Dear Ken Pom


#21

I didn’t “leave them out.” I simply addressed your specious argument.

BTW - You think I don’t know that all the win-loss based retrodictive systems have us ranked a lot higher than the score-based predictive systems?

Sheeeeee.

I’ve forgotten more about this stuff than you’ll ever know.


(Jimmy Morris) #22

That’s apparent from what splitstud pointed out.


#23

Nah, he’s right that they start with bias then drop it, just didn’t remember it was late January.


#24

All the rankings are meaningless this early, so only view the one that makes you feel the best. All the rest are bullshit :smile:


(Jimmy Morris) #25

Most of us know the bias is eventually weeded out. So I’m going to stick it to him for calling me wrong. My biggest axe to grind with KenPom and Sagarin is the huge boosts these ratings unnecessarily give in the offseason to big name basketball programs. It’s not harmless because media use these preseason rankings. Their attention and hype uses these biased rankings. Then teams that had a great season last year but aren’t media darlings have to start the next season from scratch to gain the respect all over again.


(gpropes) #26

KPI currently has us at 7.

And if I’m reading Packtex’s link correctly, KPI would therefore currently have us as a 2 seed.


(Ben B) #27

I’ll take a two seed! Now we just have to go out and not lose any games the rest of the way!!! Tough to do.


#28

I wasn’t smart enough to find the KPI site. I just read it off of the team sheet. Thanks for the KPI link.


(College of Technology Class of 1981) #29

hey, Isn’t “pom” french for “potato” ? …just askin’


#30

pomme = apple, la pomme de terre = potato


(College of Technology Class of 1981) #31

Thanks for clearing that up! I guess I was working from hearing rather than reading.

I try not to work too hard on humor, heh


#32

no problem. i am self studying french to try to keep my brain working while i recover.


(goblue) #33

You do realize that they are computer programs that don’t know anything about the name on the front of the jerseys, right?

KenPom’s preseason ranking is mostly based on how good your team has been the last several years with most recent seasons weighted most heavily. Why? Because that makes it more accurate. I mean there are other inputs like how much returning production you have, allowing for some bump for elite recruits, etc, but those are the same thing everybody considers.

It’s also ironic to complain about KenPom considering KenPom was a big fan of Houston last season ranking them #17 in the country heading into the tournament.

Houston’s lower starting spot in the preseason this year had nothing to do with not being a media darling, it had to do with losing Rob Gray who used an insane 30% of possessions while on the court and Devin Davis who used 21% of possessions. Those were the Cougars 2 highest usage players.

KenPom is what it is. It’s quite easy to see how to skyrocket up the rankings. Cougars are a 2 point favorite at Temple tonight (by KenPom). Win the game by 20+ points and they will make a big jump. Lose by 10+ points and they will drop. Anything (win or loss) within about 5 points and they will barely move.


(Jimmy Morris) #34

As a Michigan fan, I don’t blame you for supporting this reasoning. There is no reason for you to do a deep dive into recruiting rankings if you are often near the top. What we find as supporters of a G5 program, is that most of the recruiting rankings are lazy. They often identify a recruit’s worth based solely on which schools are giving offers. We have had threads almost every year that keeps track of individual recruits ratings on several services and one thing has been certain, when a recruit that has several P5 offers, commits to a G5 program his rating goes down and when a recruit switches from a G5 commit to a P5 commit, their rating goes up.

So the question is, does anyone in this conversation really believe that KenPom and Sagarin had accurate information in regards to the newcomers Alley, Hinton, Gresham and Jarreau? Their job wasn’t to replace Gray and Devin. It was Armoni and White’s jobs to replace those two, Brady, Robinson and Davis’ job to pick up any slack and the new faces job to replace or improve on Armoni and White’s contributions off the bench.

It wasn’t just a one year thing. It’s been an every year thing with many prominent teams. Less teams in football so I use it as an example as it’s easier to notice the trend. This year in preseason rankings UCF starting with a lower rating and Auburn moving up in the ratings. Two years earlier when Houston moved down in the ratings before their first game and Florida State had moved up.

It just makes more sense to let the ratings be fluid. Leave teams where they ended one year to start the next. Unless the entire team and coaches have changed, they earned those ratings from the previous year and no bias should be injected especially if they are weeded out eventually anyway.


#35

Kenpom predicted we’d win that game by 1. Just sayin’.


(Jimmy Morris) #36

I think this post is worse than the one you came after me because Alley had one great night shooting. Do you think Kenpom calculated in the crazy calls? SMH


(goblue) #37

Here’s the thing, Michigan gets almost no bump for recruits in the KenPom preseason formula. It’s really mostly top 10 or 20 recruits that move the needle to any degree and Michigan has had about a grand total of 1 of those in the last decade.

My post about KenPom’s formula has nothing to do with the recruits Houston signed. It’s a math formula. It’s easy to understand. Since it spits out almost the exact same results as the Vegas pointspread, it’s clearly got value and meaning. I mean if it was always so wrong about a certain team or set of teams, it’d be easy to go to Vegas and make a killing betting on those games.

Does KenPom potentially undervalue a given recruit or two going to a certain school? Perhaps. But the value of those gets considered on the basis of the previous seasons ratings for that school. The end of season KenPom ratings have nothing to do with any preseason guesses. So if your team always finishes in a given range, it will probably start out there in preseason no matter how little credit the recruits are given. Houston’s last 4 season finishes in KenPom (going back) were 18, 52, 61, and 214. Like I said not surprising that they slid back a bit in the preseason formula considering the loss of the 2 highest usage players on the team. To move up the rankings, simply outperform the predicted point spreads consistently.

It’s a nice simple thing to use and understand and to know it has no bias for or against any school or conference. I mean hell it usually holds a school like St. Mary’s in higher esteem than polls and tournament selection committees.


#38

Of course one data point isn’t significant but Kenpom thought it would be a close game and it was. So like almost every game we’ve had recently where our MOV was generally in line with Kenpom’s projection, our Kenpom ranking will barely change. Even with crazy calls it’s still a close game though maybe one we win.


(Ben B) #39

If you paid attention in just one set of calls the refs made a 7 point swing and that doesn’t include other atrocities they committed. So no, it wouldn’t have been a close game if it was even reasonably called.


#40

Officiating is part of why there’s an advantage for home court that is factored in. Kenpom has us as 8 point favorites over Temple at FC right now. If you want to say that a coach getting a tech for taking his tie off when he does that every game is so far outside a normal call that’s true (as has been mentioned many times, Penders getting a tech for fainting due to health issues will probably never be surpassed). But bad calls always happen and generally happen more for the road team and sure some games are worse than others. After our game with LSU which was probably right in the Kenpom margin Tiger fans were griping about the officiating. Same with the St. Louis game which was closer than Kenpom probably would have projected. We are 12 point favorites per Kenpom at home Saturday. Let’s see how Kenpom in actuality gets that margin wrong even if that’s almost exactly what happens Saturday.