This is our old blog. It hasn't been active since 2011. Please see the link above for our current blog or click the logo above to see all of the great data and content on this site.
«
»

The Finals Are Even… But Are They Really Even?

Posted by Neil Paine on June 16, 2010

The Lakers needed to win last night to force a 7th game in the NBA Finals, and they didn't disappoint, delivering a brutal 89-67 beatdown that evened the series at 3 wins apiece. In NBA playoff history, this is the 105th series to go 7 games, and the 17th time the Finals have gone the distance. It should be a fitting finish to the season, and the capper to a thrilling 7-game Championship series with a razor-thin margin between the two teams...

Or has it really been that razor-thin? In the series, the Lakers have outscored the Celtics by 20 points after 6 games. Is that a lot, or a normal amount for a 7-game series? How does that compare to the margins of past 7-game playoff series through 6 games?

Year Round Team A Pts Thru 6 Team B Pts Thru 6 Margin G7 Loc G7 Winner
2008 EC1 BOS 595 ATL 545 50 BOS BOS
1982 ECF BOS 627 PHI 580 47 BOS PHI
1959 EDF BOS 739 SYR 700 39 BOS BOS
1984 ECS BOS 656 NYK 617 39 BOS BOS
1971 ECF BAL 611 NYK 572 39 NYK BAL
1978 FIN WSB 631 SEA 593 38 SEA WSB
1962 EDF BOS 676 PHW 639 37 BOS BOS
1976 WCF GSW 672 PHO 635 37 GSW PHO
1960 FIN BOS 670 STL 634 36 BOS BOS
2003 EC1 DET 556 ORL 520 36 DET DET
1957 FIN BOS 681 STL 647 34 BOS BOS
1951 FIN ROC 504 NYK 471 33 ROC ROC
2009 WCS LAL 596 HOU 564 32 LAL LAL
1987 ECF DET 679 BOS 650 29 BOS BOS
1963 EDF BOS 725 CIN 697 28 BOS BOS
1977 ECS PHI 670 BOS 642 28 PHI PHI
2006 WCS LAC 686 PHO 659 27 PHO PHO
1988 WCF LAL 656 DAL 630 26 LAL LAL
1962 FIN BOS 714 LAL 688 26 BOS BOS
1981 WCS PHO 554 KCK 528 26 PHO KCK
1994 WCS HOU 627 PHO 602 25 HOU HOU
1984 FIN LAL 720 BOS 695 25 BOS BOS
1964 WDF SFW 667 STL 643 24 SFW SFW
1998 ECF CHI 583 IND 559 24 CHI CHI
1966 FIN BOS 732 LAL 709 23 BOS BOS
2005 EC1 BOS 544 IND 521 23 BOS IND
1960 WDF STL 646 MNL 624 22 STL STL
2006 ECS DET 527 CLE 505 22 DET DET
1996 WCF UTA 557 SEA 535 22 SEA SEA
1995 ECF IND 643 ORL 621 22 ORL ORL
1966 WDF LAL 701 STL 680 21 LAL LAL
1954 FIN MNL 436 SYR 415 21 MNL MNL
2005 WC1 HOU 607 DAL 586 21 DAL DAL
1988 FIN DET 604 LAL 583 21 LAL LAL
2006 WCS DAL 608 SAS 588 20 SAS DAL
2009 EC1 BOS 679 CHI 659 20 BOS BOS
2005 FIN DET 533 SAS 513 20 SAS SAS
2010 FIN LAL 551 BOS 531 20 LAL ???
1975 WCF GSW 573 CHI 554 19 GSW GSW
1965 EDF BOS 674 PHI 655 19 BOS BOS
1990 WCS SAS 686 POR 667 19 POR POR
2003 WC1 POR 614 DAL 595 19 DAL DAL
1981 ECS MIL 651 PHI 632 19 PHI PHI
2009 EC1 MIA 539 ATL 520 19 ATL ATL
2007 WC1 UTA 530 HOU 512 18 HOU UTA
1974 FIN BOS 575 MIL 557 18 MIL BOS
1974 WCS DET 571 CHI 553 18 CHI CHI
1978 WCS MIL 697 DEN 679 18 DEN DEN
2000 WCF POR 581 LAL 563 18 LAL LAL
1974 ECS CAP 585 NYK 567 18 NYK NYK
1994 ECS CHI 555 NYK 537 18 NYK NYK
1973 WCS LAL 611 CHI 594 17 LAL LAL
2005 ECF DET 553 MIA 536 17 MIA DET
2006 WC1 PHO 630 LAL 614 16 PHO PHO
1971 ECS BAL 659 PHI 643 16 BAL BAL
1979 ECS PHI 673 SAS 657 16 SAS SAS
2001 ECF MIL 538 PHI 523 15 PHI PHI
1968 EDF BOS 688 PHI 675 13 PHI BOS
2004 EC1 MIA 500 NOH 487 13 MIA MIA
1988 ECS ATL 638 BOS 625 13 BOS BOS
2008 ECS CLE 504 BOS 491 13 BOS BOS
2003 WCS DAL 690 SAC 678 12 DAL DAL
1993 WCF SEA 643 PHO 631 12 PHO PHO
1995 ECS NYK 568 IND 556 12 NYK IND
2009 ECS ORL 571 BOS 560 11 BOS ORL
2004 WCS SAC 571 MIN 560 11 MIN MIN
1994 FIN NYK 524 HOU 513 11 HOU HOU
1975 ECS WSB 628 BUF 618 10 WSB WSB
1973 ECF NYK 649 BOS 639 10 BOS NYK
1981 ECF BOS 638 PHI 628 10 BOS BOS
1980 WCS MIL 622 SEA 612 10 SEA SEA
1981 WCS HOU 663 SAS 654 9 SAS HOU
1994 WCS UTA 589 DEN 580 9 UTA UTA
2008 WCS NOH 563 SAS 554 9 NOH SAS
2002 WCF SAC 591 LAL 583 8 SAC LAL
1970 FIN LAL 681 NYK 673 8 NYK NYK
1970 WDS LAL 671 PHO 664 7 LAL LAL
1979 ECF SAS 650 WSB 643 7 WSB WSB
1963 WDF LAL 675 STL 669 6 LAL LAL
1976 ECS WSB 557 CLE 551 6 CLE CLE
1986 ECS PHI 670 MIL 664 6 MIL MIL
1971 WCS LAL 619 CHI 614 5 LAL LAL
1952 FIN MNL 496 NYK 491 5 MNL MNL
2000 ECS NYK 485 MIA 480 5 MIA NYK
2010 EC1 ATL 561 MIL 556 5 ATL ATL
1988 WCS UTA 605 LAL 600 5 LAL LAL
1955 FIN FTW 549 SYR 544 5 SYR SYR
1969 FIN LAL 638 BOS 633 5 LAL BOS
2001 ECS CHH 560 MIL 555 5 MIL MIL
1997 WCS HOU 610 SEA 606 4 HOU HOU
2001 ECS PHI 562 TOR 558 4 PHI PHI
1987 ECS MIL 735 BOS 731 4 BOS BOS
1994 ECF IND 522 NYK 518 4 NYK NYK
1993 WCS SEA 591 HOU 588 3 SEA SEA
1992 ECS CLE 625 BOS 622 3 CLE CLE
2004 ECS DET 517 NJN 514 3 DET DET
1979 WCF PHO 607 SEA 604 3 SEA SEA
1995 WCS HOU 649 PHO 647 2 PHO HOU
1990 ECF DET 579 CHI 577 2 DET DET
1961 WDF LAL 697 STL 695 2 STL STL
1970 EDS BAL 621 NYK 619 2 NYK NYK
1979 ECS ATL 602 WSB 600 2 WSB WSB
1992 ECS NYK 539 CHI 537 2 CHI CHI
1997 ECS NYK 509 MIA 507 2 MIA MIA
1977 WCS LAL 636 GSW 635 1 LAL LAL

A 20-point margin for one team after 6 games ranks the 2010 Finals in the middle of the pack historically. The average 7-game series sees one team outscoring the other by 16.6 through 6 games, so this series has been slightly more imbalanced than usual going into the 7th game, but not significantly so. Here's the above table again, but with the non-Finals series taken away:

Year Round Team A Pts Thru 6 Team B Pts Thru 6 Margin G7 Loc G7 Winner
1978 FIN WSB 631 SEA 593 38 SEA WSB
1960 FIN BOS 670 STL 634 36 BOS BOS
1957 FIN BOS 681 STL 647 34 BOS BOS
1951 FIN ROC 504 NYK 471 33 ROC ROC
1962 FIN BOS 714 LAL 688 26 BOS BOS
1984 FIN LAL 720 BOS 695 25 BOS BOS
1966 FIN BOS 732 LAL 709 23 BOS BOS
1954 FIN MNL 436 SYR 415 21 MNL MNL
1988 FIN DET 604 LAL 583 21 LAL LAL
2005 FIN DET 533 SAS 513 20 SAS SAS
2010 FIN LAL 551 BOS 531 20 LAL ???
1974 FIN BOS 575 MIL 557 18 MIL BOS
1994 FIN NYK 524 HOU 513 11 HOU HOU
1970 FIN LAL 681 NYK 673 8 NYK NYK
1952 FIN MNL 496 NYK 491 5 MNL MNL
1955 FIN FTW 549 SYR 544 5 SYR SYR
1969 FIN LAL 638 BOS 633 5 LAL BOS

Again, this year is basically par for the course -- the average 7-game Finals series has one team outscoring the other by 20.5 points going into the deciding game.

Now, does the margin going into the 7th game have any predictive power on who will win the rubber match? The team who outscored their opponent through 6 games went on to win the 7th matchup 53.8% of the time (56.3% in the Finals). However, by contrast, the team playing Game 7 at home won 81% of the time (81.3% in the Finals), so margin through 6 games isn't exactly the best predictor in the world. If you set up a logistic regression to predict Game 7 winners using only margin through 6 games, you get this equation:

Win% ~ 1 / (1 + exp(-0.0173*margin))

Which would yield an expected WPct for the Lakers of 58.6% in tomorrow's game. Then again, we also know that the Lakers will be at home tomorrow night, so if we add home-court information to the regression, we see this equation:

Win% ~ 1 / (1 + exp(1.4204 - 0.0144*margin - 2.8408*loc)), where loc = 1 if home, 0 if away

This yields an 84.7% probability that the Lakers will win tomorrow, which shows how little extra information the +20 margin provides once you know that L.A. is at home (remember, home teams win Game 7 81% of the time).

In other words, knowing that the Lakers have outscored Boston by 20 points tells you a little about how they will perform in Game 7, but knowing that L.A. will be playing the decisive game at home tells you a lot more.

Related posts:

    Which Position Should You Build Around For Playoff Success?
    Championship Usage Patterns III: Regular-Season Teams Built For the Playoffs
    Career Playoff Quality of Defenses Faced (1991-2010)
    Which Games Are the Most Important In a 7-Game Series?
    Best Road Performances in Opening-Round Game 1s?

This entry was posted on Wednesday, June 16th, 2010 at 11:23 am and is filed under Analysis, Playoffs, Statgeekery. You can follow any responses to this entry through the RSS 2.0 feed. Both comments and pings are currently closed.

15 Responses to “The Finals Are Even… But Are They Really Even?”

  1. Ben Says:
    June 16th, 2010 at 11:34 am

    Interesting. I know you're looking for simplicity, but would be neat to see regular season SRS difference thrown into the equation as well.

  2. Neil Paine Says:
    June 16th, 2010 at 11:43 am

    Sure, I can do that:

    WPct ~ 1 / (1 + exp(1.169572 - 0.012391*margin - 2.339144*loc - 0.116955*srsdiff))

    This says the Lakers have an 83.0% chance of winning tomorrow.

  3. Neil Paine Says:
    June 16th, 2010 at 11:45 am

    I should also note that the extra variables other than location aren't significant at a 0.05 level. Basically, all you need to know about the two teams' chances tomorrow night is that the game is in Los Angeles. 4 out of 5 times, the Lakers will win this game.

  4. Walter Says:
    June 16th, 2010 at 12:00 pm

    Nice work Neil!

    The coefficient for location, was that factor from a previous analysis based on regular season or did you derive that coefficient from the regression on the tables above?

    If based on regular season, then I would be curious as to how different the coefficient is if regressed on the actual dataset. I'm curious if homecourt has any more significance in game 7 over a regular game.

  5. Neil Paine Says:
    June 16th, 2010 at 12:02 pm

    That was just from historical Game 7 data. The problem with any study like this using playoff data is that it's tough to differentiate between how much HCA is due to being at home, and how much is due to the fact that the team with the better record gets more home games.

  6. Neil Paine Says:
    June 16th, 2010 at 12:04 pm

    That's why I suspect the extraneous "ability" variables (margin, srsdiff) weren't significant -- HCA is really important, but ability is also captured simply by knowing the Lakers had home-court in the series. The team who wins more regular-season games is usually the better team (ability-wise) anyway.

  7. P Middy Says:
    June 16th, 2010 at 12:25 pm

    If it's another blowout (which I highly doubt) this could go from the middle of the pack, to being one of the most disparate?

  8. downpuppy Says:
    June 16th, 2010 at 1:33 pm

    Only 17 out of 59 series have gone 7 games.

    If you look at the most similar series, this looks like 1969. How convenient that it's the only series where the road team was down on points & won.

  9. Bill Reynolds Says:
    June 16th, 2010 at 1:48 pm

    In the Finals, when the team that has scored fewer points in the series through 6 games is also the road team for game 7 (as is the case for Boston this year), that team has won game 7 only once in 8 tries. The one team to do it was the 1969 Celtics, who had only been outscored by 5 points total over the first 6 games. And they only won game 7 by two points.

    Great stuff, Neil.

  10. Bill Reynolds Says:
    June 16th, 2010 at 2:06 pm

    If you look at all series, including the Finals, the team in the Lakers' position (scored more points through first 6 games, have game 7 at home) has gone 44-8 in game 7. The C's can take hope from the 82 Sixers, who won game 7 in the Boston Garden despite being outscored by a whopping 47 points over the previous 6 games!

  11. AHL Says:
    June 16th, 2010 at 2:08 pm

    Err, just for simplicity, could you add a column for "Won Game 7 by this many points" or something? I guess "G7 Margin".

  12. tread Says:
    June 16th, 2010 at 3:42 pm

    I'd also like to see if Avg MoV @ Home (in the series) shows anything about how a game 7 plays out. In LA, the Lakers outscored Boston by 26 pts in 3 games (Avg = +8.66). Boston outscored LA by 6 points in the 3 games at the Garden (Avg = +2). Looks pretty good for the Lakers.

  13. Ben Says:
    June 16th, 2010 at 3:47 pm

    Thanks, Neil. Some nice additions in Bill Reynolds comments as well.

  14. marc Says:
    June 17th, 2010 at 1:24 pm

    Never underestimate to Celtics to come though in a clutch situation, Remember they where supposed to lose to Cleavland and Orlando, but here they are at game 7. Go Celtics!

  15. Coach Rowe Says:
    June 17th, 2010 at 2:06 pm

    Any way to try and figure out the total with SCHOENE type methods now that Kendrick Perkins is out.

    Any numbers that show game 7 to be higher or lower than average score or mean average of first 6 games?

«
»