FWIW: SRS Playoff data since 1990

Moderator: Doctor MJ

lorak
Head Coach
Posts: 6,317
And1: 2,231
Joined: Nov 23, 2009

Re: FWIW: SRS Playoff data since 1990 

Post#21 » by lorak » Thu Apr 24, 2014 5:21 am

sp6r=underrated wrote:
It is the exact same thing with the NBA playoffs. Each team has to play 4 rounds of basketball in the sample size. There is a minimum of 4 games per round since 2003*. The NBA doesn't force a team to play game 5 if they won games 1-4. Accordingly weighing game is irrelevant for who had a more difficult road.


It's extremely relevant. Look, "average opponent" (so SRS per round), your original question from opening post, is something entirely different than "difficulty of the road" (SRS per game). Maybe simple example will allow you to see it.

1987 Lakers in WC won 3-0, 4-1 and 4-0. Celtics in EC 3-0, 4-3 and 4-3. For the sake of argument lets assume both teams opponents SRS in each round was the same. According to your "method" both teams had equally difficult road to the finals. But that's obviously not true and Celtics had much tougher road.
User avatar
mopper8
Retired Mod
Retired Mod
Posts: 42,618
And1: 4,868
Joined: Jul 18, 2004
Location: Petting elephants with the coolest dude alive

Re: FWIW: SRS Playoff data since 1990 

Post#22 » by mopper8 » Thu Apr 24, 2014 8:13 pm

lorak wrote:
sp6r=underrated wrote:
It is the exact same thing with the NBA playoffs. Each team has to play 4 rounds of basketball in the sample size. There is a minimum of 4 games per round since 2003*. The NBA doesn't force a team to play game 5 if they won games 1-4. Accordingly weighing game is irrelevant for who had a more difficult road.


It's extremely relevant. Look, "average opponent" (so SRS per round), your original question from opening post, is something entirely different than "difficulty of the road" (SRS per game). Maybe simple example will allow you to see it.

1987 Lakers in WC won 3-0, 4-1 and 4-0. Celtics in EC 3-0, 4-3 and 4-3. For the sake of argument lets assume both teams opponents SRS in each round was the same. According to your "method" both teams had equally difficult road to the finals. But that's obviously not true and Celtics had much tougher road.


No, if the opponents were of the same SRS, then the Celtics did not have a tougher road, they just fared worse than the Lakers did on an equally difficult test.

By introducing wins/losses within a series into the calculation, your'e conflating team strength with opponent team strength.

Imagine for a moment that 1 team could play multiple playoff series. Miami's 1st round is against Charlotte, and they beat the Hornets 4-0 to advance. Then Toronto has its first found, and they also play Charlotte, and beat them 4-2 to advance. It would be ridiculous to say that Toronto had a "tougher road" to the second round than Miami because they had to play the Bobcats 6 times whereas Miami only had to play the Bobcats 4 times. They both played the Bobcats! It was only "tougher" for Toronto because Toronto is worse than Miami and thus couldn't dispatch with the competition as easily, but that doesn't mean Charlotte was somehow better or a harder team to beat all of a sudden once they started playing Toronto.

edit: another way to think about it is that if you and I both take the same exact exam, and you get 95% and I get a 90%, it would be absurd for me to claim (without something further) that I had a tougher test than you did.
DragicTime85 wrote:[Ric Bucher] has a tiny wiener and I can prove it.
lorak
Head Coach
Posts: 6,317
And1: 2,231
Joined: Nov 23, 2009

Re: FWIW: SRS Playoff data since 1990 

Post#23 » by lorak » Thu Apr 24, 2014 8:24 pm

mopper8 wrote:
By introducing wins/losses within a series into the calculation, your'e conflating team strength with opponent team strength.

Imagine for a moment that 1 team could play multiple playoff series. Miami's 1st round is against Charlotte, and they beat the Hornets 4-0 to advance. Then Toronto has its first found, and they also play Charlotte, and beat them 4-2 to advance. It would be ridiculous to say that Toronto had a "tougher road" to the second round than Miami because they had to play the Bobcats 6 times whereas Miami only had to play the Bobcats 4 times. They both played the Bobcats! It was only "tougher" for Toronto because Toronto is worse than Miami and thus couldn't dispatch with the competition as easily, but that doesn't mean Charlotte was somehow better or a harder team to beat all of a sudden once they started playing Toronto.


It does mean Charlotte was more difficult opponent for Toronto then for Miami and that's the point! If for example opponent is +3 SRS team then for weaker team (+5 SRS) it would be more difficult to defeat that +3 team than for a stronger team (+7 team), ergo +5SRS team's "road" would be tougher.
sp6r=underrated
RealGM
Posts: 17,197
And1: 8,517
Joined: Jan 20, 2007
 

Re: FWIW: SRS Playoff data since 1990 

Post#24 » by sp6r=underrated » Thu Apr 24, 2014 9:29 pm

mopper8 wrote:
lorak wrote:
sp6r=underrated wrote:
It is the exact same thing with the NBA playoffs. Each team has to play 4 rounds of basketball in the sample size. There is a minimum of 4 games per round since 2003*. The NBA doesn't force a team to play game 5 if they won games 1-4. Accordingly weighing game is irrelevant for who had a more difficult road.


It's extremely relevant. Look, "average opponent" (so SRS per round), your original question from opening post, is something entirely different than "difficulty of the road" (SRS per game). Maybe simple example will allow you to see it.

1987 Lakers in WC won 3-0, 4-1 and 4-0. Celtics in EC 3-0, 4-3 and 4-3. For the sake of argument lets assume both teams opponents SRS in each round was the same. According to your "method" both teams had equally difficult road to the finals. But that's obviously not true and Celtics had much tougher road.


No, if the opponents were of the same SRS, then the Celtics did not have a tougher road, they just fared worse than the Lakers did on an equally difficult test.

By introducing wins/losses within a series into the calculation, your'e conflating team strength with opponent team strength.

Imagine for a moment that 1 team could play multiple playoff series. Miami's 1st round is against Charlotte, and they beat the Hornets 4-0 to advance. Then Toronto has its first found, and they also play Charlotte, and beat them 4-2 to advance. It would be ridiculous to say that Toronto had a "tougher road" to the second round than Miami because they had to play the Bobcats 6 times whereas Miami only had to play the Bobcats 4 times. They both played the Bobcats! It was only "tougher" for Toronto because Toronto is worse than Miami and thus couldn't dispatch with the competition as easily, but that doesn't mean Charlotte was somehow better or a harder team to beat all of a sudden once they started playing Toronto.

edit: another way to think about it is that if you and I both take the same exact exam, and you get 95% and I get a 90%, it would be absurd for me to claim (without something further) that I had a tougher test than you did.


Thank you.

Sent from my SCH-I545 using RealGM Forums mobile app
mysticbb
Banned User
Posts: 8,205
And1: 713
Joined: May 28, 2007
Contact:
   

Re: FWIW: SRS Playoff data since 1990 

Post#25 » by mysticbb » Fri Apr 25, 2014 8:25 am

mopper8 wrote:edit: another way to think about it is that if you and I both take the same exact exam, and you get 95% and I get a 90%, it would be absurd for me to claim (without something further) that I had a tougher test than you did.


That is not even remotely close to the same. In fact, it entirely misses the point lorak makes. And lorak is correct here. The method presented by the op is actually punishing teams having to play tougher opponents, while allowing teams to get away with bad performances against weaker teams. Take a look at that example:

Code: Select all

SRS A   B
0   4   7
2   5   5
4   6   6


The first column is the SRS of the opponents. Team A needs 4 games against the 0 SRS team, Team B needs 7. Both have the same amount of games against the other opponents. Now, the method by the op portrays both schedules to be equal, while in reality team B played an average schedule of 1.9, and team A 2.3 (assuming no OT).

Yeah, you can play out the method the op presents, but it actually doesn't give you useful information. The weighted average SRS of the opponents on the other hand gives the value needed to adjust the scoring margin in order to learn how good the team really was. Anything else is just playing with numbers without a real purpose. Or other: What does it matter that the 1991 Bulls played a weaker schedule? They actually killed their opponents in 1991 by an average scoring margin of 11.7 in the playoffs. The 1991 Bulls would have also likely won the title, if the average opponents would have been much stronger (like +5 in average). The numbers in the op don't give you that kind of information. It is not important in the end what kind of opponents a team plays, but rather how good they played, and for that I want the weighted average SRS as well as the average scoring margin.
sp6r=underrated
RealGM
Posts: 17,197
And1: 8,517
Joined: Jan 20, 2007
 

Re: FWIW: SRS Playoff data since 1990 

Post#26 » by sp6r=underrated » Fri Apr 25, 2014 4:34 pm

mysticbb wrote:Yeah, you can play out the method the op presents, but it actually doesn't give you useful information. The weighted average SRS of the opponents on the other hand gives the value needed to adjust the scoring margin in order to learn how good the team really was. Anything else is just playing with numbers without a real purpose. Or other: What does it matter that the 1991 Bulls played a weaker schedule? They actually killed their opponents in 1991 by an average scoring margin of 11.7 in the playoffs. The 1991 Bulls would have also likely won the title, if the average opponents would have been much stronger (like +5 in average). The numbers in the op don't give you that kind of information. It is not important in the end what kind of opponents a team plays, but rather how good they played, and for that I want the weighted average SRS as well as the average scoring margin.


I never said I was answering the question: Who played the best in the post-season since 1990? If I did your criticism of my method would be correct. To answer that question you would have to factor in games played, point differential along with opponent strength. That wasn't my question.

I was quite surprised to see how brutal of a potential road to the finals the Clippers faced with getting GSW in the first round potential matchups with OKC and SAS. My impression was that a historically difficult road to the finals. I they were extremely unlucky when you factor in that they won 57 games and came into the post-season as a 3rd seed rather than a 7th seed.

I ran the numbers to see if I was correct. After running the numbers my interest shifted to average competition rather than the quality of the competition relative to seed. I may at some point return to my original interest. The NBA only requires in the relevant sample that you win 3 rounds to get to the finals so games played aren't of interest to my question*. I posted the numbers and created a few lists because I figured other people would be curious about that information.

The idea that the only question to ask is "who played the best" is extremely myopic. Most people who are interested in basketball also are curious about other things surrounding the sport. There are threads on this website about where would Shaq rank if he shot 90% from the free throw line or who has the best cross-over. I really don't care about it and don't post in those threads. Now I may chime in if people go from stating who had the best cross-over to listing that as the deciding factor who was the best offensive player.

That wasn't the case here. Nobody posted something similar to the 95 Rockets played far better teams than the 91 Bulls and as a result 95 Rockets > 91 Bulls. So I'm not sure what your complaint is Mystic. People have interests about basketball beyond your own interests. This is a forum to facilitate conversation. Not all conversations will be of interest to you.

*pre 2003 1st round may be a different story

That is not even remotely close to the same. In fact, it entirely misses the point lorak makes. And lorak is correct here.


Lorak is wrong. His disagreement with my OP isn't that I was asking the wrong question and that my topic is pointless which is your objection but rather that my method was wrong. As others have explained, GP isn't relevant for my question.
mysticbb
Banned User
Posts: 8,205
And1: 713
Joined: May 28, 2007
Contact:
   

Re: FWIW: SRS Playoff data since 1990 

Post#27 » by mysticbb » Fri Apr 25, 2014 5:07 pm

sp6r=underrated wrote:So I'm not sure what your complaint is Mystic.


I can imagine, because my post did not contain the complaint specifically. I was referring to lorak's point that "average opponent" and "difficulty of the road" are different things. I haven't made that clear in my post, so, sorry for that.

I understand what you are trying to do, and it makes sense to do that for average opponents strength of a bigger sample (like your comparison between EC and WC), but I have a hard time seeing that it is useful for specific teams in specific seasons.
lorak
Head Coach
Posts: 6,317
And1: 2,231
Joined: Nov 23, 2009

Re: FWIW: SRS Playoff data since 1990 

Post#28 » by lorak » Sat Apr 26, 2014 12:55 pm

:D
Chris_SoCal
Sophomore
Posts: 223
And1: 171
Joined: Jun 22, 2014
 

Re: FWIW: SRS Playoff data since 1990 

Post#29 » by Chris_SoCal » Mon Jun 23, 2014 2:08 am

mopper8 wrote:
lorak wrote:
sp6r=underrated wrote:

edit: another way to think about it is that if you and I both take the same exact exam, and you get 95% and I get a 90%, it would be absurd for me to claim (without something further) that I had a tougher test than you did.


That is not a very accurate way of looking at it. That would only work if the playoffs where single game/single elimination style tournament.



If you want something comparable it would be like, Yes your friend passes on the first go around but your teacher has high standards so notes some corrections and adds a question or two for you to bring back tomorrow. Bringing it back the next day he makes some more correction and adds one or two more questions just to make sure you got it. You bring it back the third day and you finally get your 95%. While I wouldn't want to penalise the first person and it was your own fault... you still would have to say you had a much harder test than your friend.


We need two statistics, one that is simply strength of schedule, then we need a "difficulty of path to the finals stat" that would include how many games it took you.

Return to Statistical Analysis