could an ELO system work in the NBA?

[morosis]

yay for offseason threads. lol.

like many of you i have been thinking about how to make the RS more interesting, since despite all the injuries and wear and tear on players, the league seems hell bent on keeping the RS at 82 games. this led me to think that the stakes arent high enough to play your best players and take individual games seriously late in the season. which led me to think that winning or losing any individual game simply does not have enough value as just one single game out of 82.

installing an ELO system could help this. it would add extra weight to late season matchups between good teams. it would also punish good teams for losing against bad teams, especially late in the season. it could have rewards for breaking other teams' win streaks. by the end of a given season it would provide a more accurate stack ranking of who the best teams are than raw w/l does (similar to SRS) except it could actually be used (in some faraway future) to determine playoff seeding instead of the weird tiered tiebreakers we have today.

what sayeth you, RGM friends? could it work? would it be fun or stupid? are there unintended consequences i'm missing that could/would ruin the sport?

please take this as something fun and interesting to kick around in the abstract. it is not a serious proposal or request for change

edit: for those who might not be aware https://en.wikipedia.org/wiki/Elo_rating_system

[The Master]

Considering the rapid growth in number of injuries, I truly don't think that taking lightly the end of the regular season is really the NBA's problem. Actually, the league should think about completely new system of regular season that would make it more interesting with less games played, but this is at this point unrealistic.

But in overall, yeah, I think that reshaping the regular season would be a good idea and ELO system in idealistic scenario could be a solution.

[Backcountry]

I always liked Electric Light Orchestra.

[SkyBill40]

I always liked Electric Light Orchestra.

Fire on High is a great song.

[Clav]

I always liked Electric Light Orchestra.

I've seen ELO systems for other sports, but not exactly for basketball, so it wouldn't be some novel discovery out of the blue, but with some time we could really use it to find the best player, like, a new world record of basketball prowess.

Send it.

[CptCrunch]

I mean there are already tons of ELO NBA rankings online. It takes less than <1 hour to implement for a skilled developer (even less with an LLM). Here is my version that I just spun up just now.



Interactive web app here; https://nbaelo.netlify.app

[benson13]

yay for offseason threads. lol.

like many of you i have been thinking about how to make the RS more interesting, since despite all the injuries and wear and tear on players, the league seems hell bent on keeping the RS at 82 games. this led me to think that the stakes arent high enough to play your best players and take individual games seriously late in the season. which led me to think that winning or losing any individual game simply does not have enough value as just one single game out of 82.

installing an ELO system could help this. it would add extra weight to late season matchups between good teams. it would also punish good teams for losing against bad teams, especially late in the season. it could have rewards for breaking other teams' win streaks. by the end of a given season it would provide a more accurate stack ranking of who the best teams are than raw w/l does (similar to SRS) except it could actually be used (in some faraway future) to determine playoff seeding instead of the weird tiered tiebreakers we have today.

what sayeth you, RGM friends? could it work? would it be fun or stupid? are there unintended consequences i'm missing that could/would ruin the sport?

please take this as something fun and interesting to kick around in the abstract. it is not a serious proposal or request for change

edit: for those who might not be aware https://en.wikipedia.org/wiki/Elo_rating_system

Pickleball is a lot of fun. My gym allotted two hours twice a week to competitive play. I went to that a couple times thinking it would be enjoyable.

No. There are middle aged white guys who keep talking about this 5 point system and arbitrary rankings. You're a 3.8, but if you team with a 3.2 and beat a couple 3.5's your rank will go up more. They basically ruined it, and everyone stopped showing up. It's an open play session now, and it's still fun.

The middle aged white guys, well they do whatever at the gym, and, like all middle class white guys, wind up sitting naked in a steam room together.

[morosis]

I mean there are already tons of ELO NBA rankings online. It takes less than <1 hour to implement for a skilled developer (even less with an LLM). Here is my version that I just spun up just now.

Interactive version at link: https://graceful-malabi-6348f7.netlify.app/

correct i know people use it to gauge team quality and such at present, but since its not implemented and exposed, it doesnt influence the way teams approach individual games. in the alternate timeline i'm talking about, a team would know its own ELO and the ELO of the opponent going into a game, so the thought i had was, wouldnt that change the hype or individual player participation decisions late in the season, if it were the case? and wouldnt it be a really good input to have for playoff seeding?

example: According to the model you posted, the ELO-ified WC bracket for last year would have been:
1) OKC
2) DEN
3) MIN
4) GSW
5) HOU
6) LAL
7) LAC
8) SAC

that seems like a more accurate stack ranking of the WC going into the playoffs than the current W/L + tiebreaker system produced. and it fits what we saw in the playoffs too. GSW vs HOU was evenly matched (and GSW was actually the better team since their ELO is 50% influenced by pre-trade deadline results). OKC vs DEN was very close. DEN vs LAC was close as well bc despite their seeds, they were only separated by ~50 ELO points out of like 1,600.

[CptCrunch]

I mean there are already tons of ELO NBA rankings online. It takes less than <1 hour to implement for a skilled developer (even less with an LLM). Here is my version that I just spun up just now.

Interactive version at link: https://graceful-malabi-6348f7.netlify.app/

correct i know people use it to gauge team quality and such at present, but since its not implemented and exposed, it doesnt influence the way teams approach individual games. in the alternate timeline i'm talking about, a team would know its own ELO and the ELO of the opponent going into a game, so the thought i had was, wouldnt that change the hype or individual player participation decisions late in the season, if it were the case? and wouldnt it be a really good input to have for playoff seeding?

example: According to the model you posted, the ELO-ified WC bracket for last year would have been:
1) OKC
2) DEN
3) MIN
4) GSW
5) HOU
6) LAL
7) LAC
8) SAC

that seems like a more accurate stack ranking of the WC going into the playoffs than the current W/L + tiebreaker system produced. and it fits what we saw in the playoffs too. GSW vs HOU was evenly matched (and GSW was actually the better team since their ELO is 50% influenced by pre-trade deadline results). OKC vs DEN was very close. DEN vs LAC was close as well bc despite their seeds, they were only separated by ~50 ELO points out of like 1,600.

Made this into a web app;

https://nbaelo.netlify.app

[BigDan245]

Yes an ELO system will bring in a ton of more fans...Casuals are just dying for it.

[picko]

A system that a vast majority of fans wouldn't remotely understand would absolutely be super popular.

And I can ensure you that an opaque system that few people understand definitely won't be a conduit for conspiracy theories.

[ryguy613]

I always liked Electric Light Orchestra.

Damn it man, how am i supposed to make this grampa joke if you already made it hours ago? rude, is what it is....

[morosis]

A system that a vast majority of fans wouldn't remotely understand would absolutely be super popular.

And I can ensure you that an opaque system that few people understand definitely won't be a conduit for conspiracy theories.

the reason why i think it makes more sense nowadays than it has historically is mainly because of the explosion of people who play online games and watch e-sports as well. ELO/MMR systems are very common and widely understood in that universe.

i think anyone 20 years old or younger would pretty much immediately understand what it is, and that its doing the same things that the system they know from League of Legends or whatever other competitive online game does. it also doesnt need a ton of complexity, so its not the kind of thing where the formula would need to be obfuscated or anything, either.

[Sofia]

If NBA introduced this ****, I would turn to stone

[monopoman]

Considering the rapid growth in number of injuries, I truly don't think that taking lightly the end of the regular season is really the NBA's problem. Actually, the league should think about completely new system of regular season that would make it more interesting with less games played, but this is at this point unrealistic.

But in overall, yeah, I think that reshaping the regular season would be a good idea and ELO system in idealistic scenario could be a solution.

They are never cutting down on the number of games, despite what people think most of the money teams make is by ticket sales and concessions and therefore 82 games >>>>> 60 games.

I know people think everyone would massively watch the games more if there was less, but the only way to massively boost fan interest is make it like the NFL and there is no way that is happening. I doubt fans would go ballistic and start massively hyping up a 60 game season vs. an 82 game season.

People also bitch and moan with the introduction of the play-in game, I have a feeling many people would also call this ELO system bull or claim it heavily favors certain teams that play a certain way.