Kenpom.com

Collegiate Sports
1

Someone mentioned in another thread that they held the Kenpom ranking and prediction system above the RPI. How about the rest of yall?

If you are not familiar with Kenpom.com, it is a wonderfuly bare-bones site where Ken Pomeroy has his own Men’s Bball ranking system and also keeps track of team stats, player stats, and rankings throughout the NCAA.

Predicting: Since the GT game, it has predicted our record correctly (only 3 games, but like the RPI, it takes time for Kenpom’s #s to even matter. I also don’t have earlier predictions for prior games). Directly after the GT game it predicted Richmond as a 72-57L, actual score 71-59L. Pretty close. The post-GT prediction for SBU was a 69-65L, but it isn’t going to predict us going to 3OT either. The actual difference (4pts) was the same as the predicted. The post-GT prediction for Fordham was 72-63W, actual was 71-61W. So, from this limited data it would appear that Kenpom has some great predicting capabilities.

Ranking: But I think it falls flat in its ranking. In short, I think it places entirely too much value on margin of victory. Time and time again, I’ll see teams with a no/few quality wins ranked highly because they beat much lesser competition by a great margin. Looking quickly, I pick out Washington and Baylor.

Washington - Rank: #8. I’d say they have 2 decent wins with USC, and UCLA, and then I guess mentionable UVA, Portland, and LBSU. Their losses haven’t been against bad teams either. But they really haven’t beaten anyone all that great. Does this really seem like a #8 team to you? I think what makes this #8 in the Kenpom rankings is 54pts, 26, 43, 27, 29, 22, 28, 30, 18, and 31pt wins against a bunch of mediocre teams. Not saying they aren’t good, just not close to #8 if you ask me.

Baylor - Rank: #41. They have 0 decent wins. I suppose AZ St. is mentionable. No bad losses. Wins by 35pts, 45, 44, 33, and several by 10-20 against a bunch of weak teams. They aren’t a bad team, just not #41 I don’t think.

IMO, if he dropped the value of MOV on ranking, it would be more worthwhile.

2

Aww come on

3

tl;dr

4

GD! I knew it. I even put in bold words for you people so you could know what to bypass/skim over.

tl;dr: Kenpom would be a more accurate ranking if it didn’t value margin of victory so much. Discuss.

5

I’m probably the guy who said they prefer KenPom to RPI so I will engage you in conversation. With KP the rankings are based on so much more than W/L + Margin of victory. The stats used to measure offensive/defensive efficiency are very thorough and have great influence on the overall rating.

To your point on Washington, they are considered to have the 5th most efficient offense in the country at this point, and 22nd most efficient defense. These factors combined with a few others lead to the 8th overall ranking. I agree that they don’t have an amazing win, but they’ve gone head to head with a few very good teams on the road/neutral and lost by slim margins.

Bottom line; how a team plays says more for their ranking than who they’ve played. You’ll notice teams will not take as much a hit for just playing a lower seeded in KP as they will in RPI.

By the end of the season look at Ken Pom’s top 10 and I bet you agree by and large with the order. Check out the last few seasons on record; they are spot on. It’s a brutally honest way to assess our team.

One cool feature of Ken Pom, click on a score of a past game this season and you will see a line graph chart the % chance a team has to win at any point during the game. Too bad this doesn’t stream live.

6

GD! I knew it. I even put in bold words for you people so you could know what to bypass/skim over.

tl;dr: Kenpom would be a more accurate ranking if it didn’t value margin of victory so much. Discuss.[/quote]

Margain of victory would be more interesting if it was weighted i.e. a blowout against a good team has more value than a blowout against an awful team. It’s still better than RPI because it doesn’t matter if you beat a team by 1 or 100 for RPI purposes.

Better?

7

[quote=“donkeyjaws, post:5, topic:24781”]To your point on Washington, they are considered to have the 5th most efficient offense in the country at this point, and 22nd most efficient defense. These factors combined with a few others lead to the 8th overall ranking. I agree that they don’t have an amazing win, but they’ve gone head to head with a few very good teams on the road/neutral and lost by slim margins.

Bottom line; how a team plays says more for their ranking than who they’ve played. You’ll notice teams will not take as much a hit for just playing a lower seeded in KP as they will in RPI.[/quote]
They’ve actually gone up since my previous assessment. They are now ranked #6.

I think their EO is so high because of MOVs I listed earlier. My problem, is I think once you have a team beat, I think it isn’t much of a testament of skill to run up the score on them. Granted, running up the score by such absurd margins is an indication of significant scoring ability, even against a lesser opponent; but for that to be a large part of what makes you 6th in the country?

I don’t want to over value who you’ve played like the RPI may, but they have not beaten anyone. It is one thing to run with great competition, but it is another thing to beat them, which they have not. Good teams win games. In my view, what they have demonstrated is worth a 20-30, but certainly should not be judged as one of the premier teams in the nation.

Now look at #7 Syracuse. 22pts, 20, 57, 42, and 28pt wins are their biggest and only 20+ point wins. Somehow their SOS (overall) is lower than Wash, but not by much (19 spots). The thing is, Syracuse HAS beaten a top 30* team (2), and Wash lost to both of theirs. Sure Syracuse hasn’t played a top 10 team and that is one of Wash’s losses, but Syracuse also hasn’t lost any games at all. Wash has a loss to a 72, too. And then compare top 100 and top 50 wins, Syracuse has that hands down.

I just don’t think a team like Washington with 4 losses and lesser best wins is BETTER or even that close to an undefeated team like Syracuse.

*I realize it’s bad form to use something you disagree with to bolster your argument against it, but bear with me lol. Trying to be concise.

By the end of the season look at Ken Pom's top 10 and I bet you agree by and large with the order. Check out the last few seasons on record; they are spot on. It's a brutally honest way to assess our team.
I do still really like Kenpom despite my stance here. Have used it for.. 6 seasons (wow) since when he also used RPI. It is very accurate and in more ways accurate than RPI, I just think there should be a MOV coefficient (if there isn't already) or something where you get less reward after 20pts, and 30 pts, and so on.

I’m not sure how to look at old RPIs to judge which result I agree with most, but I’m sure Kenpom is pretty close. Another problem tho, I think, is these teams’ records reflect the NCAA tourney which skews results a bit.

8

[quote=“NLP, post:6, topic:24781”]Margain of victory would be more interesting if it was weighted i.e. a blowout against a good team has more value than a blowout against an awful team. It’s still better than RPI because it doesn’t matter if you beat a team by 1 or 100 for RPI purposes.

Better?[/quote]I agree with your first point.

I disagree with the second. I don’t think a team with 4 losses and lesser wins should be ranked higher than a team with 0 losses and much better wins simply because the first team ran up the score like crazy against several teams. See: Washington v. Syracuse in my last post.

9

[quote=“donkeyjaws, post:5, topic:24781”]I’m probably the guy who said they prefer KenPom to RPI so I will engage you in conversation. With KP the rankings are based on so much more than W/L + Margin of victory. The stats used to measure offensive/defensive efficiency are very thorough and have great influence on the overall rating.

To your point on Washington, they are considered to have the 5th most efficient offense in the country at this point, and 22nd most efficient defense. These factors combined with a few others lead to the 8th overall ranking. I agree that they don’t have an amazing win, but they’ve gone head to head with a few very good teams on the road/neutral and lost by slim margins.

Bottom line; how a team plays says more for their ranking than who they’ve played. You’ll notice teams will not take as much a hit for just playing a lower seeded in KP as they will in RPI.

By the end of the season look at Ken Pom’s top 10 and I bet you agree by and large with the order. Check out the last few seasons on record; they are spot on. It’s a brutally honest way to assess our team.

One cool feature of Ken Pom, click on a score of a past game this season and you will see a line graph chart the % chance a team has to win at any point during the game. Too bad this doesn’t stream live.[/quote]

I’m no statistician, but I prefer Ken Pomeroy too - although I look at both. Kenpom has us ranked higher so far.

10

I agree that Syracuse would seemingly deserve a higher public ranking than Washington based on going undefeated, but what Ken Pom tells us is that despite the difference in records, the 2 teams are still essentially indistinguishable based on their current level of play. Look at Washington’s breakdown. Overall, they kill who they’re supposed to kill. Syracuse has let some teams hang around that had no business doing so, and have only played 2 true road games all year…Both teams at the end of the day are essentially weighted even. Washington loses to Texas A&M on the road by 1, Syracuse beats William and Mary by 3 at home…which makes you cringe more?

Here’s some insight into the basis for the ratings pulled straight from the site’s help page.

"The first thing you should know about this system is that it is designed to be purely predictive. If you’re looking for a system that rates teams on how “good” their season has been, you’ve come to the wrong place. There are enough systems out there that rank teams based on what is “good” by just about any definition you can think of. So I’d encourage you to google college basketball ratings or even try the opinion polls for something that is more your style.

The purpose of this system is 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."

11

[quote=“donkeyjaws, post:10, topic:24781”]I agree that Syracuse would seemingly deserve a higher public ranking than Washington based on going undefeated, but what Ken Pom tells us is that despite the difference in records, the 2 teams are still essentially indistinguishable based on their current level of play. Look at Washington’s breakdown. Overall, they kill who they’re supposed to kill. Syracuse has let some teams hang around that had no business doing so, and have only played 2 true road games all year…Both teams at the end of the day are essentially weighted even. Washington loses to Texas A&M on the road by 1, Syracuse beats William and Mary by 3 at home…which makes you cringe more?

Here’s some insight into the basis for the ratings pulled straight from the site’s help page.

"The first thing you should know about this system is that it is designed to be purely predictive. If you’re looking for a system that rates teams on how “good” their season has been, you’ve come to the wrong place. There are enough systems out there that rank teams based on what is “good” by just about any definition you can think of. So I’d encourage you to google college basketball ratings or even try the opinion polls for something that is more your style.

The purpose of this system is 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."[/quote]
True. I actually think I read that before too. So then can you really prefer one over the other when they are doing 2 different things?

Haha it just seems like they are doing the same thing tho, kinda. One uses past results to predict a winner. One uses past results to say who has done/is better. Seems like the optimal answer should be the same.

  • And I would still think the Kenpom predictions would show Syracuse as a significantly stronger team than Wash lol.