Projecting KU's record using Seniority Scale

One of the myriad of graphs and charts I track for KU basketball is what I call the Seniority Scale. It is very simple. I take the total minutes played by every player and weight average based on class where a freshman has a class of 1, a sophomore is a class of 2 a junior is a class of 3 and a senior is a class of 4. By this standard the youngest team under Coach Self was 2014 with a SS of 1.66 and the oldest was 2012 with a SS of 3.26. I estimated the 2022 team total minutes played by player and came up with a SS of 3.3 If this holds this will be the oldest team under Self. Here is my minutes estimated for 2022 (per game)

Seniors -- 117 (Remy Martin, Cam Martin, David McCormack, Ochai Agbaji, Lands-Coleman and Mitch Lightfoot)
Juniors -- 30 (Braun)
Sophomores -- 48 (Wilson, Harris and Yesufu)
Freshmen -- 5 (Adams, Clemence, Cuffe, Pettiford)

I know the freshman number is low and very well may be wrong. But we are so loaded with experience it seems that there will be scant minutes for freshmen.

You will notice on the chart that I have iso-talent lines. I have assessed the talent level of each team an drawn the lines. Obviously the more talented the better the winning percentage but also the more experienced the team the better the win%. In fact, it appears that each increase in seniority scale by 1 full class improves winning percentage by 6-7 percentage points. In other words an average talented KU team with half sophomores and half freshmen (a SS of 1.50) would win about 74% of their games or go roughly 30-10 whereas an average talent KU team with a SS of 2.6 would win about 81% of the games or go 32-8 and a team with average talent and a SS of 3.5 would win about 88% of their games go roughly 35-5. So purely being older with the same level of talent improves winning percentage by about 7 percentage points for each year of seniority.

I project the current 2022 team to be average KU talent. I would welcome anyone else's input on that. This is a solid team in regards to talent but not really any better than the average KU team under Self. The most average teams under Self have been 2006, 2009 2013. Remember the talent is also a weighted average based on minutes played.

With all that said, a 2022 team having a SS of 3.3 and average talent would result in a record of 35-5. Even if it turns out this team is below average talent they should win 32 or 33 out of 40 games. So if you buy all this then KU has good probability of being north of a 30 win team.

Very interesting...

Corpus, You've quite a bag of tricks there!

Thanks!

RCJH! Go KU!

thx corpus! my only nit in your analysis wld be that one cld argue that seniority has mostly been of benefit because of the time of exposure to coach self and learning to do things his way. this year half of our seniors will be in their first year with self and may not produce the kind of benefit we get from the others. will hope for otherwise of course but im actually expecting some early tension with these guys and coach as they adjust to each other

In the world of QA (quantitative analysis) the challenge is to isolate the dependent and interdependent relationships of all variables. So by way of full disclosure, rarely are any of these correlations perfect. So introducing a variable like seniority at KU or redshirting are no doubt valid and will have some impact. There are other variables that will impact the correlation as well. But the degree of impact and the skewness from not including those variables can be reasonably estimated. That is a big word salad to say that while this graph is not completely accurate it is nonetheless quite reasonably accurate enough to be meaningful and interpretively useful. Unfortunately the work necessary to refine the impact of these variables pretty much exceeds my allotted time available for this particular hobby. The benefit does not justify the cost from a time standpoint. But rest assured, whenever I have free time I tend to work to validate my data analyses so I can give you guys meaningfully useful analyses.