2015-16 Media love thread

[proshox]

JH4P - what do you think of BPI!

[1979Shocker]

nm

[DUShock]

Wow! Purcell gets away with another over the back and the recipient is called for the foul (Wing)

[DUShock]

ISUr is schooling th Aces in the second half.

[Jamar Howard 4 President]

JH4P - what do you think of BPI!

Honestly I barely ever look at it. Don't even know how it's calculated. Don't particularly care.

[ShockerFever]

Have you seen ESPN's new "strength of record"?

They're really pulling out every trick in the book.

[Cdizzle]

Have you seen ESPN's new "strength of record"?

They're really pulling out every trick in the book.

If they invent all the metrics, they can pick all the tournament teams.

[ShoxRoxMySox]

Have you seen ESPN's new "strength of record"?

They're really pulling out every trick in the book.

Surprised ESPN can afford to create anything new, massive budget cuts and all.

[ShockBand]

Why don't they just get it over with and call any new metric SONP5 - **** On Non Power 5!

[rayc]

Unlike RPI, Ken's formula is much more complex. RPI is purely based on wins and losses. KenPom is purely based on points scored and allowed per possessions played. Here is my basic understanding.

Ken calculates raw offensive and defensive efficiencies per 100 possessions. This simply means calculating how many points a team scores every 100 possessions. Nothing complicated here other than compiling the data. The average team in division 1 scores about 1 pt/possession, give or take.

Ken then takes the raw numbers and adjusts them (I'm not sure of his exact formula) to reflect the quality of opponent played. Scoring more than a point per possession against Bradley is far from the same from doing it against a good defense like WSU's. Ken's adjustments take the raw, statistical efficiencies and adjusts for the SOS that has been played so that efficiencies can be compared from team to team. Thus, when WSU's adjusted offensive efficiency is 110.4 it means that WSU would be expected to score 110.4 points in 100 possessions against the average D1 team. They would expect to score more or less than this if playing an above or below average defense.

The great thing about this approach is that it is based on efficiency, not total points scored. By taking tempo out of the equation, teams that play slow, low scoring games aren't penalized, and teams that play fast, high scoring games aren't overly rewarded. It is much harder to win by 30 when the tempo is slow than when the pace is really fast. Removing tempo allows Ken to focus on scoring, but make it fair for all teams.

Overall rank is calculated through one more level of calculation. Ken uses what he refers to as the “pythagorean winning percentage". His formula takes offensive and defensive efficiencies and calculates an expected winning percentage against the average D1 team. You may have seen elsewhere where I have mentioned WSU's KenPom rating has been hovering in the 0.90 range. It is 0.9059 currently. Simply put, this means that WSU would be expected to win 90% of their games against average (approx #175 out of 351 in D1) teams on a neutral floor.

I think using KenPom and RPI side by side is extremely helpful because they are calculated so completely differently. If the 2 ranks are similar, it gives me confidence that the ranks are accurate for a given team. If they are far apart, it helps to give a range of valid rankings. If a team is #20 in KP and #50 in RPI, I may not feel confident giving them an exact rank of my own, but very rarely do I feel that they should be #19 or #51. Usually, I'm confident they belong somewhere within the range.

(Bold added to the sentence by me.)

Thanks for the reply. This is as good an explanation of KenPom as I've run across.

One of the problems I've always had with KenPom is reflected in the bolded line in your explanation. At one point in the past he mentioned in an interview or article that he did give a bonus to teams in the power 5 conferences in his ratings. But of course did not provide the how, where and how much. I hate proprietary programs for use in something like team selection for the NCAA tournament. Black box programs are of course necessary from his financial point of view. But I'm not comfortable with them being a selection criteria, per se.

The ESPN stats are of course a total joke.

[Jamar Howard 4 President]

@rayc:, I've never heard about Ken giving power 5 conferences a special boost in his ratings formula. When I say he adjusts for SOS, I merely mean that he takes a team's points/poss rate and multiplies it by some factor to account for whether their opponent was an above or below average offense or defense. I've never heard anything to indicate that conference affiliation affects his rankings in the later part of the year. I say "later part of the year" because it is possible that he does something fishy in the first few weeks or months. I know he makes special adjustments to start the season when there is very little data available. In November, he factors in performance from the previous season. In March, he is 100% focused on the current year.

I'd be curious to find out more if you ever come across a link.

[ShockdaWorld]

What does he use to determine SOS early in the season? Is that number based on the rankings or some other metric? I've always wondered how people come up with "they have the toughest schedule in the country" when there haven't been games played. They say the rankings don't count, but how else are the figuring that SOS out?

[Jamar Howard 4 President]

Here's a good blog post Ken wrote prior to the 2014 season explaining how he develops his pre-season ratings. As the season goes on, he reduces the weighting of his pre-season ranks and eventually goes fully off of the current season's stats.

http://kenpom.com/blog/index.php/weblog/entry/pre_season_ratings_2014

[SB Shock]

WSU and Connor getting some play of their highlights at Halftime of the KU/OU game

[ShockerFever]

WSU and Connor getting some play of their highlights at Halftime of the KU/OU game

Are you and your dad back off the Frankamp train again? Bandies.