 Read the daily ramblings of Pete Tiernan as he discusses the strategies behind BracketScience.
2012 tourney will be won by one of the six teams foretold with champ stats |
|
Sun,
Apr
1,
2012,
12:38 PM
|
|
2 Comments...
|
Throughout the season, I tracked potential tourney champs on the basis of eight statistics. Every one of the last 10 champs have:
• earned a one, two or three seed
• come from a Power conference (ACC, Big East, Big Ten, Big 12, Pac-10 or SEC)
• been led by a coach with more than five tourney trips and at least one Elite Eight run
• either went to the previous year’s dance or had an All-American
• averaged more than 73 points per game
• allowed fewer than 73 points per game
• owned an average scoring margin of at least seven points per game.
• played a schedule among the 75 strongest in the country
Only six teams were on the champ list for the entire season. Five of them reached the Elite Eight, three made the Final Four--and now two are vying for the championship. The six teams were: Kentucky, Syracuse, Kansas, North Carolina, Duke and Ohio State. Only Duke significantly underperformed.
If you had used these champ credentials in your bracket building, you probably would've done alright. Interestingly, the one region that was the most unpredictable was the West. And none of the teams had the statistical chops to be champs. MSU wasn't high scoring enough, and Missouri and Marquette didn't have the coaching qualifications to cut down the nets. So that left a void with the high seeds in the region, and an opening for Louisville.
|
|
2 Comments...
|
|
Great Eight Challenge winner comes down to last game |
|
Sun,
Apr
1,
2012,
12:20 PM
|
|
0 Comments...
|
This year's Great Eight Challenge won't be decided until the 2012 tourney winner is crowned tomorrow. If Kentucky wins, then andolinojt63 wins the iPad. If Kansas cuts down the nets, then andrewchen gets the prize. As I recall, this is the first time in the Great Eight Challenge that the winner hasn't been decided until the final game. Good luck, andolinojt63 and andrewchen!
|
|
0 Comments...
|
|
2012 results will change the performance rankings of several veteran coaches |
|
Wed,
Mar
28,
2012,
9:09 PM
|
|
4 Comments...
|
It’s too early to provide definitive numbers on how the 2012 tourney has changed the PASE values the top college coaches. But there will be some big shakeups.
There are 78 active coaches with four or more dance trips—eight who just made their fourth appearance. I excluded them from this analysis and focused on the top 25 coaches who made the 2012 tourney. The following table gives an estimate of how their performance against seed expectations (PASE) will change once the dance is done. (I can’t give precise numbers until the tourney is complete.)
The first column in the table shows what the coaches new rank would be among 70 veteran coaches after 2012. Column two shows what the rank was heading into the tourney. And the third column shows the difference in the rankings. The biggest improvements are in green…and the biggest drops are in red.
Right now, Thad Matta has added the most to his PASE, not only climbing 13 positions, but also changing from an overall underachiever to an overachiever. Bill Self has improved his ranking by the second-most number of positions. And Billy Donovan and Rick Pitino are tied for third.
Of course, the PASE values of Matta, Self, Pitino and Calipari could all go higher. If the Buckeyes cut down the nets, Matta’s PASE could go up to +.313. That would vault him to 13th place, a leap of 31 positions! If the Jayhawks win the title, Self’s PASE would jump to +.265, making him the 18th best overachiever, an improvement of 20 spots. If Louisville wins the tourney, Pitino’s PASE would improve to +.720 and he’d leapfrog Billy Donovan to claim fourth place in the overachievement rankings. Finally, a Kentucky win would boost Calipari’s PASE to +.329, and he’d be the 13th best performer, climbing 13 spots in the rankings.
Six coaches suffered big slides in their PASE fortunes, none more steep than John Beilein’s. Michigan’s opening-round loss to Ohio lopped .311 off Beilein’s PASE (that’s the number in the last column). And he dropped seven positions in the rankings, from fourth to 11th. Davidson’s Bob McKillop will fall five positions, coming down to earth after he got the “Curry bump” in 2008. Jim Boeheim will also fall five positions—an unfortunate decline given the Melo injury. Leonard Hamilton, Steve Fisher and Fran Dunphy has also suffered precipitous drops.
Those advances and declines refer to the rankings. If we just evaluated the coaches on the changes in their PASE values, Matta, Donovan and Pitino would be the biggest gainers (so far). And Beilein, Fisher and Izzo would be the biggest losers (all with a connection to the state of Michgan). Other big gainers include Bill Self and Tom Crean. And the next biggest losers are Coach K, McKillop and Leonard Hamilton.
|
|
4 Comments...
|
|
Models in top 72nd percentile overall…but none would win your pool |
|
Sat,
Mar
24,
2012,
10:56 AM
|
|
13 Comments...
|
This year’s models have done better than 2011, but none of them would win a moderate-sized tourney pool. Here are the percentile rankings based on the ESPN Tourney Challenge:
Model 3 Final Four/champ – 97.2%
Model 5 Factors – 92.3%
Model 7 Pythag/Coaching – 89.8%
Model 9 Seed match-up – 85.5%
Model 10 Outcome match – 81.6%
Model 1 Gut – 81.6%
Model 6 Pythag – 77.3%
Model 8 Pulsecheck – 77.3%
Model 11 Contrarian – 67.6%
Model 12 Keeper – 51.6%
Model 4 ComboPASE – 41.2%
Model 2 Upset-toss-up – 17.6%
It’s encouraging to see that my original stats model, Final Four/Champ, is the top performing model this year. It’s really been the most reliable approach to filling out the bracket. This is probably the end of the glory road for that model though. Its Final Four is Baylor (ooh…), Michigan State (ouch), Ohio State and Kansas. The healthiest of the models is probably Factors of Pythag/Coach, which still retain three of four semifinalists.
If you divide these models between stats-driven approaches and opinion-based picks, the stats models outperform my instincts 73.3% to 66.9%. But I did change my Gut picks at the last minute based on the Melo injury. If I hadn’t done that, Model 1 would by tied for first at 97.2% and I’d still have Kentucky, Syracuse and Kansas in play. But I reserved the right to spend one minute contemplating the Melo injury, had the Orange going out against Kansas State and advancing Florida State. I should’ve stuck with my original gut.
All in all, not a bad showing for the models, both nothing earth-shattering. Hopefully, everyone took these as guidance and tweaked to suit your own ideas.
|
|
13 Comments...
|
|
2012 Madometer hovers near record insanity...but won't get there |
|
Sat,
Mar
24,
2012,
9:27 AM
|
|
0 Comments...
|
Heading into today's Elite Eight round, the Madometer is holding at 19.3%. That's pretty darned close to last year's all-time record of 19.8% madness. But it won't get there, no matter what happens. If the lower seed won from here on in, we'd have a Florida/Baylor/OSU/Kansas Final Four, or Florida/OSU or Kansas final, and a Florida champion. I don't see that happening, but even if it did, the Madometer would only nudge up to 19.7%--one tenth of a percent below the record.
If the high seeds win from here on in, we'll wind up with 16.7% insanity. That's still pretty darned mad. Only eight of the last 27 dances have been crazier. My guess is we'll end up with a Madometer reading around 17.1%, the same reading as 2010. Baylor or Florida would have to make some noise for it to go higher.
On the upset front, after seven shockers in the first round, there was only one upset in round two (NC State over Georgetown), and one in the Sweet 16 (Florida over Marquette). That's nine. The most we could have in 2012 is 12. And that would only happen if Florida wins out. No matter what happens, we won't reach the record total of 13 achieved in 1985, 1986, 1990, 2002 and 2011.
|
|
0 Comments...
|
|
My main man Herb Maeder calculates all the possible outcomes for the Great Eight Challenge |
|
Sat,
Mar
24,
2012,
8:46 AM
|
|
0 Comments...
|
I’m blessed to have some amazing members. From all the guys who keep the comments section cooking to the people who send me bracket theories, there are scores of people I’ve gotten to know over the years through Bracketscience. Some of them I know by name; others I know by their alias. One of them is my friend Herb Maeder.
Herb makes an amazing contribution to the site every year. He’s a programmer who, out of his passion for bracket pools and the goodness of his heart, calculates out all the possible winning scenarios for the Great Eight Challenge. He’s hoping that he could develop this program on a larger scale for those million-user contests, so people could see beyond their scores to their actual chances of winning.
What a great thing that he’s a member of Bracketscience. Here are the results of his work, along with a description in his words of the site’s features.
http://pooldoggers.com/pools/bs_great8/ncaa .
The main page has a link to the "Current Standings,"the "What If Scenarios" followed by the complete "List of Entries" and an indicator of whether each entry has a chance of winning the pool or not.
The pool standings will be updated automatically within about 15-30 minutes after the end of each game.
At this point in the tourney, it is feasible to compute all the possible outcomes of the tourney and of the Great 8 Challenge. Of the 588 entries, five still have some shot at taking first place in the pool. After the Sweet 16, 33 people could’ve won.
The "Current Standings" page give a list of the best possible picks at this point in the tourney followed by the list of entries, ranked in order of the current score. The number in parentheses is the "seed total" of all the picks in the entry, which will be used as a tie breaker if necessary. The entry name is link to the list of picks for that entry. There also a summary of the number of "dead" picks (e.g. 3/8 indicates that three out of the eight picks are dead) followed by a list of alive/dead seeds for the entry.
The "Must Win Games" link provides a list of games that are are absolutely critical to win in order for the entry to have a shot at winning the pool. If any of the teams listed loses, the entry will no longer have a chance to win. Some lines list multiple teams, if any of those teams wins, the entry may still have a chance at winning. The list of "Must Win Games" changes as each game completes, and the list of "Must Win" games usually increases. So check back often to see which games are critical for each entry.
At the bottom of the "Must Win Games", there is an indicator of how many ways there are for an entry to win. For example, in the snippet below, there are 128 ways the tournament can end from the Elite Eight. Of these, "cborchwis" will win 76 of them.
Based on Herb’s amazing work, we now know that these five entries can win the Great Eight Challenge (number of winning options in parentheses).
Cborchwis - 76
Andrewchen - 26
Vitamin - 16
andolinojt63 – 5
Jeffreypeters21 – 5
Thanks so much, Herb, for all your help!
|
|
0 Comments...
|
|
Coaching tourney performance could’ve guided you to a winning bracket |
|
Mon,
Mar
19,
2012,
8:37 PM
|
|
9 Comments...
|
Six of the eight multi-year stats models I run are above the 70th percentile in the ESPN Tourney Challenge. Two of them are in the 92nd percentile and have retained all their Elite Eight picks. I’m happy to see that my oldest model, the Final Four/Champ strategy, is one of those two. But the other model is more interesting to me. That’s the “Pythag and Coaching PASE” model.
I hit me late last night that every single one of the snake-bitten coaches—those with more than five tourney trips and no Elite Eight runs—was already out of the tourney. That includes Leonard Hamilton, Gregg Marshall, Steve Alford, Kevin Stallings, Fran Dunphy, Mike Brey, Matt Painter and Dave Rose. Eight coaches—and eight disappointing runs.
Then I dug a little deeper. All but five of the Sweet 16 coaches are historical overachievers. That would be OSU’s Thad Matta, NC State’s Mark Gottfried, Cinci’s Mick Cronin, Indiana’s Tom Crean and Xavier’s Chris mack. But OSU got to the second weekend through Gonzaga and Mark Few—a -.021 PASE underachiever. Gottfried beat John Thompson II, a -.275 PASE underperformer. And Cronin beat the snake-bitten -.259 PASE underachiever Leonard Hamilton. Crean managed to avoid an upset by Shaka Smart and his gaudy +3.463 PASE, but his Hoosiers just barely got the job done against a 12 seed. And Mack’s X-men struggled with 15 seed Lehigh.
Even if you go back to the first round, coaching overachievers bested underachievers: Majerus of St. Louis over Pastner of Memphis and Donovan of Florida over Bennett of Virginia. And underachievers bested bigger underachievers (Xavier’s Chris Mack over Notre Dame’s Mike Brey). And bigger overachievers best overachievers (Jim Groce’s +.833 PASE was better than Beilein’s +.705).
So if coaching PASE is the key to this year’s dance, who’s on track to advance to the Elite Eight and Final Four? If you took the coach with the higher PASE in all match-ups, here’s who would win out:
Sweet 16
- Kentucky over Indiana
- Baylor over Xavier
- Michigan State over Louisville
- Florida over Marquette
- Wisconsin over Syracuse (by a measly +.011 PASE)
- Ohio State over Cincinnati
- Ohio…yes Ohio…over North Carolina (but let’s advance NC because previous long-odds upset games haven’t gone to the overachieving coach)
- Kansas over NC State
Elite Eight
- Baylor over Kentucky
- Michigan State over Florida
- Wisconsin over Ohio State
- North Carolina over Kansas
Final Four
- Michigan State over North Carolina
Champion
- Michigan State—and the magical Izzoman
|
|
9 Comments...
|
|
Madometer still on pace for record madness |
|
Sun,
Mar
18,
2012,
11:16 PM
|
|
3 Comments...
|
The Madometer reads 22.1%, still ahead of last year's record madness of 19.8%. If there were no more low-seed wins the rest of the way, the 2012 dance would still register 15.2% insanity--tied for the tenth craziest out of the 28 modern-era tourneys.
So far, there have been eight upsets. The average tourney features 8.5 shockers. In the Sweet 16, five of the eight games have upset potential. We'll need something pretty nuts to get to the record of 13 upsets in a single dance.
|
|
3 Comments...
|
|
Madometer reaches an all-time high after seven upsets, two 2v15 shockers |
|
Fri,
Mar
16,
2012,
11:39 PM
|
|
9 Comments...
|
|
What a day! After two mild upsets yesterday, we had two--count 'em--two 2v15 upsets, a 4v13 upset, a 5v12 and a 6v11 upset. All that made the Madometer go bonkers. I had to redesign the graphic on the home page because it didn't go past 20% madness. Well...at this point, it reads a 25.8% deviation from perfect high seed dominance. Even if the higher seed won from this point forward, we'd still see a pretty sizeable final number of 13.1%. That's already more crazy than 10 of the last 27 dances. It's bound to get uglier. Stay tuned.
|
|
9 Comments...
|
|
|
|
