Wednesday, March 14, 2018

It's March, It's 2018, and That Means It's Time for Dr. Brackets' 2018 Tips and Tricks

Meet Michael Porter, Jr. More on him below. Courtesy ESPN.

Cue the music! Cue Kevin Harlan! March Madness is back in less than 48 hours, and Bruff "Dr. Brackets" McGruff is back again with 10 of the freshest tips, tricks, and statistical anomalies that will make you go "Hmm, interesting," and then probably probably go on to win thousands of real American dollars and billions of Euros' worth of bragging rights in your circle of friends.

As always, I have to lead with the disclaimer that I am not an actual doctor. I am about as much a doctor as Ron Burgundy--who once, during a prank phone call, referred to himself as "professional doctor" Chim Richalds. I took a statistics class once about 5 years ago, and also I am somewhat familiar with how to run multiple regressions. I have a database with 7 years of tourney data that helps me come up with this stuff, so if you screw something up, don't blame me, blame the numbers! (and me.)


But man do I enjoy working with big sets of data. And there's a lot of good stuff out there to dig through. A few of my favorite models seem to get more accurate every year, so if the combined wisdom of several statistical wizards and one goof can help you climb to the top of a pool or two, then the work is all worth it. So let’s get down to it:

1. It's been a pretty weird year for college basketball, with FBI investigations centered on shady back-channel high school recruit payments making front-page news right before the season started and continuing through March with the ESPN story on Arizona coach Sean Miller potentially tied to a payment for freshman stud DeAndre Ayton dropping just like two weeks ago. Several bubble teams--including Louisville and USC--were tied to the investigation and just missed the cut, but Arizona was comfortably in as a 4-seed.


This guy's shadow could beat me in one-on-one. Courtesy Bleacher Report.

It's always tough to know how to "price in" external distractions like this, but usually big stories like this don't help the preferred "EVERYONE'S AGAINST US!" narrative. 

2. I say this every year, but when putting together your bracket, you want to be a little "off" from the conventional picks. You don’t finish in the money if you just pick along with the mob. So it’s important to pinpoint which teams have been overvalued and undervalued by the Selection Committee. This helps us figure out who might steal an extra game or two, and who might flame out earlier than most expect. Both are important, because every correct pick can be the difference between winning a pool and finishing just out of the money.


Plus, you want to pick a couple teams that you think would be fun to root for, because lest we forget, the next few weeks are all about fun.


This year's Committee did a pretty fair job seeding the teams ranked from about the 1-seeds down to the 9-seeds (The big quibbles this year are about who got in and who got left out around the bubble--like Nebraska and those shady sun-splashed folks at Southern Cal). Once again, it was mid-majors who have the strongest argument for getting somewhat hosed. Last year's runners-up Gonzaga are strong enough to be a 2-seed instead of the 4 they were saddled with. Loyola Chicago flew under the radar this season but beat Florida on the road earlier this year and probably deserved an 8-seed instead of their 11. Also, the best team in the entire field is two seeds too low, but we have a separate talking point for that craziness.


The more flagrant errors this year involved overseeded teams that look ripe for early upsets. The overseeding starts at the very top, as Xavier and Kansas are more borderline 2- or 3-seeds instead of the 1 spots they received. My model also says several SEC teams were overrated: Tennessee is more of a 5 than a 3-seed, Auburn should be a 6-seed instead of their 4, and Arkansas should be down in the 10-seed range instead of the 7 they received. 


3. This year, we've witnessed a disappointing rash of injuries in the weeks prior to (or during) conference tournaments, which are always tough for teams to quickly adjust to, as it takes time for teams that usually go 8- or 9-deep to settle into new rotations and usage patterns. Injuries can turn title contenders into 2nd-week also-rans, and can take promising mid-high seeds and make them upset fodder. Just ask last year's 6-seed Creighton, who lost NCAA assist leader Maurice Watson, or...(deep, heavy, concentrated sigh)...the 5th-seeded Gophers, who lost Akeem Springs in the Big Ten Tournament and immediately bowed out to Middle Tennessee.


And if the injuries occur late enough in the season, the advanced analytics haven’t been allowed to gather enough data to effectively “price in” a key player’s absence to his team’s route through the tournament.


Here's a list of the injuries you should know about this year, and keep in mind: injuries to point guards and big men often hurt the most in March. (Numbers denote ppg/rpg/then anything else of note.)


1) Kansas - Soph C Udoka Azubuike QUESTIONABLE, knee (7-footer, 3rd leading scorer, 13.7/7.1/1.8 bpg)
1) Virginia - Fr G/F De'Andre Hunter OUT, wrist (4th leading scorer, 9.2/3.5)
4) Auburn - Soph F Anfernee McLemore OUT, ankle (one of few post presences, 6th scorer, 7.4/5.3/2.7 bpg)
5) Kentucky - Fr F Jarred Vanderbilt LIKELY OUT OPENING WEEKEND, ankle (7th leading scorer, 5.9/7.9)
5) Clemson - Sr F Donte Grantham OUT, ACL (2nd leading scorer and top big man, 14.2/6.9)
6) Miami - Soph G Bruce Brown OUT, foot (2nd leading scorer and top rebounder and assist guy, 11.4/7.1/4 apg)
6) TCU - Soph PG Jaylen Fisher OUT, meniscus (4th leading scorer, 12.3/5.4 apg)
7) Nevada - Jr G Lindsey Drew OUT, achilles (5th leading scorer, 8.1/4.1/4.3 apg)
7) Texas A&M - Sr G Duane Wilson OUT, knee (6th leading scorer and PG, 9.0/4.0 apg)
8) Creighton - Soph F Martin Krampelj OUT, ACL (2nd leading scorer, shot 68% from field, 11.9/8.1)
9) Alabama - Jr F Donta Hall QUESTIONABLE, concussion (2nd leading scorer, all-SEC defense, 10.9/6.8/2.1 bpg)
10) Texas - Jr G Eric Davis OUT, suspended (FBI investigation casualty, 6th leading scorer, 8.8/2.4)
10) Texas - Fr F/C Mo Bamba QUESTIONABLE, toe (2nd leading scorer, 12.9/10.4/3.7 bpg)


De'Andre Hunter, ACC 6th Man of the Year, and current owner of a broken wrist. Courtesy CBS 19 Sports.

4. In the last two points, I mentioned Xavier as an overseeded team this year and then listed a bunch of injured fellas. But the coolest comeback story in the tournament this year belongs to Missouri's Michael Porter, Jr. Projected as a top-five pick in next year's NBA draft, Porter played the first two minutes of his first collegiate game with Mizzou back in November, then suffered a back injury that was expected to keep him out all season.

This was supposed to torpedo the Tigers' chances of a tournament berth, but the team--which also includes Porter's younger brother Jontay as a key contributor--played well above expectations and certainly earned their 8-seed in the tourney.

And in their conference tournament, Michael returned to the lineup.

He looked rusty and will probably be on a minutes limit, but Mizzou represents a rare case where a player's return may have them underpriced heading into the tourney.


5. One of the old clichés about March is that you win with good guard play, and the stats largely bear out that steady point guard play is particularly essential for a deep tournament run. In fact, 21 of the last 24 Final Four teams, and each of the past six champions, has featured a point guard that averages at least 3.8 assists per game.  Two of the only three Final Four teams that didn’t feature a “floor general”…both, arguably, had two of them (Traevon Jackson and Bronson Koenig for Wisconsin ’15, and Joel Berry II and Marcus Paige for North Carolina ’16). The other team was last year's surprising South Carolina squad, featuring Sensational Sendarius Thornwell and his 25 ppg in the tourney.

Over the last 6 years, teams with a floor general win an average of 1.2 games in the tournament; teams without one only win 0.7 games. That’s why it’s important to note which teams are missing this piece to their tournament puzzle. This year, the top teams that lack a floor general are CincinnatiTennesseeTexas Tech, Michigan, and Clemson.


6. This year will mark the end of one of my favorite running stats since I started putting these tips together, and it's all the bracket makers' fault. Eight teams over the past six years have entered the tournament with top-10 ratings in both offense and defense, per Ken Pomeroy and his excellent site kenpom.com. All eight of those teams won at least three games in the tournament. 

This year, two teams can boast that they enter the tournament with both top-10 offenses and defenses. And they're likely going to play each other in the Sweet 16...they are
Duke and Michigan State. So my stat is going to die AND IT'S ALL BRUCE RASMUSSEN'S FAULT, DANGIT BRUCE.

This is Bruce. Courtesy NCAA.

7. One new finding from last year involved pinpointing which of those 5/12 and 6/11 match-ups feature the high-seeded teams most ripe to be upset. The sample size on this finding is still small, but was buttressed by Minnesota's struggle in the first round last year. Over the past six years, I’ve found that 5 and 6 seeds that are quite good defensively but struggle to maintain high efficiency on offense are MUCH more susceptible to the early upset than those with good offense but questionable defense. So which of this year’s 5- and 6-seeds fit the profile to have their nice defensive numbers undone by one hot-shooting mid-major opponent?

Unfortunately, this counts as another mark against Clemson, KenPom's 48th-ranked offense and 8th-ranked defense and a team already missing their 2nd-leading scorer in the tournament.


8. Advance
d statistical site 7 Overtimes has been on a roll the last two years, correctly pegging Villanova as the nation's best team before they won the title two years ago, and having runner-up Gonzaga as last year's best team pre-tourney.

Who does 7 Overtimes like this year? Why, it's Villanova once again as their top team. Of note: Every team that 7 Overtimes has rated above a "650" in their metric pre-tourney has won at least two games in the tourney. So don't knock Nova or Virginia out of your bracket too early this year.


9. My data-centric bracket showed how important it is to nail the final in making your full resume of work look good last year. After a poor Sweet 16, I rebounded to finish in the 98.5th percentile on ESPN's Tournament Challenge after correctly picking UNC to beat Gonzaga in last year's final.
 

But hey, working on this statistical model and watching this first week of games has been the best week of the year for me since I can remember, and all this pre-tourney research does help dig up some nuggets of wisdom that I hope helped people make a few of last year’s tricky picks a little bit easier. Plus, this year, I get to talk the ear off of this little nugget, who will for sure be plopped in my lap for this weekend's games.

This is Cece. She's a goofball. Courtesy her dad.

10. So who exactly do I have pegged as the top teams and the best bets for upsets this year? Well, this year's super-intricate point system is once again based on a multiple regression model that accounts for multicollinearity wherever possible. And my state-of-the-art, Sir Crunchalot 6000 statistical model spit out the finding that this year’s best two teams......are not 1-seeds.

And, as mentioned above, they'll probably play each other in the Sweet 16.

It's Duke and Michigan State, the 2 and 3, respectively, in Kansas' bracket.

Other cool bits the model spit out this year are that there's not a ton of separation from the 5 seeds all the way down to the 12 seeds in terms of team strength. the 11-, 12- and 13-lines each have a team that's considerably closer to their better-rated counterpart than we usually see in the 6-11, 5-12, and 4-13 contests:

11: Loyola Chicago
12: New Mexico State
13: Buffalo

Did I just give away my biggest upset selections? Maaaaaaybe.

---

But the overall lesson from the 1,900 words you just read? To win your bracket, use some common sense, a couple tips from above that speak to you the most, but above all, be a little different and have some fun rooting for the Penns of the world to knock off the Kansases (NOTE: I did not pick Penn to beat Kansas). For instance, I will never pick Duke to win it all, even in my “Stats Are Terrific” bracket, because Duke blows. So make your own rules, get your brackets submitted, and relax as you settle in for the best weekend of the year.

1 comment:

  1. Bài viết tuyệt vời. Bên cạnh đó chúng tôi muốn giới thiệu cho bạn một nhà cái lô đề online uy tín bật nhất. Đặc biệt với chuyên đề chốt số giờ vàng của trang web này. Click vào link bên dưới:
    https://loto188asia.com/chot-so-gio-vang

    ReplyDelete