PDA

View Full Version : Anyone else perplexed by Pomeroy rankings?



FuManShoes
12-13-2009, 04:02 PM
http://kenpom.com/rate.php

15. St. Mary's - (7-1)
40. Washington (6-2)
65. Gonzaga (6-2)
83. Oregon (4-4)
84. Portland (6-3)

That's something of a random sampling of teams of interest but what gets me is that GU's strength of schedule is ranked 149th with only Duke, Illinois and Memphis (and St Mary's) left to help. This team will need to blast through the WCC and win a couple of the remaining OOC matchups because what looked like a daunting schedule coming into the season isn't holding up.

BobZag
12-13-2009, 04:22 PM
While the Zags have a couple "nice" wins against Wisconsin, Cincy and I guess you could say WSU, some others have hacked up hairballs like Colorado. And even Sparty and Wake have not been overly impressive. Zags shoulda (woulda coulda) beat Wake, but.......

A couple marquee Ws like Duke, Illini, Ok, Memphis would sure help GU's cause.

CDC84
12-13-2009, 06:16 PM
It's too early in the season for the Pomeroy ratings to mean anything. It's like the RPI. Many analysts don't even think you should look at the RPI until early February. Conference play really needs to kick in for the ratings to mean much.

maynard g krebs
12-13-2009, 08:45 PM
If you flip a coin 10 times and it comes up heads 9/10, the odds of getting tails on the next flip are still 50%, but systems like Pomeroy's will at that point predict 90% chance of heads on the next flip.

FieldHouseFishHouse
12-13-2009, 09:29 PM
If you flip a coin 10 times and it comes up heads 9/10, the odds of getting tails on the next flip are still 50%, but systems like Pomeroy's will at that point predict 90% chance of heads on the next flip.

The performance of basketball teams is not quite as random as flipping a coin.
The reason SMC is ranked high is because they have absolutely trounced some teams. Their average margin of victory is very high. 4 of their 7 wins are by 20+, and their only loss was by 2. Like it or not, kenpom works with parameters that are more detailed than just wins and losses.
GU is ranked lower because they have a worse record, and the games they have won have been close (Cinci, Colorado, WSU even Davidson was too close for comfort).
Some people have pholosophical objections to certain rating systems. In my mind, a 1 point win is worlds away from a 1 point loss. I'm not sure Kenpom takes winning into account as much as I would like . However, I'm sure everyone agrees, the sample size is too small right now and things will sort themselves by late January.

maynard g krebs
12-13-2009, 10:28 PM
The performance of basketball teams is not quite as random as flipping a coin.


I didn't mean to suggest it is. The point was strictly to illustrate the statistical or predictive irrelevance of a small sample.

FieldHouseFishHouse
12-14-2009, 01:46 AM
I didn't mean to suggest it is. The point was strictly to illustrate the statistical or predictive irrelevance of a small sample.

I see. However, no matter what the size of the sample predicting the fate of a flipped coin is always irrelevant.
On the other hand, prediction of a team's performance can be somewhat relevant even with a small sample. That's why Ken Pomeroy has a popular site.

titopoet
12-14-2009, 05:14 AM
It is too early to look at Kenpom, even the Ken Pomeroy says its too early, http://rivals.yahoo.com/ncaa/basketball/blog/the_dagger/post/What-s-with-North-Carolina-s-weak-Pomeroy-Rating?urn=ncaab,207246 Cal is top ten. (they one away game at Pacific listed as a win. While GU has only one away game listed, MSU as a loss.) It has currently Saint Mary winning out and finishing 25-2, while GU being 21-9. Is there a 90% chance of that happening?

thebigsmoove
12-14-2009, 07:55 AM
The only reason the SOS is so skewed is due to the teams Gonzaga has played not playing anyone else yet, by the time conference play rolls around the schedule we have played out of conference will be working overtime for us. It like Gonzaga has saved up a large sum of money and during conference play they will be living off the interest. Think about all the top games that MSU, Duke, WSU, Wake, Illinois, and Oklahoma will have played. It works out in the end, our Non Conference SOS will be top 10 in the nation when all is said and done...

bullzag23
12-14-2009, 08:13 AM
Think about all the top games that MSU, Duke, WSU, Wake, Illinois, and Oklahoma will have played. It works out in the end, our Non Conference SOS will be top 10 in the nation when all is said and done...

I agree with you on every team there except for WSU. Let's be honest here, how many 'top' games can they get in the Pac 10 this year? So far the conference is looking terrible, with the so called 'best team' (UW) losing the only 2 games they play away from home in the nonconference. Granted they did play a solid Georgetown team, but the Huskies were absolutely exposed by the Hoya big men on saturday.

Drew
12-14-2009, 08:38 AM
The Zags dropped from 25 to 41 after beating Davidson in the Sagarin . . . this makes me wonder how much "strength of schedule" is taken into account (in mathmatical rankings), given that Davidson is ranked 195th.

http://www.usatoday.com/sports/sagarin/bkt0910.htm

CDC84
12-14-2009, 08:51 AM
Regarding the Pac 10......I think people on a national basis are forgetting that Cal has been dealing with injury problems all season long, and this has really effected their team performance so far. They are now finally starting to get healthy. It's a 3 man team, and if anyone of those big 3 go down for any length of time, they're going to struggle. They finally have Theo Robertson back. His first game back resulted in a 25 point blow out win, on the road, against what was a 7-1 Pacific Tiger team. Obviously, I don't expect Cal to beat Kansas on December 22, but the Bears are a better team than people think. Robertson averaged 13.1 PPG, 49% FG and 49% 3PT last season. You can lose that kind of production and expect to beat top 25 teams. I wouldn't shock me to see Cal win the Pac 10 with relative ease, but those big 3 - Randle, Christopher and Robertson - must stay healthy.

primal23
12-14-2009, 10:56 AM
I don't get it, SMC is supposed to win out, when they have beaten who? And so far have lost at home to Vandy? Vandy hasnt panned out as well as first thought they would. WOW all I can say is WOW, that and WTF

UberZagFan
12-14-2009, 11:12 AM
Too early too early too early.

whatazag
12-14-2009, 03:54 PM
Don't get too offended. Kenpom is a computer ranking. Doesn't mean a lot at this point, but it really has more to do with efficiency statistics, and the Zags honestly haven't been amazingly efficient. Sure, we've pulled out some tough wins, but that isn't necessarily going to translate in kenpom ratings.

FieldHouseFishHouse
12-14-2009, 04:42 PM
I don't get it, SMC is supposed to win out, when they have beaten who? And so far have lost at home to Vandy? Vandy hasnt panned out as well as first thought they would. WOW all I can say is WOW, that and WTF

The numbers say that because they play who? the rest of the way (besides GU). To be fair, they are favored to win each of their remaining games, but they are projected to have 1 more loss by the end of the regular season.

MickMick
12-14-2009, 08:25 PM
KenPom doesn't watch a game and think things like:

"That player has high baskeball IQ"

"This player is all out hustle"

"That player has great court vision"

"This team is improving"


It is pure number crunching with insufficient data. Nothing more.

People like Bilas and Lavin (fromer player and coach) look at Gonzaga and remark that the Zags are "tougher" than years past. Or the Zags have a top notch front line. Or the Zags have a great player in Matt Bouldin.

This is called "passing the eye test" and the Zags are doing just that.

Ultimately, it comes down to March. Unlike football where public perception is critical, the Zags will get their chance on the floor....as it should be.

FuManShoes
12-14-2009, 08:44 PM
Agreed the sample size is small, but Pomeroy has done well in projecting March success based on offensive and defensive efficiency and the Zags are not standouts in either at this point in the season. Allowing 91 and almost getting outrebounded by Davidson won't get it done and the Zags can't look at Bouldin's absence as an excuse for lax D. Thankfully there is time to shore things up. So I guess I wasn't so much perplexed by the Pomeroy rankings as surprised that despite the wins, the Zags aren't the bee's knees statistically (BTW: Duke is 3rd in O, 17th in D efficiency)

23dpg
12-14-2009, 09:43 PM
A little off topic but I just read that the Pac10 is 0 for 18 against the top 50 teams in the country! I know we poke at UW, but I'm a Pac10 guy overall. I can't believe how quickly the West Coast bb deteriorated.

alaskazagnut
12-14-2009, 10:29 PM
is that the best way to determine it is to average the AP, the ESPN, the RPI, the Pomeroy. At least for the top 25. Come February I can fill out my bracket almost as good as Lunardi. I have always been surprised that Lunardi can get the brackets almost 100% right, but he can't get a solid sweet sixteen. At least not enough to break Vegas though. Lunardi amazes me sometimes but it should be easy enough if that is the life you live: Bracketology.

Lunardi's brain sounds like Homer Simpson for donuts: Bracketology, mmm...... bracketology.