All stats are important, of course, but turnovers and FT shooting are going to be huge as we get to conference play and beyond. Let's hope the Tigers shore those up quickly...

(11-18-2021 04:49 PM)Fluke Wrote:  

(11-17-2021 03:38 PM)Tiger87 Wrote:  ESPN has us tied for 348th in TO's per game

While I’m sure it’s still not in our favor and we need to improve it.. people need to stop using per game numbers for most stats. We were listed at the #3 fastest pace of play the other day. More relevant are the per offensive possession numbers to get a truer idea of where we stand. A slow team that has better total turnover numbers does not mean they actually take care of the ball better and vice versa.

I get your thought process behind weighing on a ratio of turnovers and number of possessions. However, number of turnovers as a stand along stat is still massively important, especially live ball turnovers. Live ball turnovers almost always provide a favorable scoring advantage for the offense and a significant disadvantage for the team that just turned the ball over.

Even if we have the same ratio of turnovers to possessions as a slower paced team we still put ourselves in a disadvantaged situation several more times.

(11-19-2021 08:21 AM)Hoots Wrote:  All stats are important, of course, but turnovers and FT shooting are going to be huge as we get to conference play and beyond. Let's hope the Tigers shore those up quickly...

I cannot wrap my head around the terrible free throw shooting...

Timerberlake's form looks stiff and awful. He bends, then straightens/stiffens up and then tries to shoot. Instead of it being one fluid motion to shoot it is three distinct choppy movements.

(11-18-2021 04:36 PM)salukiblue Wrote:  

(11-17-2021 03:53 PM)cmt Wrote:  

(11-17-2021 03:44 PM)Tiger87 Wrote:  Looking over some of ESPN's stats and rankings.
To show you how screwed up their BPI is...
3 of their top 5 ranked BPI teams have a loss already...
including their top 2 teams of Villanova and Illinois.

Yeah, Kenpom puts Michigan at 3 for losing and dropped us to 15 for winning. Same BS every single year. How does being in the P5 in football guarantee you a better starting spot in an objective, math-based, preseason ranking? Only 6 P5 teams are ranked below 100, lowest Washington at 134.

Actually KP dropped Michigan 1 spot for losing to a top 30 KP team and Memphis moved up 2 spots after beating STl.

Stl actually moved UP 5 after losing to Memphis.

Memphis is 13 currently.

(11-18-2021 04:33 PM)salukiblue Wrote:  

(11-18-2021 04:17 PM)Tiger87 Wrote:  I agree that it's way too early to get excited, and there is definitely some conference bias built in.
My main point is that the ESPN BPI is the worst of all of these.
I've never been impressed with it, and it often has strange results.
Having your #1 and #2 ranked teams sitting there with 2-1 records this early in the year is an absolute joke.

How is it a joke? (Ok illinois is a joke) but from a pure comparison standard, Nova losing in OT (factually a tie at regulation) on the road to the #7 team would mean they are better on a neutral court and much better at home than the #7 team.

In all reality, poll wise, if Memphis were to go the the Kennel and lose on a last second shot to the Zags, the pollsters should actually move Memphis UP in the rankings, not down just because Memphis took an L in a particular column.

According to that line of RPI thinking, theoretically, a team should be ranked higher if they played the #1 ranked team away close 30 times and lost rather than win close games against the #40 ranked team 30 times at home.

(11-19-2021 09:25 AM)ballhog Wrote:  

(11-18-2021 04:33 PM)salukiblue Wrote:  

(11-18-2021 04:17 PM)Tiger87 Wrote:  I agree that it's way too early to get excited, and there is definitely some conference bias built in.
My main point is that the ESPN BPI is the worst of all of these.
I've never been impressed with it, and it often has strange results.
Having your #1 and #2 ranked teams sitting there with 2-1 records this early in the year is an absolute joke.

How is it a joke? (Ok illinois is a joke) but from a pure comparison standard, Nova losing in OT (factually a tie at regulation) on the road to the #7 team would mean they are better on a neutral court and much better at home than the #7 team.

In all reality, poll wise, if Memphis were to go the the Kennel and lose on a last second shot to the Zags, the pollsters should actually move Memphis UP in the rankings, not down just because Memphis took an L in a particular column.

According to that line of RPI thinking, theoretically, a team should be ranked higher if they played the #1 ranked team away close 30 times and lost rather than win close games against the #40 ranked team 30 times at home.

Pretty much. Kenpom and other advanced stats sites will set their own “line” based on what their algorithms say. If you are supposed to lose by 10 but instead lose by less than that margin then you have outperformed their algorithm. So you would move up.

(11-19-2021 09:25 AM)ballhog Wrote:  

(11-18-2021 04:33 PM)salukiblue Wrote:  

(11-18-2021 04:17 PM)Tiger87 Wrote:  I agree that it's way too early to get excited, and there is definitely some conference bias built in.
My main point is that the ESPN BPI is the worst of all of these.
I've never been impressed with it, and it often has strange results.
Having your #1 and #2 ranked teams sitting there with 2-1 records this early in the year is an absolute joke.

How is it a joke? (Ok illinois is a joke) but from a pure comparison standard, Nova losing in OT (factually a tie at regulation) on the road to the #7 team would mean they are better on a neutral court and much better at home than the #7 team.

In all reality, poll wise, if Memphis were to go the the Kennel and lose on a last second shot to the Zags, the pollsters should actually move Memphis UP in the rankings, not down just because Memphis took an L in a particular column.

According to that line of RPI thinking, theoretically, a team should be ranked higher if they played the #1 ranked team away close 30 times and lost rather than win close games against the #40 ranked team 30 times at home.

RPI wouldn't do that. The main measure of RPI is a team's winning %, then a teams opponent's winning percentage, then opponents-opponents winning %.


Kenpom and sagarin both are efficiency based models.

I do agree with 87, broadly speaking, that early season rankings aren't very good, BUT like in the case with Villanova, just because a team loses doesn't mean it wasn't the better team that night (advanced stats wise).

(11-18-2021 04:49 PM)Fluke Wrote:  

(11-17-2021 03:38 PM)Tiger87 Wrote:  ESPN has us tied for 348th in TO's per game

While I’m sure it’s still not in our favor and we need to improve it.. people need to stop using per game numbers for most stats. We were listed at the #3 fastest pace of play the other day. More relevant are the per offensive possession numbers to get a truer idea of where we stand. A slow team that has better total turnover numbers does not mean they actually take care of the ball better and vice versa.

Very good point.

But another way of looking at it is: unforced/careless turnovers.

All those computer rankings are flawed because you have to have a starting point. They had us ranked 16 to start. How do they know how any team is going to play ? Still a human element involved which brings in bias.

(11-19-2021 10:13 AM)Eagleonpar Wrote:  All those computer rankings are flawed because you have to have a starting point. They had us ranked 16 to start. How do they know how any team is going to play ? Still a human element involved which brings in bias.

Agree.

(11-19-2021 10:29 AM)salukiblue Wrote:  

(11-19-2021 10:13 AM)Eagleonpar Wrote:  All those computer rankings are flawed because you have to have a starting point. They had us ranked 16 to start. How do they know how any team is going to play ? Still a human element involved which brings in bias.

Agree.

Yes they are flawed to start...KenPom and I think BartTorvik both remove any preseason Bias by game 13.

At game 13 the starting point is the current season...After that point there is no human element outside of the person who wrote the algorithm.