ESPN should hire a Data Engineer that understands hoops

[dhsilv2]

https://www.nbarapm.com/ is a pretty good site, I'm sure someone's going to get hired from there to take it down lol

[eyriq]

NBA Team 4FA - ESPN Analytics https://share.google/Uf9rD7Ls6Q6St6ibs

Most people will be (wisely) uncomfortable clicking on Google Share link.

https://espnanalytics.com/nba-team-4FA

Thanks! I posted from my phone.

[Bad Bart]

Net points and BPI are the best advanced analytics around. Stop sleeping on their recent improvements.

NBA Team 4FA - ESPN Analytics https://share.google/Uf9rD7Ls6Q6St6ibs

Wow, did not know this exists. Thank you

[magee]

As a data person, it kills me how ESPN sucks in data that's not a random "Lebron is the first ever 6'8 to score 27+ points with 10 boards and under 1 TOs on a random Tuesday night in Charlotte when it rains outside and the opposite team wears white jerseys" fact thrown on the screen during a timeout.

Exhibit A:
Scoot is yet to play this season.
On his player page, it says
https://www.espn.com/nba/player/_/id/4683678/scoot-henderson
Stats GP MIN FG% 3P% FT% REB AST BLK STL PF TO PTS
Regular Season 66 26.7 41.9 35.4 76.7 3.0 5.1 0.2 1.0 2.7 2.7 12.7
Career 128 27.5 40.1 34.0 79.3 3.1 5.2 0.2 0.9 2.9 3.0 13.3


So they they still show his stats from last season in a row deserved for current season stats.


Exhibit B:
Dylan Harper has played so far in 4 NBA games in his young career.
https://www.espn.com/nba/player/_/id/5037871/dylan-harper
Stats GP MIN FG% 3P% FT% REB AST BLK STL PF TO PTS
Regular Season 3 26.0 48.6 28.6 83.3 5.0 4.3 0.0 1.0 4.7 2.3 16.0
Career 4 26.0 47.8 33.3 85.7 5.3 4.8 0.0 1.0 4.3 2.0 14.8



How the hell he has 3 games in this regular season and 4 games in his career?
And his last game wasn't even tonight - it was more than 24 hours ago.

It's obvious the Career row includes that last game, while the RS row - doesn't (thus the difference in all the stats).
It's clear to me they have two separate processes bringing RS and Career data and they aren't aligned - might be diff scheduling, even diff logic/filters.
It's not knew, I saw it at least last season as well, but only for last night games, never for the prev night not being updated.

And in Scoot's example, since he's an active player on the roster, but yet to have any stats this season, they somehow didn't nullify those Reg Season stats on the first day for everyone - they probably do it only when a new data arrives or something.
Any Junior Data Engineer who understand NBA will make it work easily.

/rant

Did Harper not play in the first game of the season? I wonder if they are counting the total number of games the team has played while factoring in the player being on the team even if they are hurt and not playing.

[madmaxmedia]

The "Worldwide Leader In Sports" appears to be perfectly content being a mediocre public source of sporting statistics. They should have just bought sports-reference years ago and merged it into ESPN.com. Or they could have used one drop from their wealth bucket to build an even better repository of statistics than the great sports-reference but they are just too creatively lazy and operationally conservative.

God please don't. ESPN keeps cutting costs anyway. They'd by now have just fired the staff for sprots reference and we'd be without. Just like they dropped 538, grantland, and even the RPM stats. ESPN just doesn't want to spend a dime on sports stats.

Good thing they probably care more about presenting fantasy stats for bettors anyway.

[Invictus88]

ESPN doesn't particularly care about data or facts or real analysis in its TV content. Why should its website be any different?

[Clav]

I agree, but ESPN is like baseline research for me. I like some of the interface and overall design but I rarely use it for anything other than looking at some season aggregates for players or the standings, or schedule. They have a consistent design which is fine with me. I apprecaite it when casually perusing other sports other than NBA.

But I get your complaint, if you want to use the world of statistics there is more to do for the user to dig and get that information. ESPN uses stats for their narratives like you commented about "a certain player this age in October that got 20/10/5/5 in 30 minutes" lol It's selective and obsfuscates from in depth statistical analysis. Unfortunately ESPN shows they don't want to pay for that service, it's not in their business model...

[BruttoNostra]

As a data person, it kills me how ESPN sucks in data that's not a random "Lebron is the first ever 6'8 to score 27+ points with 10 boards and under 1 TOs on a random Tuesday night in Charlotte when it rains outside and the opposite team wears white jerseys" fact thrown on the screen during a timeout.

Exhibit A:
Scoot is yet to play this season.
On his player page, it says
https://www.espn.com/nba/player/_/id/4683678/scoot-henderson
Stats GP MIN FG% 3P% FT% REB AST BLK STL PF TO PTS
Regular Season 66 26.7 41.9 35.4 76.7 3.0 5.1 0.2 1.0 2.7 2.7 12.7
Career 128 27.5 40.1 34.0 79.3 3.1 5.2 0.2 0.9 2.9 3.0 13.3


So they they still show his stats from last season in a row deserved for current season stats.


Exhibit B:
Dylan Harper has played so far in 4 NBA games in his young career.
https://www.espn.com/nba/player/_/id/5037871/dylan-harper
Stats GP MIN FG% 3P% FT% REB AST BLK STL PF TO PTS
Regular Season 3 26.0 48.6 28.6 83.3 5.0 4.3 0.0 1.0 4.7 2.3 16.0
Career 4 26.0 47.8 33.3 85.7 5.3 4.8 0.0 1.0 4.3 2.0 14.8



How the hell he has 3 games in this regular season and 4 games in his career?
And his last game wasn't even tonight - it was more than 24 hours ago.

It's obvious the Career row includes that last game, while the RS row - doesn't (thus the difference in all the stats).
It's clear to me they have two separate processes bringing RS and Career data and they aren't aligned - might be diff scheduling, even diff logic/filters.
It's not knew, I saw it at least last season as well, but only for last night games, never for the prev night not being updated.

And in Scoot's example, since he's an active player on the roster, but yet to have any stats this season, they somehow didn't nullify those Reg Season stats on the first day for everyone - they probably do it only when a new data arrives or something.
Any Junior Data Engineer who understand NBA will make it work easily.

/rant

Did Harper not play in the first game of the season? I wonder if they are counting the total number of games the team has played while factoring in the player being on the team even if they are hurt and not playing.

No, they count only games you actually played, and Harper legit played in all 4 games, and the difference in stats is exactly that 4th game (i.e. calculate his PPG after 3 and after 4 games)
It just caught my eye as someone who build a career making sure the data is refreshed, verified, there is a SSoT etc - and ESPN is so amateurish with that, like some stident project, not a group of professionals. There is no QA, no definition of done, no edge case checking. Amateurs.

[BruttoNostra]

I see Harper's stats finally got updated (until the next game and next delay, idk).
Looking for similar errors from player who played last night - check Edgecombe:
https://www.espn.com/nba/player/_/id/5124612/vj-edgecombe
He has 4 RS games, but only 3 Career games - a mirror case of Harper's data issue.

Scoot's bug is still there though.

[BruttoNostra]

I think that ship has sailed, NBA.com and BBref are much better sources for stats.

nba.com still shows 3 games and 25 PPG for Edgecombe who played his 4th game (and lowered his PPG) +- 20 hours ago.
https://www.nba.com/stats/player/1642845

Bball ref are fine though.

[JDR720]

I see Harper's stats finally got updated (until the next game and next delay, idk).
Looking for similar errors from player who played last night - check Edgecombe:
https://www.espn.com/nba/player/_/id/5124612/vj-edgecombe
He has 4 RS games, but only 3 Career games - a mirror case of Harper's data issue.

Scoot's bug is still there though.

If a rostered player hasn't played a game the current season, then their previous season stats are used (or the last season they played a game). It's how it is for all sports/players, not just the NBA. I'm pretty sure all of ESPN's (and NBA.com and basically all other sports data) is automated though. They just take data from the league or wherever and repost it. They usually lag behind at least a day or so of the last games end time.


ESPN shows Grant Williams having played 16 games too, https://www.espn.com/nba/player/_/id/4066218/grant-williams

Grant Williams on NBA.com displays his last seasons stats too.

https://www.nba.com/stats/player/1629684