SRS Upsets, Conference Finals Style
Posted by Neil Paine on May 26, 2009
Well, this past week's games haven't exactly gone according to the numbers, have they? In both of this year's conference finals, we're headed for some pretty serious upsets, at least if you believed the Simple Rating System -- in the East, the Orlando Magic (SRS: 6.49) are really wearing out the Cleveland Cavaliers (8.68), and while the Denver Nuggets (3.13)-vs.-L.A. Lakers (7.11) West finals are technically tied, you can't help but see the Nuggets as the more dominant team in the series right now.
How common is this? I mean, according to SRS, Denver only had an 18.5% probability of winning a 7-game series vs. L.A. without home court advantage, and Orlando's chances were only marginally better at 26.9%. So how many times in the modern (16-team) NBA playoffs have the conference finals produced upsets of this magnitude?
(All probabilities assume a constant 60% home-court advantage.)
| Year | Winner | SRS | Loser | SRS | p(Upset) |
|---|---|---|---|---|---|
| 1986 | Houston | 2.11 | LA Lakers | 6.84 | 0.153 |
| 2001 | LA Lakers | 3.74 | San Antonio | 7.92 | 0.165 |
| 2009 | Denver? | 3.13 | LA Lakers | 7.11 | 0.185 |
| 1995 | Houston | 2.32 | San Antonio | 5.90 | 0.217 |
| 2009 | Orlando? | 6.49 | Cleveland | 8.68 | 0.269 |
| 2006 | Miami | 3.59 | Detroit | 6.23 | 0.269 |
| 2005 | Detroit | 3.31 | Miami | 5.76 | 0.285 |
| 1999 | New York | 1.45 | Indiana | 3.86 | 0.298 |
| 2003 | San Antonio | 5.65 | Dallas | 7.91 | 0.331 |
| 2004 | LA Lakers | 4.35 | Minnesota | 5.86 | 0.348 |
| 1988 | Detroit | 5.46 | Boston | 6.15 | 0.410 |
| 2002 | LA Lakers | 7.15 | Sacramento | 7.61 | 0.424 |
| 1998 | Utah | 5.73 | LA Lakers | 6.88 | 0.431 |
| 2007 | Cleveland | 3.33 | Detroit | 3.68 | 0.441 |
| 2004 | Detroit | 5.04 | Indiana | 4.93 | 0.476 |
| 1990 | Portland | 6.48 | Phoenix | 7.09 | 0.476 |
| 1989 | LA Lakers | 6.38 | Phoenix | 6.84 | 0.491 |
| 1993 | Chicago | 6.19 | New York | 5.87 | 0.496 |
| 1993 | Phoenix | 6.27 | Seattle | 6.66 | 0.497 |
As you can see, a Denver win over the Lakers this year would be the 3rd-most improbable conference finals upset since the league expanded to a 16-team playoff bracket in 1984. And an Orlando win (which is looking more and more like a reality with each passing game) would rank as the 5th-biggest upset.
To further heap perspective on what these two teams are attempting to do, a Nuggets win would be, going into the series, approximately as probable as if the Bulls had upset the Pistons in 1989, or if the Nuggets had upset the Lakers (sound familiar?) in 1984. And a Magic win would be an upset on the magnitude of the Knicks upsetting the Pacers in 2000, or the Bulls upsetting the Pistons in 1990 (which they nearly did, pushing the defending champs to 7 games).
But while these are improbable outcomes so far (before the series, the probability of Orlando being up 2-1 on Cleveland after 3 was supposedly 26.1%), upsets like this are not quite as rare as we might think. People love to say "the better team will win a 7-game series," but that's really not always the case. Does the better team usually win? Yes. But over the past 26 years (52 conference finals), you can see that upsets like the kind Orlando and Denver are currently trying to pull off, while rare, are not completely without precedent.
By the way, SRS now says Denver has a 24.5% shot at eventually winning their series with L.A. (most likely in 7 games), and that Orlando has a 53.6% probability of knocking off Cleveland, most likely in 6 games.
May 26th, 2009 at 8:25 am
Houston beating the Lakers in 1986 prevented Magic and Kareem from playing 8 finals in a row (!!) from 1982 to 1989. The Rockets were led by 3rd year Samspson and 2nd year Olajuwon so I guess that was indeed as big an upset as it can get in a conference final.
Funny to see that upsets n°2 and n°3 actually are defending champs making it back to the finals : the Lakers in 2001 and Houston in 1995. For some reasons these teams had underperformed in the regular season (well, especially Houston), but were able to turn it on in the play-off to eventually win it all again. These teams were much better than their SRS showed, and I think eveybody was well aware of that for the 2001 Lakers so I would not call their win an upset.
May 26th, 2009 at 12:14 pm
Romain - 2001 Lakers had an okay regular season really (they were top 3 in the league), but Shaq, Kobe, and Fisher who all had injuries during the season all got healthy for the playoffs, and they went on a tear. Also, if memory serves, didn't Duncan have a bum ankle in those playoffs? I could be thinking of 2004...
1995 Rockets had some trouble getting Clyde integrated in the last 3rd of the season when they had a very tough schedule, and they wound up losing more games than they should have. However, once he got in the flow, Clyde added a ton of punch for the playoffs. The saving grace for them advancing though was Hakeem's ability to dominate Barkley and Robinson to get them through a very tough conference to the finals.
Funny that #s 6 & 7 (which I guess should be considered 4 & 5 until the current conference finals are over) are the Detroit beating Miami (thanks to injuries to Wade and Shaq) one year and then getting beat by Miami the next. Tough to go from the being the conquering underdog to the conquered favorite against the exact same team!
May 26th, 2009 at 2:30 pm
Three of the Shaq and Kobe Laker teams achieved upsets, eh? Was it coincidence, or providence?
Cleveland has some serious matchup problems with Orlando that the stats rooted in point differential and their general dominance of the league just didn't take into account. (hence Hollinger's apology article at ESPN.com) The Lakers will probably still beat Denver due to the back and forth officiating which typically helps the home court team, but even with Lebron taking 24 free throws in game three, the Cavaliers lost. It's clear that without Mo Williams playing like an all-star, even the refs may not be able to save Cleveland. That means the dream matchup which Nike and Vitamin Water started promoting prematurely may be in danger of not happening. They should've known better than to sell Denver and Orlando short like that.
May 26th, 2009 at 3:42 pm
To play devil's advocate: does that mean the playoffs really determine the best team, or merely the one that lucked into the most favorable matchups? And, if so, is a -- gasp! -- BCS-like approach actually the better solution to the championship question?
May 26th, 2009 at 4:33 pm
Maybe so. Would Dallas have advanced farther into the 2007 playoffs had they avoided Golden State? Probably. Same goes for Portland and how they got screwed by a first-round matchup against Houston. In the case of Cleveland against Orlando, one would think that being the best team would give Cleveland a shot of overcoming matchup problems through sheer determination and team chemistry (which to me is the reason a Cleveland team with arguably only one legit all-star managed to win 66 games; by contrast, I think that a lack of chemistry has kept L.A. from playing the way everyone thinks they should, and yet their talent and size carried them to 65 wins), but if that chemistry doesn't show up soon, Orlando will ride their matchup advantages to the finals, and the best team won't have a championship trophy by season's end.
Determination and chemistry has carried Denver to a 2-2 series against a Laker team that should have a clear advantage in the frontcourt, so you never know. If Cleveland wins tonight, it's a whole new series. If they lose, then everyone will chalk up their imminent doom to the matchup problems and/or the lack of help for Lebron.
May 26th, 2009 at 4:59 pm
"And, if so, is a — gasp! — BCS-like approach actually the better solution to the championship question?"
This would make a great future post Neil, but as someone who's witnessed countless "BCS fights" between rabid college football fans, you really, really, REALLY don't wanna go there anytime soon :)
May 27th, 2009 at 6:07 am
Orlando is now up 3-1 and Lebron is averaging 42 ppg over the 4 games
--> It would be interesting to know how many times has a team lost a play off series during which one of its players averaged 40+ ppg.
I know the Bulls lost to the Celtics in 1986 with MJ averaging 43.7 ppg but are there other examples?
May 27th, 2009 at 9:15 am
Romain - Good question! Maybe Wilt? AI? King?
Neil - I would say it's a little bit much to chalk it all up to matchups. It's pretty tough to fall into 4 consecutive lucky matchups, and it's also hard for a lesser team to maintain enough focus to take advantage of favorable matchups to "steal" 4 games against a significantly better team. Generally speaking the NBA isn't a "rock / paper / scissors" league. It doesn't go - Boston can beat LA, LA can beat Cleveland, Cleveland can beat Boston and that's how it works every time. It's more of a case of all the top teams can beat all the top teams and the difference is usually focus and effort.
The Dallas / GS example is interesting, but I gotta think a true championship team doesn't let one chink in its armor get exploited to the point of losing a seven game series (two series in a row actually). Blame the coaches. Blame the players. Blame rules change that made it illegal to defend on the perimeter. This goes beyond luck and matchups in my mind.
May 27th, 2009 at 12:53 pm
Romain - I suspect that Jerry West and/or Elgin Baylor might be among those players. It's rare for a player to score 40 ppg in ANY series, however, nevermind limiting it to series which have been lost.
"Blame rules change that made it illegal to defend on the perimeter."
Indeed. And I agree that a great team should be able to overcome matchup problems over a lengthy series, but you never know when the 'inferior' team will get inspired and surprise you. Cleveland played well enough to win last night, but Rafer Alson came up with a career night and played spoiler. Not to mention the timely shots from Rashard Lewis, which are happening consistently enough that one might think it's a pattern and not just good fortune. Orlando doesn't fear Cleveland at all. They believe they can win, and they're one game away from proving it.
When you think about it, three of the games have been decided a point or two, and unfortunately for Cleveland, they've been on the losing side of two of them. People say that Cleveland should've been down 0-3 after the first three, but it could've been 2-1 (and now possibly 3-1) for Cleveland just the same. Only one game was decided by more than a one or two point margin, and that was game three. Otherwise all the games have been really close, which might suggest that Cleveland is the better team (and heck, Cleveland was up by 23 in one of the games, but they just couldn't sustain the lead), but that the matchup problems have closed the gap enough for Orlando to make each game a 50/50 proposition.
I seriously doubt that Cleveland would be down 3-1 to this postseason's Boston team, so I do think that matchups affect the percentages, if nothing else. It's not as simple as RPS, but matchups are definitely a factor, and matchups are forcing Cleveland to beat Orlando with offense (it's crazy; Cleveland held Atlanta and Detroit to 78 ppg whereas Orland is averaging 104 ppg against them), which isn't Cleveland's bread and butter despite the improvements Cleveland has made on that end. I think the biggest problem is that Cleveland doesn't have a Perkins to defend Howard one-on-one in the paint. The result is that Cleveland is being forced to pick their poison, and so far, both Orlando's perimeter players as well as Howard have taken advantage of their opportunities.
May 27th, 2009 at 2:53 pm
The answer to the question is Elgin Baylor, in 1962.