Jake Flancer

August 29th!

I'm still cautious about comparing this season to past years like @kenpomeroy noted yesterday, but here are the ratings in historical context (since 2013). Count by season: 2013: 1, 2014: 4, 2015: 1, 2016: 4, 2017: 2, 2018: 5, 2019: 2, 2020: 6

Final end of season sRAPM update now up at jflancer.shinyapps.io/rapm/. Some movement at the top off of the final game as Butler climbs to number one... but still practically the same as the others given the variance on these ratings.

And updated player ratings as well...

Final model update of the tournament

My top 25 team ratings from 2013-14 to present (pre elite 8)

Updated probabilities using round 2 data now live at jflancer.shinyapps.io/march_…. Model didn't have the best day yesterday + dropped to 24th in Kaggle but still very happy with its overall performance.

Updated probabilities on jflancer.shinyapps.io/march_… after yesterday's games. Still very cautiously optimistic and there are a lot of games left, but now up to 17th place in Kaggle!

Updated sRAPM-based predictions are live at jflancer.shinyapps.io/march_…. These include the use of round 1 data (you can also toggle to pre-tournament). I have some concerns but the model has done well so far. #41 on kaggle- I forgot to submit spreads but will compare them eventually

Update after the first day of games- note I'm just updating probabilities based on pre-tournament strength and not actually re-calculating anything (although I might for round 2)

Bracket predictions are now live! Check out my prediction table, team ratings, and matchup predictions at jflancer.shinyapps.io/march_…

Here's vanilla rapm without any prior... Getting the prior alone is more difficult to do based on how my code is set up

It's fairly similar but values do look like they shrink quicker

Final pre-tournament RAPM update is live at jflancer.shinyapps.io/rapm/. I think Gonzaga may have officially broken my model...