Demystifying SOS: MOV for top 4 teams against Winning Record Teams

[spursfaninla]

Many claim that GS is the better team, and according to them, it is not even close.

Proof: Scoring efficiency through the roof, record, MOV, and the best player (almost all advanced metrics support).

What can we use as an objective measure of Spurs chances/success?

Margin of Victory is the best predictor of current strength, and the granddaddy of advanced stats, Hollinger, argues it is the best predictor we have for success in the post-season and for odds for championships. (lets put aside health and matchups, which may confound this).

MOV To Date top 4 teams:
GS 11.8
Spurs: 13.5
OKC 8.4
Cleveland: 5.3

The above MOV is largely why ESPN predictor today gives spurs 43% of winning the championship, and GS 40%.

However, unless MOV is adjusted for SOS, it may merely reflect a good team beating up bad ones. At the end of the year this tends to event out, but if some teams have player substantially harder schedules mid-stream, and everyone knows the spurs SOS has been the easiest of any team in the League.

Lets see if the SOS argument holds up.

To get the most data, I compared top 4 team records against all teams with a current winning record.

In summary: Spurs are tied for best MOV against winning teams.

They played the fewest of such games, but are within 1 game of the # played by GS and CL, who are within 2 games of OKC.

Record if score with 6 points (imo, stats guys seem to think this usually means luck)

Spurs: 2-3
GS: 5-0

Note: of spurs 3 losses to these opponents, all 3 are by 6 or fewer points.

Top4 teams v. winning teams

GS 9-1
Spurs 6-3
OKC 7-5
CL 6-4


Average MOV against winners:
GS 4.3
Spurs 7.1
OKC 6.1
CL 7.1
ESPN playoff predictor and ESPN Basketball Power Index both place Spurs as stronger than GS and/or more likely champions. These tools use MOV, but are also adjusted by SOS.

Bottom line, I don't claim that Spurs are necessarily better than GS, just that GS is not "clearly better". It is probably a toss-up today. Of course, everything could change in 3-4 months.

[Cry Havoc]

MOV really isn't relevant to the Spurs and Dubs. They are both ending games by the 4th quarter, sometimes it's over at halftime.

Does MOV have any value when 8-15 minutes of the game are complete garbage time? I think it takes a lot of the edge off, because the Spurs and Warriors are shaping up to be historically dominant teams. A lot of the standard metrics just won't apply to them.

[SAGirl]

I think you point an important aspect of our 3 losses to good opponents, they were winnable games but execution down the stretch was an issue and other has been much whining here about it with some good reason. This team is nit yet all it can be bc they need to figure out those issues.

[Mr Bones]

I think MOV is a good general guide as a stat, but it isn't an exact science at all because of variables like playing time in blow outs, strength of schedule, injuries, coaching experiments, etc... Teams with a 10.5 MOV are going to be consistently better than teams with a 1.2 MOV (especially as the season rolls on and the sample size grows), but for me MOVs of 8,9,10,11, or 12 are all about the same.

[spursfaninla]

MOV really isn't relevant to the Spurs and Dubs. They are both ending games by the 4th quarter, sometimes it's over at halftime.

Does MOV have any value when 8-15 minutes of the game are complete garbage time? I think it takes a lot of the edge off, because the Spurs and Warriors are shaping up to be historically dominant teams. A lot of the standard metrics just won't apply to them.

Well, to your point, arguably mov >10 gets some noise because it is a blowout and our 2nd or 3rd string are playing, so true strength is no longer represented, so both SA and GS you may not get accurate data on true strength.

However, I think you are missing the point of checking MOV against winning opponents, because those games are averaging a MOV less than 10. Thus, is is more likely to have less blowouts and to have more of our regular rotation in the whole game. Therefore, these games are better at representing accurate portrait of our (and other top 4) strength.

[spursfaninla]

Dude, that is the point of my thread.

I admitted SOS is the common counter to MOV, but when you reduce the sample MOV vs. winning teams, you get more meaningful data.

The result: GS and Spurs are about as successful against winning teams; GS slightly better at winning, Spurs better at MOV vs. those teams.

[spursfaninla]

I think you point an important aspect of our 3 losses to good opponents, they were winnable games but execution down the stretch was an issue and other has been much whining here about it with some good reason. This team is nit yet all it can be bc they need to figure out those issues.

Nope. Winning close games is basically luck was my point. Unless you have a large enough sample, it is arguably just luck that we are attaching meaningfulness to (clutch).

http://bleacherreport.com/articles/374519-the-clutch-myth-and-why-we-buy-into-it

http://www.theatlantic.com/entertainment/archive/2012/06/why-clutch-is-overrated/258179/

http://research.sabr.org/journals/the-statistical-mirage-of-clutch-hitting

http://bleacherreport.com/articles/923262-debunking-the-myth-of-clutch-in-the-nba-once-and-for-all

https://www.psychologytoday.com/blog/extreme-fear/201006/the-science-sports-is-there-such-thing-clutch-performer

"In Slate (link is external) today, writer Alan Siegal poses the burning question: "Is Kobe Bryant really the best clutch player in the NBA?" That is to say, does Bryant possess that ineffable quality, so highly prized among athletes, of being able to respond to the highest degree of pressure by pulling out the stops and performing at an even higher level of performance than usual? Which, as Siegel acknowledges, raises a corollary question: does such a quality even exist? A growing consensus among sports statisticians is that the answer is no, as attempts to identify clutch players based on their average performance under certain high-stress conditions (the last shot of a game, say) have so far come to naught. "

[SAGirl]

It's not luck if you don't execute well, have chemistry issues, don't get the ball to you best player, etc. Luck is whether the ball goes in or not and that is more just probability than luck.

[spursfaninla]

If the Spurs did not execute well, had chemistry issues, and could not get the ball to their best players, how are they winning so often and by such a large margin?

I don't want to discount that end of games are different in some ways:
teams player tighter defense,
(all) players are under more pressure and tend to perform worse
defense is set so they can game plan in the half court about denying your best player the ball, etc.

Even so, we won 2 of 5 of the close games against winning teams. That is almost half, and the sample size is too small to be significant anyway, but if we take it at face value, it completely supports my luck argument; close games are a coin flip, and good teams make sure they have less close games. Teams that rely on being "clutch" for their wins have an inflated record.

[SAGirl]

We are talking about close games. It's an entirely different issue in games decided by large margin and even in those games they didn't necessarily execute consistently well, just well enough that combined with their awesome defense they were winning. Then you have some garbage time crew guys that would frankly be rotation players for other teams and often they expanded leads in garbage time.

It's a different issue in close games. If you don't think there is a way to affect the probability that you win close games with good execution, what is the point if coaching? . Yea sometime you get the best shot you could have possibly gotten given circumstances and it doesn't go in, but I don't think in close games we have exactly been executing well.

[spursfaninla]

We are talking about close games.

Yes, and I said we have won 2 of 5 against winning teams. Its not like we won none of them.



It's an entirely different issue in games decided by large margin and even in those games they didn't necessarily execute consistently well, just well enough that combined with their awesome defense they were winning. Then you have some garbage time crew guys that would frankly be rotation players for other teams and often they expanded leads in garbage time.



My contention is that there is some luck when it comes down to a handful of plays to determine if you win or lose. Good teams make sure it does not come down to that often, and that is why they have better records.

Essentially, my contention is this: we put more value in winning close games that we should; we should instead rely on MOV, and wins based on "stomping" your opponent, to determine championship chances. The following study supports my argument:

The Miami Heat have been under fire recently for dominating weak foes and losing close games against good teams. The conventional wisdom is that this reveals a major gap in Miami's armor -- they just can't close the deal against stronger opponents.

Is this really true, though? And does it even matter?

http://www.basketball-reference.com/blog/?p=8159

In 2005, our friend Aaron Schatz wrote a piece for Football Outsiders called "Guts and Stomps", wherein he tested the same phenomenon in NFL football. Guts are close wins against good teams; stomps are blowouts of bad teams. As it turned out, the team with more regular-season "stomps" (that's big wins vs. bad teams) tended to win the Super Bowl and the conference title game more often than the team with more "guts", the team with more "skates" (close wins over bad teams), and even the team with more dominations (big wins vs. good teams). In other words, being like the Heat -- winning big over bad teams and not being able to close the deal against good opponents -- was actually predictive of Super Bowl success!

In the NBA, dominating good teams is clearly the best indicator of postseason success. Teams that had more regular-season dominations (big wins over good teams) won 64.8% of their "final four" series, including 73.3% of their Finals matchups. But the second-most predictive attribute of "final four" success was having more stomps -- that is, destroying the league's weaker teams. And having more stomps was actually a better indicator of success than having more guts (close wins against good teams), just like Schatz found in football.



It's a different issue in close games. If you don't think there is a way to affect the probability that you win close games with good execution, what is the point if coaching? . Yea sometime you get the best shot you could have possibly gotten given circumstances and it doesn't go in, but I don't think in close games we have exactly been executing well.

So, basically I now think the data supports that unclutch player/team is 1) not predictive of success, or lack of success, 2) is more indicative of luck (even though we like the idea of being clutch), and 3) is a symptom of falsely attributing meaning to what we see as a pattern of events (like the full moon effect).

[SAGirl]

My contention is that there is some luck when it comes down to a handful of plays to determine if you win or lose. Good teams make sure it does not come down to that often, and that is why they have better records.

Essentially, my contention is this: we put more value in winning close games that we should; we should instead rely on MOV, and wins based on "stomping" your opponent, to determine championship chances. The following study supports my argument:

The Miami Heat have been under fire recently for dominating weak foes and losing close games against good teams. The conventional wisdom is that this reveals a major gap in Miami's armor -- they just can't close the deal against stronger opponents.

Well you have a good point overall bc defense wins you championships and stomping over the weak/mediocre teams is a sign of good defense. You have a better chance in the postseason if you have relied on your defense all through the season to come out ahead rather than hoping for a Marco Belinelli 3 to save you (see what I did there? props to TD21, Raven, ElNono, etc.)

I do not use that as some final barometer overall bc I think we can still get better. You can't win all close games as there is an element of luck involved yes, (although GSW has their execution down pat). I just think overall, we don't have our own execution down pat in close games, and we didn't get the best chance to win some of the close games we lost precisely because of that. There are still some hierarchy issues with the Big 3 and LMA/Kawhi. They have at different times come through. All of them in different games, but they have also bumped up each other, or left a guy to take a shot who was not the best option, or put some guy in a position to take a shot that was probably the worst possible one. There are things like that you can look at in the games we lost, which should give us hope since we can improve and its not all.. oh yea, just LMA, Kawhi, Manu, Tim or Tony missed a shot, tough luck.