Showing posts with label Data Sheet Analysis. Show all posts
Showing posts with label Data Sheet Analysis. Show all posts

Tuesday, January 21, 2014

Data Sheet Analysis: WCC Ratings Roundup, January 21st


Don't Understand This Chart? I'll Make Sense After Reading This Post...Hopefully



Though it's not even February yet, it's never a bad time to talk about postseason expectations. One thing I liked to do leading up to tournament time is looking up different ratings of basketball teams that are on "the bubble". In the WCC, it's usually only 2-3 teams, but nonetheless, I wanted to look at the conference as a whole and the ratings of each and see where there were major differences between ratings systems for certain teams. For a lot of the top teams in conference, the ratings were pretty consistent across the board (as I will show in a graph via the Data Sheet), but things got interesting as you look at the middle of the pack to the bottom of the WCC.

Before I go into the analysis of each team, I wanted to give people a brief synopsis of each ratings system I used leading up to the team analysis.

RPI (Rating Percentage Index)

  • The most common and widely known ratings system. Used in the NCAA selection process. The formula for RPI is  (WP * 0.25) + (OWP * 0.50) + (OOWP * 0.25) (where WP is Winning percentage, OWP is opponent's win percentage and OOWP is Opponents' Opponents' Winning Percentage).
  • ESPN's own ratings index similar to RPI created in 2012. BPI accounts the most factors in games in comparison to the other Major Indexes, as it includes scoring margin, diminishing returns for blowouts, pace, game location, SOS beyond Opponent W-L, Wins (not really quality losses), and Key Players missing. While the positive is that it accounts all factors, the negative with BPI is that it can be the most subjective of all the ratings indexes with all the factors it accounts. For a more detailed explanation, check out this post here.
  • More mathematical than the above, but in my opinion very informative in terms of "pure prediction" as he states on his "Ratings Explanation" page. Pomeroy seems less concerned with past results (since he knows upsets will happen and it's impossible to predict every upset or win), but in his words "to show how strong a team would be if it played tonight, independent of injuries or emotional factors. Since nobody can see every team play all (or even most) of their games, this system is designed to give you a snapshot of a team’s current level of play." The core of Pomeroy's system is based on the "pythagorean calculation for expected winning percentage", though he has incorporated some changes to his algorithms going into this year. Overall, Pomeroy's seems to be for the most mathematically or statistically inclined, but I like it the most because I find the data accurate in reflecting a team's profile.
  • Sagarin's ratings are not only used by the NCAA Tournament committee, but they have also been a staple among oddsmakers in terms of determining lines for basketball games (and football, as his football ratings have been part of the BCS system since its inception). While the exact formulas for Sagarin's ratings aren't exact, according to Wikipedia (yes, I am quoting Wikipedia, shoot me English Teacher Nazi) his system is "the difference in two teams' rating scores is meant to predict the margin of victory for the stronger team at a neutral venue...teams gain higher ratings within the Sagarin system by winning games against stronger opponents, factoring in such things as home-venue advantage." Thus, Sagarin is more aimed toward the NCAA Tournament as the idea of playing on a neutral court is something that is rarely important in the regular season and more important come Tournament time (where all the teams play on neutral courts).
  • Massey's ratings are more predicated on measuring past performance rather than predicting future results. This emphasis varies from other ratings such as Pomeroy and Sagarin (which state their goal is to predict future results, albeit in different ways). Massey states two challenges, as he looks to eliminate "noise which obscures the true strength of a team" (which includes officiating, luck, not playing to potential, etc.) and differences in schedule. Thus, strength of schedule is heavily incorporated into the ratings, though Massey only factors in score, venue and date and no other statistics or factors in the strength of schedule rating.

After looking up all those ratings on their various Web site, I composed an aggregate rating which basically was an average of all their ratings combined. That way, the aggregate rating gives a true "real" rating of each WCC team, since some fluctuate surprisingly a lot.

Okay, with that out of the way, let's analyze each team in the WCC and their ratings thus far through the season.


Gonzaga Bulldogs (16-3 overall, 6-1 in conference)
  • RPI: 25
  • BPI: 21
  • Ken Pom: 26
  • Jeff Sagarin: 21
  • Ken Massey: 29
  • Aggregate Rating: 24.4

BPI and Sagarin love the Zags, as they rate them at 21 in both ratings. BPI is not surprising considering all the subjective factors they take in. The Zags have huge margin of victories in many games (especially during their first 3 games in WCC play against St. Mary's, USF and Santa Clara), have been hit hard by injuries (Gary Bell and Sam Dower), and have played some tough neutral court games (Maui Classic and Kansas State in Wichita). Sagarin also takes in margin of victory, and that too probably also explains why that rating also correlates similarly with ESPN's BPI.

Massey and Pom seem less in love with the Zags, as they both rate them higher than their aggregate rating. The Zags rate 17th in AdjO and 71st in AdjD, so when you factor in both those factors, as well as their 67.4 pace factor, and the Zags rating at 26 seems about right (SMU, which is rated 25, is similar in their AdjO and AdjD difference, though is flipped, as their D is better than their O). As for Massey, which had the lowest rating for the Zags out of any of the ones listed, it relies so heavily the teams' records and opponents records, and the strength of many of the opponents outside of Arkansas, West Virgina (though that's questionable), Kansas State and Dayton isn't all that great (New Mexico State started well early on but have faded and that has hurt the Zags' strength of schedule, which in return has hurt the Zags).


BYU Cougars (13-7 overall, 5-2 conference)
  • RPI: 39
  • BPI: 49
  • Ken Pom: 49
  • Jeff Sagarin: 41
  • Ken Massey: 55
  • Aggregate Rating: 46.6

RPI loves BYU and it's not surprising. The Cougars played the best out of conference schedule (7th hardest in the nation) out of anyone in the WCC. Stanford, Wichita State, Utah State, Iowa State, Texas, Utah, Oregon and UMass were all on BYU's non-conference slate and all those teams are Top-100 teams in almost every rating listed above. Though the Cougars went 3-5 against them, three of those teams started out the season with double digit wins before suffering their first loss (Iowa State, Oregon and Umass) and one is still undefeated (Wichita State). Thus, it makes sense that BYU's highest rating came in the RPI.

Massey was the least loving of the Cougars (though that tended to be the trend for many WCC teams, as the conference as a whole rated low in Massey's ratings). My guess for the low ratings is the Cougars' opening conference losses to LMU and Pepperdine (who rank low in the Massey Ratings) hurt, as well as that Dec.14-Dec.30 stretch where they lost four straight. Ken Pom and BPI also rated the Cougars above their aggregate rating, which makes sense considering the Cougars' mediocre defensive rating (86th in KP) and high pace (76.0). Still, the Cougars have had a relatively easy road so far in conference play (10th rated in conference SOS on Ken Pom), so it'll be interesting to see if this improvement is legit or just the product of an easy stretch.


St. Mary's Gaels (14-5 overall, 4-2 in conference)
  • RPI: 50
  • BPI: 47
  • Ken Pom: 56
  • Jeff Sagarin: 51
  • Ken Massey: 65
  • Aggregate Rating: 53.8

If you want to make the NCAA Bubble Case for the Gaels, you certainly can. They sneak in when it comes to RPI and BPI. Their opening win over Louisiana Tech may have been their most underrated win of the season so far (though LA Tech's loss to Southern Miss dampened it a bit) and wins over Boise State and North Dakota State are solid wins over solid mid-major teams who have a chance to make the NCAA Tournament (albeit by winning their conference tournaments, but still a chance nonetheless).

But, the Gaels look bad in Ken Pom and Ken Massey's ratings. In KP's ratings, their soft non-conference schedule (only two away games and two neutral games in the non-conference) and mediocre defensive efficiency (175th in the nation) hurts their case big time. As for Massey, a home loss to Santa Clara (it takes heavily into consideration home record) and a slow start in conference (4-2, unimpressive difference in Points Score and Points Allowed in Conf Play) probably is the key reason why they are barely on the bubble in his ratings.


Pacific Tigers (10-7 overall, 1-5 conference)
  • RPI: 84
  • BPI: 115
  • Ken Pom: 121
  • Jeff Sagarin: 116
  • Ken Massey: 117
  • Aggregate Rating: 110.6

I really believe the Tigers are due for a second-half turnaround in WCC play. Though they currently rank in the basement in the WCC with a 1-5 conference record, their RPI is surprisingly high for a last-place team, and they played well in the non-conference slate as well with good wins over Utah State and UC Irvine and tough games against Oregon and Princeton (who really is a contender to Harvard, though it is Harvard's conference to lose at this point). Furthermore, they have already played some of the toughest WCC opponents thus far, as they have played St. Mary's twice and Gonzaga in Spokane (all boosts to that RPI).

Now, I don't believe the Tigers are a top-100 team. But even when you look at their other ratings, they are obviously seen as a middle of the pack WCC team, not a last place team like the one you see currently. Even Ken Pomeroy, who has the Tigers rated the lowest out of all the ratings, still has them as the fifth highest rated team in the conference, and did rate their conference schedule as the hardest thus far for any WCC team. So, don't give up on this Tigers team yet, even though they certainly didn't do themselves any favors with that 1-5 start in conference play.


San Francisco Dons (12-8 overall, 5-3 conference)
  • RPI: 102
  • BPI: 121
  • Ken Pom: 124
  • Jeff Sagarin: 126
  • Ken Massey: 114
  • Aggregate Rating: 117.4

A pretty consistently rated team overall. The RPI is solid for their record, but it's not egregiously different from the other ratings like Pacific (where they had a 37 point difference between RPI, their highest ranking and Ken Pom, their lowest rating). USF was hurt from questionable losses early on, and it shows in their out of conference RPI of 177, which is third-worst in the conference. Ken Pom also rates their defense 263rd in the nation AdjD efficiency, so that is also another reason why KP has them hovering in the 120 range, well out of NIT territory and maybe even CIT or CBI consideration as well.

That being said, USF's offense is good (40th in the nation in AdjO efficiency) and they have played much better as of late with four wins in their last six conference games, and the two losses being on the road at St. Mary's and a tough loss to BYU (who is playing the best ball in conference). Those losses won't hurt, and the Dons only other conference loss is to Gonzaga on the road (though that was a BAAD conference loss as they were killed). If the Dons can keep winning WCC games against the middle of the pack (they already got a huge boon in a road win over Portland which is looking better and better with Portland's win over Gonzaga at the Chiles Center), and maybe pull one home upset over St. Mary's or Gonzaga (Winning at BYU would be tough), then I think the Dons could get a high enough status in the ratings to earn some kind of postseason berth.


Portland Pilots (11-8 overall, 3-4 conference)
  • RPI: 152
  • BPI: 107
  • Ken Pom: 104
  • Jeff Sagarin: 109
  • Ken Massey: 126
  • Aggregate Rating: 119.6

The Pilots earn the award for the most bi-polar team in the Ratings Roundup. RPI does NOT like Portland and rightfully so. Their projected OOC RPI is 169 and projected OOC SOS is 290. They are 7-2 against teams rated 200 or higher, but 6-6 against teams in the 51-100 range. The Pilots boast a profile with good wins over Gonzaga and Princeton, tolerable losses to Oregon State and Michigan State, and downright crappy losses to North Texas and Montana State. Portland may be a good team underneath all that fluff, but RPI is going to have a hard time seeing it, especially with those strength of schedule numbers hanging out there.

But sans RPI, the Pilots still profile as a decent team and have a good shot for the fourth spot in conference if they can use the momentum from the Gonzaga win to turn it around. Much like Pacific, they suffered from some bad losses early (though the conference strength of schedule wasn't nearly as difficult as Pacific's; it is rated as the 8th toughest so far) and while they don't excel in anything, they aren't really terrible in anything either (they rank 116th in AdjO and 123rd in AdjD according to KP). So, they'll always be a competitive team, though they'll probably lose some games they probably should win (as evidenced by losing to Pepperdine on Saturday, though that was in Malibu). While I like USF more, I think Portland has a decent chance to make a postseason berth if they can get some consistency.


Pepperdine Waves (12-8 overall, 5-3 conference)
  • RPI: 118
  • BPI: 135
  • Ken Pom: 126
  • Jeff Sagarin: 121
  • Ken Massey: 113
  • Aggregate Rating: 122.6

As we get to the bottom of the WCC, we start to see some trends emerge. When it comes to the top of the conference, the RPI is consistently higher than the Massey ratings. For the lower-tier WCC teams, the Massey ratings are actually better than their RPIs. Because BPI is so subjective and incorporates so many variables, it goes a variety of ways, and Pom and Sagarin are more scientific, so they kinda fall in line with BPI as well.  Pepperdine is pretty typical of what we see from the lower tier teams, though to be honest, they may be on the cusp of being a middling to borderline upper tier team, if they could just pull some sort of consistency and give fans a real idea what team they're going to be for the rest of the season.

Pepperdine suffers from many of the same problems of the lower-end WCC teams: bad OOC RPI (202, second worst in WCC), bad OOC SOS (282, third worst in WCC) and lack of quality wins (2-8 against top-100) along with some REALLY bad losses (San Jose State and Cal State Fullerton are rated 252nd and 241st respectively in KP's ratings). What boosts Pepperdine (or why you might argue in favor of Pepperdine) over Portland though is that the Waves have actually been better against teams ranked in the 101-200 range (8-4 to Portland's 6-6 mark). Add that with the fact that the Waves actually are better than average in something (86th in Adjusted Offense), and that gives the Waves a bit of an advantage over a team like Portland who is just pretty average all across the board. Yes, the Pilots have the better ratings now, but check back in two weeks and I have a feeling Pepperdine will have leaped them.


San Diego Toreros (11-9 overall, 2-5 conference)
  • RPI: 189
  • BPI: 160
  • Ken Pom: 150
  • Jeff Sagarin: 156
  • Ken Massey: 171
  • Aggregate Rating: 165.2

Initially, the feeling is to say that the Toreros are the worst team in the WCC. While Pacific is last, there is the feeling that they have just been unlucky. The Tigers played some decent teams in the non-conference, and as stated before, played the toughest conference slate of any WCC team thus far. Add that with the adjustment that comes with playing in a new conference (especially a conference upgrade to the WCC from the Big West), and it makes sense why the Tigers have struggled this season. But as for the Toreros? They played two Non-DI teams (inexcusable in my mind any way you cut it), have a projected OOC SOS of 276, and are 1-9 against Top-100 teams this season. Their 189 RPI is second-worst in the conference, and by all means, it is easy to categorize San Diego as the bottom feeder in the WCC.

But surprisingly, the Toreros aren't as bad as advertised. Sans a bad loss to Illinois Chicago on a neutral court (282nd in KP's ratings), the Toreros have taken care of business against bad teams (7-1 against teams rated in the 200 plus range). Their OOC RPI isn't as bad as expected at 170 (better than both USF and Pepperdine), and the advanced guys (Pomeroy and Sagarin) like them more than most, which means statistically, they have some potential (though certainly not a lot of it). San Diego's slower pace (second-slowest in WCC) also helps them in the more variable ratings (especially BPI and KP), so I can see why they rate higher in ratings other than RPI. So, the Toreros are bad...they're just not as bad as we want to think just yet.


Loyola Marymount Lions (10-10 overall, 2-6 conference)
  • RPI: 164
  • BPI: 181
  • Ken Pom: 161
  • Jeff Sagarin: 165
  • Ken Massey: 166
  • Aggregate Rating: 167.4

Oh how they mighty have fallen. Remember how sky high people were riding on the Lions after they whooped BYU in Los Angeles? Remember how the Lions were threatening to make their case as the conference's second or third best team? Well those days are long gone. According to the aggregate ratings, the Lions rank as the second-worst team in the WCC, and with their 2-6 record in the WCC, the ratings seem pretty spot on in this case. The Lions aren't just bad in one set of ratings, but go for all-around effect in terms of their lower tier in the ratings (though BPI seems to be the most lackluster at 181), thus cementing their spot currently in the basement of the WCC.

If you want to find any silver linings for LMU, Ken Pom rates them the highest at 161st, noting that the overall AdjO rating at 106.1 is actually not bad at 134th in the nation despite them ranking last in the conference in offensive rating in WCC games alone (100.4 offensive rating in conference games only). And, the Lions' OOC SOS was actually better than most WCC teams in the lower tier (241st, which would rank as the 6th toughest OOC SOS in the WCC). Add that with an OOC RPI that was rated as 117th and the Lions probably have been a victim of bad luck rather than just mere bad play (though there has been plenty of that going on for Max Good's squad). RPI-Forecast projects the Lions to finish with a RPI of 158.5. While that is not a huge difference, I think that is a sign people should expect the Lions to trend up not down during this second-half stretch of WCC play.


Santa Clara Broncos (10-11 overall, 3-5 conference)
  • RPI: 193
  • BPI: 184
  • Ken Pom: 178
  • Jeff Sagarin: 184
  • Ken Massey: 184
  • Aggregate Rating: 184.6

Thank God Kerry Keating signed that contract extension this year or he'd be in serious trouble. I get it, he has a couple of CBI titles to his belt, but he hasn't made the NCAA tournament, kind of a big deal considering the coach the administration pushed out (Dick Davey) to make room for him had numerous on his resume. That being said, it makes sense though that the athletic department signed him to a deal so early this season: it was going to be a rebuilding year after losing Kevin Foster and a few other key players, and Even Roquemore simply isn't the player efficiency-wise to carry a team. Thus, it makes sense why the Broncos rate as the worst team ratings wise so far, even though they're holding their head above water a little bit with a 3-5 conference record.

Let's look at the good: the Broncos have played the second toughest conference schedule so far according to Ken Pom, so things are bound to get easier down the stretch (they have already played Gonzaga and St. Mary's on the road and they played BYU recently at home). Their win over St. Mary's on the road may be what plays spoiler to the Gaels' chances for an at-large bid in the NCAA Tournament, so at least they can torment their Bay Area brethren in that regard. Other than that though? 9-9 record against teams rated 100-plus (i.e. not good), 292nd in OOC SOS (worst in WCC) and 256th in OOC RPI (also worst in WCC). And, terrible losses to Cal State Fullerton (241) on the road and Rice (300!) at home. Yes, as bad as it is to hear Rough Riders (all 50 of you...sorry, just kidding), you are indeed probably the worst team in the WCC this year unfortunately. Just hope it's only a one year thing.


Candidates to trend UPWARD in ratings:
  • Pacific
  • Pepperdine
  • Loyola Marymount

Candidates to trend DOWNWARD in ratings:
  • Portland
  • BYU
  • Gonzaga


Wednesday, January 8, 2014

Data Sheet Analysis: A Look at How WCC Coaches Utilize Their Benches

If bench minutes history over the past four years means anything, those Gaels in the warm ups know their time on the floor is few and far between with coach Randy Bennett


I was looking up some data on Ken Pom.com (I know I'm addicted to that site), and I found something interesting. This season, many WCC programs rank near the bottom of the nation in terms of utilizing their bench. Only one team ranks in the Top-100 in terms of bench minutes percentage (Pacific), a stark contrast in comparison to years past from WCC teams. While you can access the Data Sheet on Bench Minutes on the "Data Sheets" Page, I'll post the numbers right here to give you a context of the lack of time the bench players see in the WCC.

TeamBench Min %Nat'l Rank
Pacific37.365
Santa Clara34.2120
Loyola Marymount33.8130
Pepperdine33.7132
Portland30.6201
San Francisco30.1217
Gonzaga28.7253
BYU28.6256
San Diego27.5274
St. Mary's23.1333

Now, there are a variety of reasons why certain coaches utilize their bench less than other rival coaches. For some, it all can depend on a coaches' style. A team that presses more, plays more in transition is more likely to use their bench to keep fresh legs on the court to play to their system. This is the case for a program like Arkansas who leads the nation in bench minutes percentage at 45.3 percent. Mike Anderson, a former Nolan Richardson disciple, plays a full court, high-pressure defensive style (i.e. "40 Minutes of Hell"). In order to maximize his team's effectiveness (especially on the defensive ends), constant bench usage is necessary. Of course, the correlation between tempo and bench isn't always the case as BYU plays the fastest tempo in the nation (77.6) but is in the bottom when it comes to bench minutes usage.

One big misconception though is that talent depth is a reason for more bench minutes usage. Teams who have more talent on the roster, should surely use their bench more than programs who are thin talent wise. So, if this thought is true, a team like Duke would be utilizing their bench than say a program like Montana State. That isn't necessarily the case. When it comes to the Top-50 for teams who use their benches the most, only six (Arkansas, Purdue, Iowa, Texas A&M, Houston, Georgia, Louisville, Oregon State) came from major conferences, and only one (Louisville) made the NCAA Tournament a year ago (though Iowa was certainly deserving). The same proves to be true in the WCC. The top-3 favorites (Gonzaga, BYU and St. Mary's) in the conference rank in the bottom of the conference in terms of bench minutes percentage while the ones at the top were relative afterthoughts going into the season (though Pacific did make the NCAA Tournament a year ago albeit as a Big West representative, a much weaker conference than the WCC).

But one year is just one year, right? After all, talent changes quite often and systems may differ depending on personnel. So, I wanted to take a look at which coaches tended to use their benches the most, and which ones did the least. To find this out, I went to Ken Pom.com, found each team's bench minutes percentage from 2011-2014 and then averaged out the four year span. The results are shown in this graph:























As you can see the four year results provide some interesting trends. For starters, Randy Bennett seems to be the leading coach in the WCC who trusts his bench the least. His 26.4 percent four year average for bench minutes percentage correlates to the 23.1 percent bench minutes percentage we have seen from his squad this year (though even for his standards he trusts his bench even less so this year, as it is 3.1 percent below his four year average). Rex Walters has sort of employed the same "shallow rotation" with his Dons squads, as his 29 percent four year average is was second-least in the conference. Considering the amount of transfers Walters has seen in his tenure on "The Hilltop" it's interesting to wonder if his "strict rotation" (or lack of bench utilizaiton) was a reason for a lot of his transfers or just one of the effects.

LMU leads the WCC in bench minutes percentage over the four year span with an average of 34.1, barely edging out Eric Reveno and Portland who finished with a four year average of 34. Portland's finish is interesting because they never led the conference in bench minutes percentage from 2011-2014, but they finished second in the conference in 2012 and 2013, so that explains their close finish to LMU, who led the conference in bench minutes percentage in 2012 (they had the 37th highest bench minutes percentage in the nation that year as well).

The most eye-popping case is at Gonzaga with Mark Few as his bench minutes percentages have fluctuated wildly over the past four years. As you can see by his four year average, Few obviously numbers and minutes-wise utilizes his bench a decent amount, but that sharply contrasts from what we have seen this year (hence, their third-place finish in the four year average with an average of 33.7). In 2011, the Zags led the conference in bench minutes percentage at 38.5 percent (30th highest in the nation). Last year, they ranked fourth in the conference but still finished 81st in the nation (2013 was a year for bench players in the WCC as only BYU was outside the Top-200 in terms of bench minutes percentage) with a percentage of 34.8. But Few at times has showed years where he has relatively little trust in his reserves. In 2012, he trusted his bench less than the 4 year average, as the bench only averaged 32.7 minutes and this year, he showed even more of a lack of trust, as the Zags reserves are averaging only 28.7 minutes per game. Now, the lack of depth in the front court could be the main cause of this, but it is interesting to see how minutes have plunged at times in the Zags program, compounded by the fact that they have had many transfers over the past few years which ultimately led to this lack of depth for this Zags squad (just look at Gonzaga transfer Ryan Spangler who is starting for Oklahoma now). Few has showed a short leash with players on the bench, even though his teams have been the deepest in the conference talent-wise for years. Even this year's team has a lot of depth on the perimeter, but as displayed from the numbers, he doesn't utilize them as much as other, lesser-talented teams in the conference.

All these numbers are up for interpretation and as stated before, there are lots of reasons why certain coaches utilize their bench more or less. Whatever the reason, it is interesting to see the numbers laid out, and see how each coaches' process will fare out for them over the conference season and into conference tournament play and beyond.