Samstag, 14. Juni 2014

Uruguay vs Costa Rica: Goalimpacts of Lineups


PlayerGoalimpactPeak GIAgeTeamNo. GamesNo. Minutes
Fernando Muslera109.1109.128.0Galatasaray26624768
Diego Lugano107.1137.233.6West Bromwich Albion30427372
Diego Godín123.1128.328.3Atletico Madrid33830431
Maxi Pereira136.1142.430.0SL Benfica31827198
Martín Cáceres109.5112.427.2Juventus24620588
Walter Gargano110.1116.229.8Parma FC29423869
Cristian Rodríguez107.7113.128.7Atletico Madrid32318173
Egidio Arévalo87.6112.432.4Monarcas Morelia20417574
Diego Forlán86.2129.335.0Cerezo Osaka55640703
Christian Stuani97.9102.027.7Espanyol Barcelona20613634
Edinson Cavani122.1125.327.3Paris Saint-Germain34628093
Rodrigo Muñoz123.8123.832.3Libertad20919381
Martín Silva108.0108.031.2Olimpia Asuncion21419831
Jorge Fucile128.8134.729.5FC Porto16914320
José Giménez85.4124.119.3Atletico Madrid282410
Sebastián Coates118.4126.723.7Nacional12310660
Álvaro Pereira121.6126.928.5Sao Paulo FC29024489
Nicolás Lodeiro113.8116.625.2Botafogo RJ1328804
Diego Pérez82.7115.234.0Bologna FC34024965
Gastón Ramírez106.8115.723.5Southampton FC1519591
Álvaro González109.5115.429.6Lazio Roma26617885
Abel Hernández103.2110.823.8US Palermo1378107
Luis Suárez144.5147.827.3Liverpool FC37732230

Costa Rica

PlayerGoalimpactPeak GIAgeTeamNo. GamesNo. Minutes
Keylor Navas96.996.927.5Levante UD14713579
Giancarlo González103.2103.926.3Columbus Crew736435
Michael Umaña85.4107.331.9CD Chivas918059
Óscar Duarte106.0109.525.0Club Brugge KV483976
Júnior Díaz107.9120.130.81. FSV Mainz 0519815776
Cristian Gamboa112.5117.324.6Rosenborg BK988692
Celso Borges110.4111.325.8AIK Solna20117668
Christian Bolaños101.5108.130.0FC Köbenhavn24918799
Yeltsin Tejeda98.4112.822.2131001
Joel Campbell110.6126.521.9Olympiakos Piräus1217725
Bryan Ruiz115.6121.028.8PSV Eindhoven29723106
Patrick Pemberton100.8100.832.1LD Alajuelense201860
Daniel Cambronero93.193.128.4CS Herediano9837
Johnny Acosta100.9114.430.8LD Alajuelense292728
Dave Myrie101.2101.226.0CS Herediano8641
Waylon Francis100.2108.323.7Columbus Crew191733
Roy Miller106.1112.029.5New York RB17013788
Michael Barrantes101.1112.830.7Aalesunds FK13811310
Esteban Granados99.4104.828.6CS Herediano11777
Diego Calvo93.5103.823.2Valerenga IF452722
José Cubero101.4104.727.3CS Herediano251836
Randall Brenes92.0104.930.8CS Cartagines583087
Marcos Ureña103.4109.424.3Kuban Krasnodar441950


  1. Interesting to see now with the averages: both Colombia and Urugay had a better bench than the starting 11! With the former it didnt matter, although now Greece was even a GI favourite. With the latter maybe it played a role in the defeat.

    Anyway, I would assume slight GI differences would mean pretty even success probabilities. Interesting to see that the guys from spielverlagerung have pretty strong - albeit elaborated - opinions about success rates. For example they didnt give Greece any chance - and were right so far against the GI. But wrong in Uruguay vs Costa Rica, where GI was closer.

    Excited for the validation of the various predictions after the tourney with a bit (still small) more sizable database than these two games.

    1. Ok, now I see the low starting 11 GI was influenced by the low GIs of older players, particularly with Colombia´s Mario Yepes at a staggering 38 at the age of 38. How come GI´s age curve takes a free fall at these ages? Aren´t there yet any counter examples of still good, old players in the database who would shape the curve differently?

    2. The forecasts were pretty much spot on so far as I can see. I'll post a backtest of different predictions at the end of the tournament. Current standings

      Spielverlagerung > Goalimpact > BSports > Goldman Sachs

    3. I think the key disadvantage of GI is that it is a player rating system. However, the team success is driven by more factors than just the players (e.g. tactical system) and hence GI is missing a part of the story.

  2. Btw, where do u get the Spielverlagerung predictions? I was only referring to the ebook they published, but in there are only rough predictions about the employed tactic, no match results.

    1. @nobilor did some great work to extract that into a prediction.

    2. Here is the link