{fitzRoy} 1.5.0 has been released on CRAN.
Many thanks to @random_effects for a mountain of work optimising code and squashing bugs over the last few weeks! Read about the changes here https://t.co/SltC7eINUZ
I've collated the publicly Brownlow model predictions + some private ones. Will do a write-up tomorrow with the results and discuss the big surprises. If you have a model, hit me up in DMs/comments! Good luck!
https://t.co/esvbQelXYG
@MatterOfStats@ChewTheStat@WheeloRatings
Feelin Fuzzy? @OSPpatrick and I work through identifying players using fuzzy joins to standardize names of sports ball players against that in your database.
We use two different techniques, using base #rstats agrep and #tidyverse -esq {fuzzyjoin}!
https://t.co/MPjB4xcUyy
You can rerun my AFL Win Probability Added (WPA) analysis from my stream and thread here. If you're statistically inclined, try to build a model that's better than mine! #RStats https://t.co/ozkfj0pQck
AFL stats fans wanting stats reminds me of the classic image below
Also a reminder of how poor the league treats it's fans and how reliant a billion dollar industry is on volunteers
Ok folks, I'm going to try out Twitch. I'll be streaming some coding because I am bored. Please let me know if the stream stays on rstudio and if my microphone isn't terrible.
CD: We have turned this binary variable {Goal | Not-Goal} into a continuous value
Statisticians: Bravo! Less noise in this metric
CD: Now we will dichotomise the totals into this bigger binary variable {Win | Loss}
S: Wait … what?! NO!