http://www.basketball-reference.com/blog/?p=8159
Is this what you were looking at?
"In the NBA, dominating good teams is clearly the best indicator of postseason success."
Soap pointed it out.
In general, Final Four (postseason) success is best predicted by:
1. Blowing out Good teams
2. Blowing out Bad teams
3. Winning close vs. Good teams
4. Winning close vs. bad teams
'losing close vs. good teams' not really having much bearing compared to the above 4.
Does the article say anything at all about how records vs good or bad teams affects your ability to win the title?
Well, considering that in the course of an NBA season every team plays the same amount of games and overall ends up playing similar amounts of good vs. bad teams, # of wins vs. good or bad teams, which is what this is measuring, is a surrogate for record vs. good or bad teams.
Essentially the article is saying that the best predictor for playoff success is winning more games by a large margin against good teams. This makes sense; if you are winning big vs. good teams, then you're a good team and will have playoff success.
The crux of the article is the 2nd best predictor. Most people would assume winning small vs. good teams indicates future success. But not to the extent that winning big vs. bad teams does. Essentially, through the years, teams that win big vs. bad teams have more success than teams that win small vs. good teams.
I guess the point of the article is record vs. good or bad teams doesn't matter as much as winning big vs. anybody predicts conference finals/nba finals victories as compared to winning close vs. good teams.
But again, these are population trends. There are exceptions all the time. But what these do say, is that if you put money on every playoff game, and ignored everything (scouting, "clutchness," matchups, etc.) and just bet the team that had the most blowout victories (or largest average margin of victory), you'd end up with more money than you started.