tothemean.com

Data science + sports

[Jesse Fischer @jessefischer33]

However...if we segment by age when drafted then things look a little different. Here is correlation between NCAA RAPM+ and NBA Peak Wins by year in school:
FR = 0.39
SO = 0.29
JR = 0.26
SR = 0.21

12/

[Jesse Fischer @jessefischer33]

Top NCAA RAPM for NBA draft prospects (RAPM variations: luck-adjusted, vanilla, box score prior)

Jaylin👀Elite impact with the age of a freshman (prediction: goes higher than #42)

[tothemean.com @tothemean]

Just got https://t.co/7NSzI9468S for you - let’s make it happen! $71940 in honor of Peyton and I’ll donate it to charity of your choice @SnoopDogg @PeytonnManning @EliManning #ManningCast

[Jesse Fischer @jessefischer33]

Added various #NBADraft2021 content (projections, statistical profiles, historical comps) I have been sharing as a blog post on @tothemean https://t.co/poK4ERFN9l

[Jesse Fischer @jessefischer33]

Check out my work on NFL Ghost Defenders where I used CNN + NFL tracking data to learn optimal defender positioning and movement #BigDataBowl
https://t.co/s07jcvfAqQ

[Jesse Fischer @jessefischer33]

In this example Russ increases play success by stepping outside the pocket and throwing a dime to Lockett. Interestingly, we can see real-time EPA drop for a few frames as the pass goes over the head of the defender highlighting an opportunity to improve (football z-axis data!)

[Luis Clausín @luisclausinmb]

.@tothemean: When do estimated statistics differ from actual (play by play)?
1 Individual players impact team pace
2 Individual players change team dynamics
3 Players with ⬇️ playing time have ⬇️ influence on team totals
4 Efficiency impacts opportunities
https://t.co/HHPh4trszL

[tothemean.com @tothemean]

#32 Carey is a potential top 10 draft pick in another era and one of the highest starter+star probabilities in the class but does lack wingspan compared to similar archetypes

[tothemean.com @tothemean]

Pick #31 Terry has some of the highest star probability remaining with his shooting and usage comparable with top guards in the class, but also high bust potential partly driven by weak BLK%

[tothemean.com @tothemean]

Pick #30 Bane had noticeable changes to his game his senior year (+AST/USG/STL/3P%) but dropped off in other ways (-FTR/EFG) - model didn't like net result but might be due to just age increase

[tothemean.com @tothemean]

Pick #29 Flynn has played his way into the first round with consistent improvement every year (RAPM last three years of 1.7...2.5...9)

[tothemean.com @tothemean]

Pick #28 Jaden McDaniels is the player you use to know which draft model(s) are/aren't heavily weighting scouting rank (high school, projected draft pick)

[tothemean.com @tothemean]

Pick #27 Udoka somehow improved an already impressive BLK% his last 3 years at Kansas but lacks star probability upside. Length is only second to Tacko comp

[tothemean.com @tothemean]

Pick #26 Pritchard has a lower starter probability but higher role probability when compared to similar players projected in his range - but does have a path to playing time as seen with his comps

[tothemean.com @tothemean]

Pick #25 Quickly improved his 3P and FT shooting out of nowhere last year (relative to prior data points). Model isn't excited though because he doesn't do much else, however his high RAPM (+5 on offense) highlight he might be doing more than the counting stats imply?

[tothemean.com @tothemean]

Pick #24 RJ Hampton has a path to formidable NBA player with high USG/STL and low TOV but lacks wingspan relative to more successful comps

[tothemean.com @tothemean]

Pick #23 Bolmaro has height (6-7) but plays like he doesn't (all comps 6-3 or under). Model has been lower on him as time has gone on

[tothemean.com @tothemean]

Pick #22 Zeke Nnaji has one of the lowest floors (high role player + starter probabilities) but lacks upside. Every single comp was drafted (rare) but no all star on the list highlight his likely future as only a solid contributor

[tothemean.com @tothemean]

Pick #21 Maxey showed he can shoot in AAU (52.9% on 34 attempts) but regressed at Kentucky 29.2% on 113 attempts). Draft model isn't high on him but is optimistic given his high (and improving FT%) and low TOV

[tothemean.com @tothemean]

Pick #20 Precious showed glimmers of elite defense (4.2 defensive RAPM) but isn't projected well because of his high turnover rate and one of the lowest EFG for a big in the class