FanPost

Selected DC United Numbers for Game 5 vs Metros + a call for help

Let us examine the DC United game using soccer statistics. I'll be using selective statistics that I find potentially very interesting largely taken from the chapters of The Number Game: why everything you know about soccer is wrong by Chris Anderson & David Sally and Soccernomics by Simon Kuper & Stefan Szymanski.


I know not all statistics are useful for soccer. But empirical evidence shows selected statistics are predictive of results, goals, and player value. I will entertain the notion of doing more complicated analyses or examining additional statistics provided the data are available. I am using data from WhoSocred.com and MLSSoccer.com.

Hopefully these numbers provide useful game insight and predictive utility.


The Weakest Link

Soccer is a team game. According to Anderson and Sally the worst player is more determinant of a team’s performance than the best player; hence my focus on the weakest link in a chain. I’ll be using player index ratings from WhoScored*. I am not thrilled about using indices where the calculation details are vague but those two are both established and I don’t see enough available data available to calculate my own. *An explanation for WhoScored’s ratings can be found here **I’ll look to incorporate the Castrol ratings in the future. They just come out too late for these game-by-game postings and they don’t break down by game.


player

vs. NJ 4-12

vs. NE 4-5

vs. CHI 3-29

vs. TOR 3-22

vs. CLB 3-8

Davy Arnaud

7.2

7.1

7

6.06

6.1

Bobby Boswell

8.4

7.3

7.5

6.64

6.85

Alex Caskey

6.1

DNP

DNP

DNP

DNP

Nick DeLeon

6.9

7.1

7.7

6.69

6.35

Conor Doyle

6

6

6.2

5.99

5.92

Andrew Dykstra

7.8

6.9

DNP

DNP

DNP

Fabian Espindola

7.3

7.8

8

5.88

5.68

Cristian Fernandez

7.1

7.3

7

6.1

6.98

Sean Franklin

6.9

7

6.8

6.67

7.05

Bill Hamid

DNP

DNP

6.8

7.17

5.8

Jared Jeffrey

6.1

DNP

5.9

DNP

DNP

Eddie Johnson

7.3

6.6

7.4

5.94

6.14

Perry Kitchen

6.7

6.8

8

6.79

5.8

Lewis Neal

DNP

6.5

6

DNP

5.9

Jeff Parke

7.3

8.6

7.2

6.63

6.3

Kyle Porter

DNP

DNP

DNP

6.04

5.93

Chris Rolfe

7.1

7

6*

N/A*

6.37*

Luis Silva

DNP

DNP

6.7

6.59

6.62

*Chris Rolfe's first three ratings (including the DNP) are from his performances with Chicago.

Inferences

Just like last in game 4 a central defender has the best player rating. This time it is Bobby Boswell. Undoubtedly, the assist helps him quite a bit since not other DC player is in the 8’s. Kitchen and the most of the midfield rated in the mid 6’s which is not good. Arnaud and Rolfe rated well as did Espindola and Johnson. Dykstra really picked it up. Probably because he faced many more shots this week.

In comparison New Jersey only had two players rated at or above 7.0 (Henry and Olave- I think Henry got at least .5 for he consistent applauding his teammates for trying but clearly fucking up). New Jersey averaged 6.50 while DC united averaged 7.0, slightly below their previous game. This was not improvement in player ratings over their last game but was the third consecutive time the team averaged better than the opposition..


KEEP THE BALL!

Controlling the game and how that control is achieved is quite predictive of league points. The overall possession statistic is, as we know, not useful for determining what happened. However, having possession is generally important. As a consequence turnovers are indicate how well a team controlled the game. Statistically, the turnover ratio (# DCU turnovers/ (#DCU turnovers + #Opp. turnovers) ) is a good number to examine. Passing data is informative. The sheer raw number of passes (more is better) is basic number to compare the two teams. The type of passes (Longball to Short Ratio LSR = Longballs/(longballs + shortballs); lower number is better) and passing completion percentage correlate (not cause) with the overall number of points a team earns.


NEW STAT ADDED. Through ball.I’m getting the number from Whoscored.com. It reflects an attempted pass from one player to another that goes through the opponents defensive line. I’m adding this statistics because the offense is still a concern for this team and so many of the other statistics I examine are about the long term.





Possession: 43% (black part of the pie, red being New Jersey 57%)
Total passes: 368 vs 487
LSR = .20
Turnovers: 43%
Through Balls = 1


Inferences

DC United did well on its turnovers and it’s through balls (NJ had 0). And in terms of ball possession and management that is it. The metro’s had much more of the ball and far more passes. In comparison to the New England game DC did worse. In fact DCU had some of the their worst possession numbers and passes of the season so far. Their 43% possession ties the Columbus match and the 368 passes is their lowest amount of passes per game by 50 (also the Columbus match). The one good ball possession number here was the low amount of turnovers. DC United’s season low, both in terms of raw numbers and the percentage above.

*I would also point out that the above statistics are correlated with league points earned but that is for an entire season (with European data too - may not apply in MLS) and not necessarily predictive of individual games.


SHOOT THE BALL!

Scoring two goals is the best predictor of winning points. Scoring a third goal only slightly improves a team’s odds above two goals of earning a draw or win. Since, goals were so rare for DC United last year I’m also looking at what predicts goals. An average of 9 shots per game is what it takes to score a goal. Can DC United manage 9 shots per game? A new (maybe just popular) statistic that predicts points is the Total Shots Ratio TSR = Total shots for/(Total shots for + total shots against). TSR is an indirect indicator (as a latent or an emergent variable - I’m not sure) assessment of team strength. The best English clubs have TSR’s around .60.


NEW STAT ADDED. Big Chances. I’m taking this statistic from mlssoccer.com. The statistic reflect the number of passes that lead to a clear cut scoring chance (usually something inside the box). I am incorporating this stat. for the same reason I added the Through Balls statistic (above).


DC United:  12 Shots (2 on goal)
BIG CHANCES: 0
1 Goals
TSR = .39


Inferences

Their total shots improved from the New England game (still below the Chicago number) and is about the same as the Columbus game. Unlike the New England game New Jersey had a ton of shots (19. 3 on goal - DC had 5 on goal). The .39 TSR is bad but not as bad as the season low (.30) against Columbus.And the shots number dropped by 8 from last week. DCU had 0 big chances this week (NJ had 0 - according to MLSsoccer.com). However, here’s the only statistic that matters and the only one DC United dominated in: goals.


Change We Can Believe In

When behind the best times to substitute a player are first by the 58th, second by 73rd, and finally the 79th minute. MLS and national team coaches adhere to those substitution times more than any other league coach in the world. As a consequence of substituting by those times MLS coaches salvage or gain more points (i.e. score game tying or winning goals) than other teams across the world. The weakest link also re-appears here as the weakest link should be the first player substituted regardless of position.

Subs came in at 66th, 72nd, 84th.

Inferences

Once again the substitute were not impactful. But they were proactive. By the numbers NJ should have subbed earlier.


Summary

DC United did not win this game statistically. But, DC United did win the game. Again. That’s two weeks in a row. Last week it was an evenly matched game in terms of the above selected numbers. This week it was not evenly matched. Many of these selected numbers are for predicting overall performance, and DC isn’t look good. Their giving up a lot of shots and not taking that many. Recently, however, they are scoring goals and holding on. We’ve seen that before. This was not a performance improvement from last week or the week before. In fact it looks similar to the Columbus game overall. Except of course for the one all important statistic. New Jersey should have won this game. Yet, 3 points. Again.


Next week I’ll do a comparison versus the entire league on most of the selected numbers. Also, week 6 could be a good week for predicting playoff qualification (more on that next week) but first:


Call For Help

I need help finding mls weekly results. I want to see if I can predict playoff qualification based on weekly results (using a ROC Curve from SDT). My problem is I cannot find the data. MLS soccer.com and Wikipedia only has a few years worth of data. Specifically, I need the results to look like from this page or this Results Table from 2009 on wikipedia would be ideal. Does anyone know where I can find this data?

X
Log In Sign Up

forgot?
Log In Sign Up

Please choose a new SB Nation username and password

As part of the new SB Nation launch, prior users will need to choose a permanent username, along with a new password.

Your username will be used to login to SB Nation going forward.

I already have a Vox Media account!

Verify Vox Media account

Please login to your Vox Media account. This account will be linked to your previously existing Eater account.

Please choose a new SB Nation username and password

As part of the new SB Nation launch, prior MT authors will need to choose a new username and password.

Your username will be used to login to SB Nation going forward.

Forgot password?

We'll email you a reset link.

If you signed up using a 3rd party account like Facebook or Twitter, please login with it instead.

Forgot password?

Try another email?

Almost done,

By becoming a registered user, you are also agreeing to our Terms and confirming that you have read our Privacy Policy.

Join Black And Red United

You must be a member of Black And Red United to participate.

We have our own Community Guidelines at Black And Red United. You should read them.

Join Black And Red United

You must be a member of Black And Red United to participate.

We have our own Community Guidelines at Black And Red United. You should read them.

Spinner.vc97ec6e

Authenticating

Great!

Choose an available username to complete sign up.

In order to provide our users with a better overall experience, we ask for more information from Facebook when using it to login so that we can learn more about our audience and provide you with the best possible experience. We do not store specific user data and the sharing of it is not required to login with Facebook.

tracking_pixel_9353_tracker