Bucks Diary

Tuesday, March 04, 2008

Historically, it still looks like the Celtics

Despite their success this season, very few experts seem to believe in the 2007-08 Boston Celtics. But, looking back at those teams who have won NBA Championships in the past, the Celtics look like a pretty solid bet to win one this year, barring some unforeseen injury or some odd collapse.

Extending my NBA Finals analysis past 1974

A couple of weeks ago I did a team and opponent "Win Score" analysis of every NBA Finalist from the modern statistical era, which dates to 1973-74. I couldn't go any further back than that because the NBA did not keep 3 necessary statistics until that season: turnovers, steals, and blocked shots.

However, fooling around a bit with the numbers, I quickly discovered that both turnovers and steals have been declining at an oddly steady and predictable "per minute" rate since 1974, and that blocked shots per minute have remained, during that same period, fairly constant. Realizing this, I calculated all the pertinent per minute numbers for every available season, plugged them into the linear graphing function featured on this University Physics Department website, and, after much trial and error (I'm not a mathematician of any sort) generated a regression formula that allowed me to speculate on the missing statistics for seasons prior to 1973-74.

Now granted, the statistical totals the formulas produced would never stand up to academic or legal scrutiny, but that doesn't mean they might not be pretty accurate. In fact, for amateur purposes, I think they are unusually accurate... or at least accurate enough to be useful.

Because even if the numbers are wrong, they have the virtue of being consistently wrong, which indicates they might be pretty close to right. In other words, of the teams I happened to analyze over several seasons, like the 60s Russell Celtics, the late 60s and early 70s Chamberlain/West Lakers, and the Bradley/Frazier Knicks, all of them show similarly structured results from year to year, which is what you would expect given their similar rosters across the seasons.

Moreover, when the results did diverge, they did so for explainable reasons (for instance, the Lakers numbers are consistently similar except that they change dramatically in 1969-70. That is a season in which Chamberlain only played 12 regular season games. Same thing for the 1971-72 Knicks, except the missing player was Willis Reed). So I think the numbers tell a believably accurate tale.

Anyway, enough with process. Here are the results in 3 parts, plus a repeat of this season's numbers.

Why the results point strongly to the Celtics

1. Defense as good as the Celtics almost always wins championships

The Boston Celtics "Opponent Win Score" for this season is 29% below the NBA average. With the exception of the 1993-94 New York Knickerbockers, every team that has held their opponents Win Score average at least 25% below the prevailing NBA average has won the NBA championship. And I consider the 1994 and 1995 NBA Finals a statistical Bermuda Triangle... I can't explain either of those Rocket championships. I can at least say that both of them were won against lopsided teams (the Knicks were all defense; the 1995 Orlando Magic were all offense)... something the Celtics are not -- see number 3 below.

2. Dominant defense trumps dominant offense

Even if the Celtics opponent should be the emerging offensive dynamo the Los Angeles Lakers, historically speaking, the Celtics should still be favored to win. Going back to 1962-63, -20% "Opp Win Score" defenses trump +20% "Team Win Score" offenses (see for examples, the 1969 Finals, the 1973 Finals, and the 1989 Finals).

3. The Celtics combine it with good offense

Not only do the Celtics play dominant defense, they also pair it with an offense that is 12% above average. That kind of +40% combination, if sustained, would almost certainly carry them to a championship. The only time it hasn't is twice.

First was the 1974 Bucks team, and that team was (a) not as dominant defensively as the Celtics; (b) lost Lucius Allen late in the season and suffered greatly for it offensively; and (c) came up against a Celtics team that was actually better defensively.

The other example was the 1997 Utah Jazz, and they were unfortunate enough to run up against another +40 team in the Chicago Bulls... and the same situation probably won't exist this season (though the Lakers or Pistons may change that).

Some other Random Historical Observations

1. Before 1974 (and in the late 80s) you could win with just defense

If you look at my results, the Russell Celtics and the 70s Knicks each won multiple titles with average and often below average offenses. In fact, stifling defenses carried the original Celtics dynasty to championships time and time again. If my numbers are right, the "Russell did it with defense" legend is absolutely correct. (Another thing I admire about the 60s Celtics is their obvious heart and "will to win". If you notice, they peaked in 1965, declined precipitously -- but won again -- in 1966, were dethroned by the awesome Chamberlain 76ers of 1967, and then somehow found a way to get off the mat and win two more championships with far lesser teams in 1968 and 1969. In that way, the Celtics dynasty follows an eerily similar trajectory and storyline to that of the 1960s Lombardi Packers.)

2. The 1971 Bucks were the third greatest champions ever

When I did my original analysis of NBA champions, I ranked each champion according to their Cumulative Win Scores. If I took my new numbers as read, and continued my rankings backward, the Bulls of the late 90s would still be number one, but the 1972 Lakers would be number two, and the 1971 Bucks would be a close number three. The Bucks were awesome that year, and they should have won more than one championship in the Kareem Era (the Celtics and Knicks won multiple championships with statistically far lesser teams). Situational misfortune (they played at the same time as two great champions -- the 1970 Knicks and the 1972 Lakers) and the aging nature of that team's rosters robbed Bucks fans of at least one more championship banner above the Bradley Center.


At March 4, 2008 at 11:48 AM, Anonymous Brian said...

Very true points, and great post... but you're basically describing both the Celtics *and* the Pistons.

At March 4, 2008 at 1:12 PM, Blogger TCW said...

You're right. They both have essentially the same makeup (Good Offense / Excellent Defense), except that the Celtics have it to a greater degree.

Another thing that makes it interesting is their numbers appear to be converging.

Still, if the numbers basically hold, history suggests the team with the superior numbers should win... though I can't say that with certainty since I've only looked at the Finals, and not every playoff series.


Post a Comment

Links to this post:

Create a Link

<< Home