Odinn21 wrote:sansterre wrote:SRS has plenty of faults. It really loves winning by big margins: there are several teams on my list that jumped 20+ positions because of one series where they obliterated another team by 15+ a game. And it has a really dim view of teams with low MoV over weak teams; a lot of good teams really seem to mail in those round 1 games, and SRS hates that (because, sometimes, really good teams lose those series, paging the 90s Sonics). So there are plenty of times where SRS will be in love with a team where I don't really agree.
And yet. Do you know who playoff SRS loves? The '96 Bulls. Also the '71 Bucks. Also the '86 Celtics. Also the '01 Lakers. And so on. While it may seem spurious, beating the tar out of other teams is something that all the teams we *know* were awesome did. SRS is what says that the '88 Lakers were a worse team than the '88 Pistons, even though the Lakers won the series.
Isn't it possible that we're conflating the success of the '89 and '90 Pistons with the '88 Pistons? The '90 Pistons posted a +10.48 playoff SRS (52nd on the list); the '89 Pistons posted a +13.16 playoff SRS (24th on the list). And, as you may infer, they're both ranked much higher than the '88 version. The reason that the '88 team is so much lower than the other two isn't because they didn't win the Finals (though that doesn't help), it's because by any objective standard they simply played worse. It's a natural thing to see how close the '88 Pistons got, then see how they played in '89 and think "I know this team is much better than that." It's understandable, but it doesn't make the '88 Pistons as good as the '89 Pistons; they weren't.
At the end of the day, my favorite thing about SRS is that it forces us to reconsider our assumptions. If an SRS-driven system thinks that the '88 Pistons were only the 97th best team of the shot clock era, maybe we were wrong. Maybe the '88 Pistons were only a really good team that used failure as the motivation necessary to become an all-time great team in '89.
But of course, your mileage may vary. This is just one attempt to make a comprehensive list that's as objective as possible, with all the warts that come with it.
I beg to differ cause of a test I ran awhile back.
In the three point era, the playoffs series after 1st rounds those went to a game 7 ended in favour of the team with negative MOV on overall 45% of the time.
So, your approach to use regular season SRS as the postseason SOS would correct some of it but not all.
Forgive me, but if I understand what you're saying, your position is "higher MoV through the first six games only predicts the winner of game 7 fifty-five percent of the time." It's one game - wouldn't we expect the predictive margin to be pretty small? I'd have expected larger than 55%, but for one game that isn't crazy.
Also, SRS entirely misses the quality loss from injuries. For example 1991 Lakers were pretty crippled in the NBA Finals. They were down Worthy and Scott in game 5, at which point the Bulls were already up 42 points in 4 games. Though I'll happily stipulate to the fact that injuries throw off SRS some. And this SRS-driven system would do nothing to account for it. Actually it looks like it'll boost 1991 Bulls due to MOV being bigger than what should've been under injury-free situation.
Heck, you gave the example yourself with 1989 Pistons. The Pistons swept the Lakers with almost 7 ppg MOV because Magic was not healthy in the NBA Finals. And then you turned to say "1989 Pistons were much better than 1988 version". Ah! I didn't know about it honestly. That said, it's not a great example in the specific, since Detroit won the first two games by 15 points when Magic was healthy, then won the second two games by 12 when Magic was injured. So I don't know how much of a swing that was.
No, they were not. I know that for a fact. I would be a fool to argue with what you know for a fact. The both versions were on the same level. One of them got a call (arguably) that shouldn't happen and one of them got not one but two teams with a unhealthy superstar to sweep (Boston and LA). I don't really get your point. The '89 Celtics were a +1.3 SRS team, so the Pistons don't get a ton of credit for crushing them. But they do get a ton of credit for blowing out the '89 Bucks (+4.6 SRS) by 12 points a game, and then another solid amount of credit for beating the improved '89 Bulls (+5.1 SRS) by 4 points per game. They may have had an edge against the Lakers on account of Magic's injury, but they were already winning before the injury. The foul call really doesn't enter into SRS at all, and the "unhealthy" stars in Boston and LA really didn't change much.
Another thing is, MOV and any +/- data have a different scale in different seasons.
1988 Pistons; 5.46 SRS (2nd), 6th in ORtg and 2nd in DRtg, 3rd in NRtg
1989 Pistons; 6.24 SRS (4th), 7th in ORtg and 3rd in DRtg, 4th in NRtg
So, looking at solely SRS would overlook the scale. Rankings should matter as well.
They do matter some; that's why standard deviations enter into my calculations. But scales can also be fairly misleading. The 1985 Celtics had the second best SRS of their year. By 0.01 points per game. The 1996 Sonics had the 2nd best SRS in the league, 4.4 points behind the Bulls. I don't know how persuasive rankings are, though I try to include that by adjusting for standard deviation from the league average.
1989 Pistons had one of the easiest routes to an NBA title and this SRS-driven system of yours doesn't make me reevaluate 1988 Pistons by looking how well 1989 Pistons did against a joke of a competition.
Did they? They faced a +1.3 SRS team, a +4.6 SRS team, a +5.1 SRS team and a +9.8 SRS team (+20.8 total). The '88 Pistons faced a -0.2 SRS team, a +3.5 SRS team, a +5.8 SRS team and a +6.3 SRS team (+15.4 total); these are my overall SRS figures. I'll grant that the '89 Lakers weren't quite as good as they looked on account of Magic's injury halfway through the series, but I still don't see how you see the '89 Pistons' opponents as a joke compared to the '88 Pistons' opponents.
Either way. I understand that you don't like SRS, and that's fine. I appreciate your reading and your contribution to the discussion.
