Ian Barnett
@Barnett_Speaks
Followers
45
Following
41
Statuses
19
California native, now Assistant Professor of Biostatistics at UPenn. SF sports fan. Boba enthusiast.
Joined January 2013
@billjamesonline @tangotiger If anything I'd presume 35-37 would be more predictive than 1-3 of overall reg. season performance. Is adv. scouting more outdated at season start than it is mid season and therefore performance is less generalizable? I'm speculating but data could answer this more definitively
0
0
1
@tangotiger Well can't argue with that huge increase in R squared relative to expected wOBA. It clearly works! Cool stuff.
0
0
1
@NinersNation Can we just wait until a SB win to ask this question? We've been burned before with seemingly invincible niners teams within the past 15 years...
0
0
1
@tangotiger As a follow up, I put together some more details attached. It boils down to correcting for unequal season length in the estimates of winning percentage. The degree of certainty in a teams winning percentage is also only one of many ways to quantify certainty in talent evaluation.
1
0
1
I looked at the last three years of Bleacher Report NBA top draft prospect player comparisons (multiple comps per prospect were allowed). Sure enough there is statistical evidence that player comps are driven by race. My personal favorite: Corey Kispert-Joe Harris.
0
0
4
@evilempire925 @getnickwright @FTFonFS1 "...It's going to be 3-1 going back to San Francisco and it's not going back to Dallas after that..." If you ignore everything he said outside of that he nailed it!
0
0
3
@FaxSem @ShamsCharania If nuggets had Murray they wouldn't have been the 6 seed. Warriors played a 6 seed caliber team..
0
0
1
@jacklich10 Just saw your presentation on this and was super impressed, particularly with how you accounted for the many possible outcomes. It's not all about return yards! Have you thought about publishing your approach?
0
0
0
@NFLFootballOps I loved the approach here. Treating launch angle and direction as the free variables gives a nice framework with enough complexity while still yielding a result that is easily digestible. I don't see why every punter in the league wouldn't like to see this applied to themselves.
0
0
0
RT @NFLFootballOps: It's #BigDataBowl week, powered by @awscloud! This year's challenge is to create innovative approaches analyzing the sp…
0
28
0
RT @StatsbyLopez: It’s #BigDataBowl week. Stay tuned for videos from each of this years finalists! And we’ll announce the winner on Frida…
0
3
0
RT @WMoneyball: Today's Show is available in Podcast form NOW! 0:00: This Week in COVID-19 30:32: This Week in Sports 1:09:15: @UPennDB…
0
3
0
@StatsbyLopez any theme that let's us use nexgen player tracking data for all 22 players on the field again!
0
0
5
@StatsbyLopez Had a blast working on this! The data was so pristine it was amazing not having to fill up my code with error/missing data checks everywhere. A true pleasure. Also, by my count that is 2 UPenn open finalists and 1 honorable mention! What a showing!
2
0
8
It seems everything #BigDataBowl happens on here, so I might as well make the leap and start up with twitter
0
0
8