How Losing To The Celtics Changes The Warriors’ Odds Of 73 Wins

https://espnfivethirtyeight.files.wordpress.com/2016/04/ap_509524628100.jpg?w=575



How Losing To The Celtics Changes The Warriors’ Odds Of 73 Wins



By NEIL PAINE and CARL BIALIK






The Golden State Warriors’ Stephen Curry during a game against the Boston Celtics on Friday in Oakland, California. MARCIO JOSE SANCHEZ / AP




On Thursday, we pegged the Golden State Warriors’ chance of winning 73 games — thereby breaking the 1995-96 Chicago Bulls’ all-time NBA record for most victories in a season — at 85 percent. But Friday night, the improbable happened: The Boston Celtics beat Golden State in Oakland, snapping the Warriors’ 54-game winning streak at Oracle Arena and handing them their eighth loss of the season.

Needless to say, this complicates the Warriors’ Quest for 73 a bit.

CHANCE OF…
RECORD...............EXACTLY THIS MANY WINS............AT LEAST THIS MANY WINS
68-14..................0.0%.........................................100.0%
69-13..................0.2............................................100.0
70-12..................1.7............................................99.8
71-11..................10.7...........................................98.1
72-10..................32.8...........................................87.4
73-9....................40.1...........................................54.6
74-8....................14.4...........................................14.4


According to our “CARM-Elo” forecast model, the Warriors’ probability of winning 73 or more games is down to just 54.6 percent — the loss to Boston dinged Golden State’s chances of making history by about 31 percentage points. Above, you can see the current breakdown of possible records Golden State could finish the season with; with six games left, now there are only two combinations (74-8 and 73-9) that would result in the Bulls’ record falling.

ODDS IN THIS GAME OF…
DATE................OPPONENT..............WIN PROB..........TYING RECORD........BREAKING RECORD
April 3..............vs. POR..................89%..................0%........................0%
April 5..............vs. MIN..................95.....................0...........................0
April 7..............vs. SA....................64.....................0...........................0
April 9..............@ MEM...................82.....................45..........................0
April 10.............@ SA.....................31.....................12.........................15
April 13.............vs. MEM.................93.....................30.........................39


Upcoming Warriors games

Golden State’s schedule isn’t doing them many favors, either. Upcoming games against Portland and Minnesota are extremely winnable, and a home-and-home with Memphis shouldn’t pose too much trouble — but it’s worth noting that even their chance of sweeping that group is only about 64 percent.1 And the Warriors also have to face the San Antonio Spurs twice in the final week of the regular season, including one date in Texas on the second leg of a back-to-back. San Antonio’s Elo rating is only a hair below Golden State’s, so assuming the Spurs play at full strength (never a certainty with Gregg Popovich) they’d be 69 percent favorites over the Warriors at the AT&T Center on April 10th.

As things stand now, that means Golden State would most likely tie the Bulls’ record on April 9 against the Grizzlies. Meanwhile, their quest for breaking the record will probably come down to the very final day of the regular season, on April 13 (also against Memphis).

After a season entirely spent ahead of pace to break the record, how did the Warriors find themselves in such a precarious position?




There were troubling signs for Golden State even before last night’s loss. On Thursday, they barely escaped with a win at Utah. Their previous three games all were home wins by smaller margins than our CARM-Elo-based point spread suggested. And since the All-Star Break, the Warriors have outscored opponents by an average of 3.6 fewer points per game than our spreads predicted.




This, if anything, understates the Warriors’ decline since the All-Star break. Our point spreads are based in part on teams’ Elo ratings. The Warriors’ Elo at the All-Star break was 1838 — just one point shy of their peak this year after winning in Boston by five on Dec. 11. Today the Warriors’ Elo is 1790 — below their level at the end of last year’s playoffs and just ahead of where the Bulls ended the 1995-96 season. As the Warriors’ Elo has declined, the dominance expected of them by our spreads has declined with it — yet they’ve disappointed even relative to those moving goalposts.

The Warriors are still great, still the best team in the NBA and one of the best in its history. But lately they haven’t been as good as they were when they started the season 24-0. That diminishes their chances of breaking the wins record, of beating the Spurs in their two remaining regular-season games — and of beating them if they meet again in the Western Conference finals.

Jay Boice contributed research.

Footnotes

We say “about” because we computed this using the game-by-game win probability for each upcoming matchup. A more accurate method (which we use for our NBA interactive) would account for the possibility that each team’s rating — and therefore each game’s odds — will change dynamically after each result. But that more involved calculation wouldn’t yield a result that differs significantly from our back-of-the-envelope estimate. ^



bob



.