alright then
https://mwwire.com/?usat_social_embed=2 ... om-ratings

Interesting that we'd drop that much. We were 38 to finish the season last year.

I heard an interview with him at the end of last season. He was asked about Utah State and he said surprise teams that have a lot of returners almost never take the next step the year after their surprise season according to his statistics. It made me wonder if he knew how much turn over we really had. I think that is where this team will improve is the depth.

Ranked in top 20 in both polls, so go figure.


Sent from my iPhone using Tapatalk

I always wonder kenpom establishes a baseline when nobody has played a game yet.

nvspuds wrote:I always wonder kenpom establishes a baseline when nobody has played a game yet.

It's pretty much based completely on last season but that doesn't even fully explain the rankings right now.

Sent from my Pixel 3 XL using Tapatalk

If it was based upon last season Nevada would be higher..I think he is just guessing.

nvspuds wrote: ↑

October 28th, 2019, 7:08 am

If it was based upon last season Nevada would be higher..I think he is just guessing.

It's based on previous years, but does factor in departures & newcomers. Say what you want about Kenpom, but he doesn't guess. The formula adjusts quickly, though. USU jumped from ~150 to ~50 in just a couple games last year

Advanced analytics in general were very high on Quinn Taylor. The computers see his departure as a bigger deal than most fans.

I think kenpom does a great job. But all analytics have to have a baseline. Before games have been played he has decided what would be a good win or a bad loss. Since this is pretty arbitrary until actual data enters the algorithm he has to make guesses to start the ball rolling. Educated guesses but guesses nonetheless. Nevada has only 2 guys returning from what was a very good team. They also have a new coach and a new system. Based upon that alone they should be ranked far lower or not even ranked at all because no one anywhere knows how this season will go for the Pack.

You have a team that returns almost your entire rotation from what was an excellent team a year ago. Based upon that you should be ranked far higher.

This all works out and about a month into the season the data gives an accurate picture. It is just the baseline that is nebulous.

I'm sure the baseline includes the AP and coaches ranking. Maybe even an AD budget number.

Kenpom's preseason rankings are indeed nebulous, but they are not arbitrary. Nor do they include any polls.

It's all math that's way above my head, but it's all math based on prior seasons. I believe he uses recruiting rankings, similarity scores and that sort of thing to evaluate the newcomers. There's nothing subjective about his rankings.

One might argue that more subjectivity would help, since his computers are still taking the tail end of the Duryea era into account. Humans understand that the change from Duryea to Smith explains a large portion of the improvement, but his computers see that improvement as luck and expect regression.

Regression to the mean or just regression to being bad?..It is two different things..

If it is regression to the mean then you would probably be ranked higher since the mean under Smith is 28 wins, conference title, tourney title and a top 25 ranking including a berth in the NCAA's.

If he suspects that USU won't be as good this season the that is just regression..

I think there could be non scientific or math related reasons he has dropped you a bit. He doesn't think much of the MW.

Maybe he thinks with your tougher ooc you will lose more games.

Maybe he is factoring in Quetta missing some time.

However, those are all essentially guesses before any games are played.

The base may have true basis in science and math but it is indeed nebulous at best until real data comes in.

Morale of the story, all of the preseason rankings are just hype until you lace up the shoes and compete on the court.

ProvoAggie wrote: ↑

October 27th, 2019, 10:11 pm

nvspuds wrote:I always wonder kenpom establishes a baseline when nobody has played a game yet.

It's pretty much based completely on last season but that doesn't even fully explain the rankings right now.

Sent from my Pixel 3 XL using Tapatalk

I'm sure he uses more than 1 previous season as a baseline. Especially given his explanation that teams who unexpectedly do well in a given year normally fall back to the average the following year. That's why he rates us so low. He's still factoring in the Duryea years and treating last year as a fluke.

He will be wrong.

I don't care what he thinks right now. I care how we play right now. These ranking systems are always interesting to evaluate after it is all said and done.