View Single Post
Old 07-05-2016, 06:58 PM   #13
Doppleganger's Avatar
Join Date: Mar 2007
Location: Yukinomiya City, Fukushima Prefecture
Posts: 11,629
Send a message via AIM to Doppleganger
Originally Posted by phoopes View Post
See Dopple normally I am all about stats and numbers like that, but Elo doesn't always tell the whole story. Like for example, in the World Cup Elo had Brazil as the favorite to take down Germany. Even when adjusting Elo for the absence of Neymar and Tiago Silva (I don't know how Nate Silver and friends adjusted exactly though) they still came out as slight favorites... and then proceeded to lose 7-1. Now yes, they were just favorites, not guaranteed to win, but if Brazil truly was favored then the percentage of them actually losing 7-1 was so minuscule that I think the issue was that the rating system didn't do a good job rather than Germany pulling off a really unlikely victory as an underdog. Yeah that example wasn't basketball, but I think it's one of the best examples of Elo not being the best system all the time, which is something to keep in mind. I love Nate Silver and FiveThirtyEight, but I had to point that out.
Oh, I totally agree.

Elo isn't perfect and isn't even close to it. The models predicted the Spurs had a 71% chance of beating the Thunder even before Game 6 of their semifinal match-up. That jived with the intuition most people felt watching the Thunder's play leading up to Game 6. Similar projection against Golden State in the finals.

The model's bias is the same problem plaguing most - they tend to gloss over sample size variance because it's not predictive. Lay people can and do make bets on short term trends, but in the long run the models win out. But the priorities are different. The 2007 Patriots, for example, had a great season on paper but we all know it was a colossal failure for reasons models still don't get.

But I don't have to use Elo. The Warriors came back from 3-1 to beat the Spurs. The Cavaliers came back from 3-1 to beat the Warriors. The Spurs won 67 games. The Warriors won 73 games.

All four of those events at once, compared to the history of the NBA, seem like kismet (or rigging). That's why stuff like the Warriors/Spurs defeat seemed impossible. But I see it as changing times, and I would bet that my interpretation is the right answer.

Originally Posted by phoopes View Post
Back to basketball though, even if you had "the stats" to show that the Warriors had a better regular season record and won the title against better playoff competition, there's still the fact of the whole "legacy" thing that people remember. It's very subjective but I do think it has some weight. Off the top of my head I'm pretty sure Jordan's Bulls were never taken to seven games in a playoff series in the six years they won titles. And IIRC the worst situation they were ever in was down 2-0 to a team. The Warriors had to go to seven games twice this postseason, and that included a 3-1 comeback to the Thunder, who looked like the better team for much of the series if you ask me. Yes, we just brought up the point that the Warriors faced better competition (something I agree with) but they didn't have the same level of success in the postseason as those Bulls teams, which is part of why people consider Jordan's Bulls the greatest team of all time.

Basically what I'm saying is this: in order to be the greatest team of all time in the court of public opinion, you need to win 72+ games and win ever postseason series in six games or less. And even then, people might still not give them credit because people are such Michael Jordan worshippers. Or, even more stupidly, they'll use the signing of Kevin Durant to somehow discredit them. It's the whole "I hate the Yankees, they're only good because they bought success" thing... Even though before this signing the Warriors top four players (Curry, Thompson, Green, Barnes) were all drafted.
That's a problem of comparing eras, but I think the momentum is shifting toward more advanced metrics and understanding as fans get smarter. It's a revolution that started in baseball and is only just starting to take over the NBA. In time, I think people will start to add asterisks to Jordan's time as opposed to Curry's*.

*Babe Ruth dominated in pre-integration, pre-international MLB. In modern MLB, he'd be closer to David Ortiz than Barry Bonds.
今 信じあえる
あきらめない 心かさね
Doppleganger is offline   Reply With Quote