The Bad Luck of the Irish and Some Unlucky Ducks
08.21.06
It’s time for the annual look at luck. In order to keep things fresh, I’m throwing my old buddy Pythagoras to the curb and leaning on a new pal, Karl Friedrich Gauss, to show us the way to the luckiest and unluckiest teams in the game last season.
You see, past versions of this post used the Pythagorean method to estimate what a team’s winning percentage should have been. But now, I’ll be using the correlated Gaussian method, invented by Dean Oliver. The difference between the two is that Pythagorean averages each of a team’s offensive and defensive efficiency to come up with a winning percentage, while the Gaussian uses the distribution of a team’s game efficiencies. Since most teams have a tendency to play to the level of the competition, the latter method is little more accurate.
See last year’s episodes (1, 2), and previous versions from there for the utility of this investigation. We’re measuring luck in terms a team’s record compared to what they deserved based on their game-by-game efficiency. The idea being that any analysis of the future relies on knowing the past, which is largely determined from a team’s record. But in a few cases, that record is misleading.
With those formalities out of the way, let’s get right to it. Oh yeah, one other thing. Unlike years past, I’ll be looking at the whole season instead of just the conference slate.
It turns out the unlucky teams are more interesting than the lucky ones. With that tease, let’s start with the lucky group. Teams are ranked by the difference between expected winning percentage (as determined by Mr. Gauss) and actual winning percentage. In parentheses is the amount of wins this represents.
1. Sam Houston St. .159 (4.3) 2. Gardner Webb .158 (4.6) 3. Grambling .136 (3.4) 4. UMKC .118 (3.2) 5. Chattanooga .107 (3.0)
Yeah, nothing all that noteworthy here. Sammy H. finished 22-9. Four of those victories were against non-D1 competition, which I am not including in my analysis. Of their other 18 wins, only 2 were by more than 10 points.
Some teams to comment on farther down the list:
6. G. Washington .105 (3.2)
Winning 27 of 30 games will always rank high on the luck-meter. But the Colonials won all four of their overtime games and their 3 losses were rather decisive.
12. Hofstra .096 (3.2)
Went 5-1 in games decided by 5 points or less, just 7-3 in games decided by at least 15. Even with everyone returning, the Pride will have trouble claiming CAA supremacy.
24. Connecticut .078 (2.7)
We sort of think of UConn as underachieving, but lucky? Their 30-4 record came on 27-7 results.
Here are the five unluckiest:
330. Northern Colorado -.129 (-3.5) 331. San Jose St. -.141 (-4.4)
OK, now the good stuff…
332. Oregon -.144 (-4.6)
Another disappointing season for Malik Hairston, Aaron Brooks, and Co., right? But the season wasn’t necessarily as bad as it looked. Not when you consider that the Ducks led the nation in losses by luck. Had the breaks merely evened out for Oregon, they would have been 20-13 and perhaps an at-large team instead of postseason-less. Oregon was involved in nine games decided by 3 or less, and was victorious just once. And that doesn’t include the season-ending double-OT loss to Cal.
Don’t be surprised if Oregon’s record improves dramatically in the finale for Brooks (and the junior Hairston?), even if their play doesn’t. And if their performance improves as much as it did between ‘05 and ‘06, we’ll be talking about a team that wins 12 or 13 Pac-10 games and gets a high seed in the NCAA Tournament, despite a pathetic non-conference schedule. You heard it here first: 2007 is the year of Chamberlain Oguchi (or Maarty Leunen - I haven’t decided).
333. Notre Dame -.145 (-4.3)
Notre Dame will look a little different in ‘07, so I don’t think there’s much about ‘06 that we can apply to this season. Nonetheless, they were a team to be revered for their bad luck. The Irish were an astounding 3-12 in games decided by at most 6 points, and they dropped all five of their overtime games (including 2 double-OT games). They were better than a bunch of at-large teams, but history will record them as a second round NIT loser.
334. Virginia Tech -.148 (-4.5)
Luck comes in many forms, some obvious, some not. In the case of the Hokies it was obvious at times, as they actually lost twice when leading as the buzzer sounded. They suffered one defeat by scoring a game-losing basket on their own hoop, and in another were felled by a successful halfcourt heave. Everyone of note is back for Tech, who should have higher than normal expectations for a team that won just four of 17 conference games.
One final note: the luck factor will be added to the scouting report pages pretty soon, so you’ll be able to track this stat in real-time this season.
Fun With Database Queries
08.14.06
I think it’s safe to say that normally, when a team loses just one starter and returns its entire bench corps, we can expect that team to improve the following season. Especially when the departing player wasn’t drafted and wasn’t picked on the top two all-conference teams in his conference. (OK, technically, he was second team Big East. But the Big East mysteriously fills its teams with 10 players each. Well, actually 11 were on the first team. Anyway, media covering the Big East basically named this guy no better than the third-best point guard in the league.)
By now, you may have guessed the team in question is the Pitt Panthers. They may well be ranked in or close to the top 5 in the preseason polls. But their new team has to go through a serious adjustment period, despite being nearly identical to the old one.
Let’s look at the lone departing player, Carl Krauser, and his most likely replacement at the point.
%Min ORtg/%Poss Carl Krauser 79.2 104.5/27.4 Levance Fields 53.8 116.1/15.1
For all the rookies out there, percentage of minutes is the playing time as a percentage of the team’s total minutes. Offensive rating is a measure of a player’s efficiency. The most important figure here is the percentage of offensive possessions used. The average player uses 20% while he is on the court. (Further explanation of these metrics here.) Say what you want about Krauser’s effectiveness, but he ate up a lot of Pitt’s possessions. Among all regulars in D-I, Krauser ranked 126th in possession usage, and was 6th in the Big East.
OK, one point of clarification. Krauser and Fields shared time at the point last season, with Krauser occasionally moving to the 2 with Fields on the floor. And based on simple math, we can see that the playing time Krauser leaves behind won’t just be picked up by Fields. At best, only a third of it will. A good chunk of the rest will go to these backcourt players:
%Min ORtg/%Poss Ronald Ramon 79.2 124.6/14.4 Antonio Graves 45.9 116.1/15.8
The bottom line here is that the minutes of a high-usage player are now going to be shared by a group of role players. And that will take some adjustment. Consider that of the 191 players that used fewer than 16% of their team’s possessions in ‘05 (like Fields, Graves and Ramon did in ‘06) and returned in 2006, only 14 had a usage of at least 20% last season.
So the raw data would suggest these players aren’t likely to pick up most of the slack left by Krauser. There were exceptions to this, of course. Florida’s Taurean Green and Notre Dame’s Chris Quinn being the most notable among point guards. But again, those were exceptions.
It’s true that much of each of these players’ time on the floor was with Krauser, so they didn’t need to use a lot of possessions. Undoubtedly, their usage will increase, but at what cost? All three were efficient in their limited involvement in the offense. You have to wonder how a guy like Ramon, whose 124 ORtg was built on 41.5% 3-point shooting, will adjust without Krauser to attract the defense’s attention.
To throw another wrench into the mix, this guy is probably going to use a few less possessions this season:
%Min ORtg/%Poss Aaron Gray 69.2 105.3/28.3
Of the 50 players in 2005 that had a usage between 27% and 30% and played significantly last season, their usage on average dropped slightly in ‘06 (28.1% to 27.5%). Some of these players did increase their usage, but blind odds say we shouldn’t expect the seven-footer Gray to do so. In reality, he may pick up a few possessions out of necessity, and a little more playing time will help bridge the gap, but he doesn’t have room to get much more involved.
Maybe Carl Krauser wasn’t an outstanding point guard, but there were a lot of teams that would have liked to have him. His lack of draftability was affected by the fact he was one of the oldest college players this side of Utah.
But Krauser was a not-so-inefficient possession eater on a team full of role players. And how do I know this wasn’t a case of Krauser using possessions that his teammates really wanted? Well, I don’t, but given that Aaron Gray could go from a guy getting 12 mpg in 2005 to a guy that took 28% of possessions in 2006, I think the rest of the team would prefer to stay out of the way unless absolutely necessary. Pitt was one of only 11 teams to have two players that used at least 27% of their team’s possessions, and now that they lose one part of that duo, I can’t help but think their offense is going to sputter a little, at least for the early part of the season.
This is a case where I’d really like to know how Pitt performed with Levance Fields on the court and Krauser on the bench. Those extra possessions went somewhere during that time, but we don’t know where or how efficiently. And as usual in our offensive-minded universe, we’re ignoring defense, where Krauser had a good reputation.
Pitt will get some help on the frontline as sophomore forward Sam Young appears ready to pick up some of the offensive load. Young only played half of the team’s player, but has the stats to confirm he was not a role player and has above-average athleticism. Guys like Joakim Noah and Aaron Gray had similar profiles in ‘05 before breaking out in ‘06 (more on this to come).
With everyone on the frontline returning for Pitt, there isn’t the obvious playing-time vacuum that was filled by Gray and Noah last season. But given the uncertainty with the guards, Jamie Dixon may well find that Young needs to be on the floor more if the offense is to function as well as it did in ‘06, which it will need to if the Panthers wish to play in Atlanta in April.
Page 1 of 1 pages