot of people picking UCLA to win it all this year, even though the Bruins may not be a popular choice at ESPN.com. Or, if back in 2010 you were in a pool with a bunch of Duke graduates, you wouldn’t have been able to make the assumption that the Blue Devils were going to be a contrary play in that particular pool.
1. Avoid teams that rely on the 3-pointer
When making the crucial decision of who you think will win the NCAA tournament, Feng's best advice is to stay far away from teams that depend heavily on the 3-point shot.
"In general, teams that shoot a high rate of 3-pointers do not win the tournament," he points out. "The data suggests not picking these high-volume, 3-point-shooting teams."
Now, we realize the hardcore college hoop fans are jumping through their computer screen right now, screaming about last year's champion, Villanova. Yes, the Wildcats were a prodigious 3-point shooting team. Feng admits in this year's book that by winning the title they broke the trend, but there's more to the story.
Villanova came into the 2016 tournament taking an alarming 44% of its shots from behind the arc, which had yellow caution lights shining all over the place for Feng and his followers last year at this time. But it seems Head Coach Jay Wright may be a fan of The Power Rank, because during his team's incredible run to winning the title, the Wildcats changed their offensive philosophy, taking only 33% of their field goals from 3-point land.
The result? They nailed an astounding 51.2% of those long-range shots while sustaining a torrid pace on two-pointers at 58.3%. Even after shooting the ball at that staggering pace, Villanova was still very fortunate to take home the crown, needing a basket at the final buzzer to beat North Carolina in the championship game.
So, if we call Villanova's run an anomaly and take that team out of the equation, dating back to 2002, only three teams that entered the tournament shooting more than 33% of their field goals from 3-point land have won the tournament. No team has done so with a 3-point field goal rate of more than 35%.
"There's just such a high variance with these teams that live by the 3 and die by the 3," Feng adds. "It's definitely something you have to consider when making that all-important champion pick."
6 March 2017Your airfare and hotel room reservation were secured months ago. You’ve spent the last few weeks hydrating, padding your bankroll and preparing your body for the pandemonium. The countdown is on for your trip to Las Vegas for March Madness. Congrats. You’re almost there. All the proper advance planning is done. ... (read more)
3 March 2017Spring Break, St. Patrick's Day and March Madness. All three events collide in two weeks in Las Vegas, bringing millions of people to the area for four days of self-indulged debauchery. Talk about a perfect storm. Take it from someone who was in Las Vegas last year when this very same trifecta hit the city. ... (read more)
24 February 2017No cheering in the press box. It's one of the first rules a young journalist learns, and the figurative and literal meaning is vitally important. You should never root for — or against, for that matter — the people or teams you are covering. Last week, however, we had a rare exception here at Casino City. ... (read more)