In 2011 MLS Had A Small Positive Correlation Between Salary and Points

Discussion in 'MLS: News & Analysis' started by EvanJ, Oct 31, 2011.

  1. EvanJ

    EvanJ Member+

    Manchester United
    United States
    Mar 30, 2004
    Club:
    Manchester United FC
    Nat'l Team:
    United States
    http://www.soccermetricsblog.com/2011/10/mls-2011-payroll-vs-points.html is titled "How financially deterministic is MLS?" The regression line was 1.190 points per game played with 0 payroll (obviously you need to pay your players, but that's the constant term in the equation) + 0.033 points per game for every $1,000,000 spent. Based on that, the two clubs with by far the highest payrolls, New York and Los Angeles, have about an 11 point advantage over the rest of MLS for a 34 game season. New York and Los Angeles averaged 56.5 points this season and the other 16 clubs averaged 43.7, for a difference of 12.8.
     
  2. ojsgillt

    ojsgillt Member

    Feb 27, 2001
    Lee's Summit MO
    but you took the avg. between 2 playoff teams versus the avg between a 50-50 split of those who did make the playoffs and those who didn't
     
  3. VioletCrown

    VioletCrown Member

    FC Dallas
    United States
    Aug 30, 2000
    Austin, Texas
    Club:
    Austin Aztex
    Nat'l Team:
    United States
    It's not EvanJ. He's just linking to a blog post by Howard Hamilton. For the record. Comment on Howard's blog to get more direct feedback.:)
     
  4. EvanJ

    EvanJ Member+

    Manchester United
    United States
    Mar 30, 2004
    Club:
    Manchester United FC
    Nat'l Team:
    United States
    Actually the last sentence of my post came from me, not Howard. What ojsgillt said is true but I don't see how that's a problem with what I did.
     
  5. ojsgillt

    ojsgillt Member

    Feb 27, 2001
    Lee's Summit MO
    its just one of those ......."well, duh" statements. Taking the top point earner and a middle table team avg is going to be higher than the avg of all others.
     
  6. sidefootsitter

    sidefootsitter Member+

    Oct 14, 2004
    On paper, the league results came out as they should have.

    Attacking/playmaking DP's made the difference for LA. Defending/GK (not the best idea to spend the big money on the defenders in a low scoring sport, which is why most teams on a budget spend much heavier on forwards and why forwards generally make more money) for the NYC didn't.

    Otherwise, teams with a similar budget ended up knocking each other out (aka, taking points from each other) and allowed LA to raise its total above the MLS middle-class.
     
  7. LongDuckDong

    LongDuckDong Member+

    Jan 26, 2011
    Club:
    FC Schalke 04
    Nat'l Team:
    United States
    Well it can't be that "duh." The Krafts told Rev fans for years that spending isn't a factor, and teams that spend less are more successful.
     
  8. Allez RSL

    Allez RSL Member+

    Jun 20, 2007
    Home
    It isn't a factor. Here's the important phrase from that blog post:

    That is, sure, the regression fit a parameter value of 0.033 for the salary variable, but the standard error on that value is so large that you can't confidently say it isn't actually equal to zero.

    You don't even have to look at the fine print. One look at the data itself is enough to tell you this is worthless:

    [​IMG]
     
  9. superdave

    superdave Member+

    Jul 14, 1999
    VB, VA
    Club:
    DC United
    Nat'l Team:
    United States
    Which two teams had the most points? Seattle and LA. How many DPs, combined, do those two teams have? 5, right?
     
  10. Knave

    Knave Member+

    May 25, 1999
    Bingo. I've made a few of these graphs over the years, and the problem you end up with is always that the two outliers (NY and LA), which are the teams you're actually interested in (does their money = more points), are actually wildly extreme outliers compared to the rest of the data. The effect is that their presence has an exaggerated effect on the regression, which looking at the data points themselves is obvious.

    The entire appearance of a positive correlation between salary and points results from LA's presence in the data set. Remove that one outlier from the data set and you'll have a totally different regression -- and probably actually a negative correlation between salary and points.

    Statistically speaking, if you can remove one case and change the entire conclusion, the conclusion is bogus.
     
  11. Knave

    Knave Member+

    May 25, 1999
    Actually, I have the numbers, so I'll go one better. Take out one case (the Galaxy) and you get this:

    [​IMG]

    (Bigger)

    The removal of one case flips the direction of the correlation.

    Does money buy success in MLS? The only statistically valid answer is that it's still impossible to know because the data is inconclusive. But venturing beyond what's really valid, I'd have to say no it does not.
     
  12. Yoshou

    Yoshou Moderator
    Staff Member

    May 12, 2009
    Seattle
    Club:
    Seattle Sounders
    Nat'l Team:
    United States
    The bottom three teams on the table had 6 DPs between them.. What's your point?
     
  13. rslfanboy

    rslfanboy Member+

    Jul 24, 2007
    Section 26
    Yep. There's no information at all to be garnered from this. Over it.
     
  14. JazzyJ

    JazzyJ BigSoccer Supporter

    Jun 25, 2003
    Or you could take out both NY and LA, since they are arguably "outliers" - way above everyone else in terms of salary expenditure, stretch out the scale to accommodate the narrower range, and I think just by eye-balling, the line might still go negative or flat. But really I think it would just be that there is no line that could be reasonably drawn. There is no correlation.
     
  15. Knave

    Knave Member+

    May 25, 1999
    Sure. You could definitely make that argument. But I just wanted to demonstrate that the sole cause of the blog post's perceived positive correlation was LA.

    And, yes, it continues negative if you take out NY, and also if you take out both NY and TFC. It flips to positive only once you take out Vancouver.
     
  16. Yoshou

    Yoshou Moderator
    Staff Member

    May 12, 2009
    Seattle
    Club:
    Seattle Sounders
    Nat'l Team:
    United States
    So.. What you're saying is that if you take out all the high payroll teams that did poorly and the one that did well, you get a positive correlation? So, could the argument be made that payroll has a positive impact up until a certain point, then impact turns negative?
     
  17. troutseth

    troutseth Member+

    Feb 1, 2006
    Houston, TX
    As alluded to in the blog, I think the only conclusion is that spending does not have a statistically significant impact on success but the data set is too small to accurately draw long term conclusions. Quite frankly, with so many teams so tightly grouped on spending, you would need to see consistent success from most if not all of the higher spenders to prove a correlation over time.

    * As a side note, I do not know the complete source of the data but one thing to consider is the data is also LIKELY very limited. For example, if it uses the published salaries they do not take into account transfer fees paid. With such small salary budgets, those fees could represent a large percentage in any given year for any given team. Use Landin with Houston as an example, you needed to add a minimum of several hundred K for his fee on the annual salary number.
     
  18. Knave

    Knave Member+

    May 25, 1999
    If you take out LAG, the correlation is negative.

    If you take out the top 2 (LAG, NYR) or the top 3 salary teams (LAG, NYR, TFC) the correlation is negative.

    If you take out the top 4 salary teams (VAN, TFC, NYR, LAG), the correlation is positive.

    What I'm saying is that the data does not indicate any correlation between salary and performance.
     
  19. Mucky

    Mucky Member+

    Mar 30, 2009
    Manchester England
    Club:
    Manchester United FC
    What is the argument here?

    That high paid players are not better than low paid players or teams with higher salaries are no stronger than those with lower salaries?

    The only argument that makes any sense is that MLS wages are still at such a level that higher wages do not yet mean significantly better teams.
    $6 million is 50% of $12 million for sure but so is $25 million of $50 million and I am willing to bet that as wage bills increase but disparity in percentage terms stays much the same there will be higher correlation between wage bill and on field success.

    Then again one of the fun things about soccer is that the team is more important than the individual and the coach can make a difference too but it is impossible to put meaningful figures to all the variables.
     
  20. xbhaskarx

    xbhaskarx Member+

    Feb 13, 2010
    NorCal
    Club:
    San Jose Earthquakes
    Nat'l Team:
    United States
    Take out both LA and NY and the line would pretty much be vertical! :eek:
     
  21. superdave

    superdave Member+

    Jul 14, 1999
    VB, VA
    Club:
    DC United
    Nat'l Team:
    United States

    DPs don't guarantee you win, but not having them makes it very hard. The last 4 Open Cup winners all had at least 1 DP. The last several Shield winners had at least 1 (if you count Schelotto in 2008, which you should, because he was only bought down by allocations. He functioned as a DP as far as the cap, it was just the Hunts using league money rather than their money to pay the extra.) MLS Cups are a bit of a different story, but then, the last couple of winners have scraped into the playoffs and won a combined 3 PK shootouts (and 1 overtime.)
     
  22. Yoshou

    Yoshou Moderator
    Staff Member

    May 12, 2009
    Seattle
    Club:
    Seattle Sounders
    Nat'l Team:
    United States
    It's that having players with higher salaries do not make a team stronger. With how the MLS cap works having three players eat up roughly half a team's salary cap results in that team means having more lower salaried players to back up the higher paid players and those lower salaried players are generally of less quality than medium salaried players. So while the higher salaried players could be superior to other players, the trade off is having lower salaried players that are likely inferior to other players. End result is that the lower salaried players cancel out the benefit of the higher salaried players.
     
  23. barroldinho

    barroldinho Member+

    Man Utd and LA Galaxy
    England
    Aug 13, 2007
    Ex-pat in HB, CA
    Club:
    Manchester United FC
    Nat'l Team:
    England
    Is LA's salary the cause of their success, or does it's success and the coincidental factor of their salary just make it look that way?

    Consider this... LA Galaxy's worst season happened when they had the highest paid player and first DP (Beckham) plus a DP-in-waiting (Donovan, due to being Grandfathered).

    The following season, with Beckham missing on loan for much of the season and unable to be replaced during that time, they finished first in the west, second overall and lost the MLS Cup in a penalty shoot out.

    The season after that, they won the Supporters Shield, having one of their two DPs on loan and then injured until the final weeks of the regular season. The second DP was almost a non-factor after June, due to being at the World Cup and then being burnt out.

    In fact, at no point thus far, has a Galaxy player been the club's MVP while holding DP status.

    Based on this, I think there's a strong argument that LA is a strong outfit beyond it's DPs. As those players individual salaries grossly inflate the wage figures, yet the club has repeatedly shown the capacity to cope without them, there's a sound argument that the Galaxy's success has more to do with good coaching and squad composition than any financial element.
     
  24. Yoshou

    Yoshou Moderator
    Staff Member

    May 12, 2009
    Seattle
    Club:
    Seattle Sounders
    Nat'l Team:
    United States
    Except, the MLS Cup winner has never had a DP. The last or second to last team for the last five seasons have had at least one DP on their rosters. Having DP on a team does not make it "easier" to win and, in some cases, it actually makes it harder. What is important is the supporting cast around the DP.
     
  25. superdave

    superdave Member+

    Jul 14, 1999
    VB, VA
    Club:
    DC United
    Nat'l Team:
    United States
    The only reason Schelotto wasn't a DP the year the Crew won it was because the Crew used allocation money to buy down his cap hit. Under the exact same contract, he later WAS a designated player. The Crew just didn't have enough allocation money anymore.

    If the argument is about DP status, well, you got me. If the argument is about how DPs affect the salary cap and roster composition....

    To me, the interesting discussion is the relationship between the cap hit of DPs and winning, not the relationship between roster rules and winning.
    You're missing a season there. 2008.
    Um, I'm assuming you mean that Landon didn't win it last year or this year. But Landon actually WAS a DP, he just didn't count toward the DP limit in 2007, 2008, and 2009. In fact, during the 2007 season it was reported that the grandfathering was going to end after that season and the Gals would have to either trade Landon or acquire another DP slot. (Don't get me started on THAT piece of Galaxy favoritism, one of, IMO, the genuine examples of bias the league has had toward the favored franchises.)
     

Share This Page