Leskošek B., Čuk I., Bučar Pajek M. TRENDS IN E AND D SCORES AND THEIR INFLUENCE... Vol. 5 Issue 1: 29 - 38 TRENDS IN E AND D SCORES AND THEIR INFLUENCE ON FINAL RESULTS OF MALE GYMNASTS AT EUROPEAN CHAMPIONSHIPS 2005-2011 Bojan Leskosek, Ivan Cuk & Maja Bucar Pajek Faculty of Sport, University of Ljubljana, Slovenija Original research article Abstract The aim of the study was to investigate the trends in execution and difficulty scores of routines on all apparatus and in both qualification and final rounds of male European championships just before and during a 5-year period after the introduction of new "open-ended" Code of Points (CoP) in 2006. It was found that the new CoP solved the problem of invariant difficulty scores, most efficiently toward the end of the observed period (2011). Execution scores showed a clear decreasing trend, both in absolute value and also in it's ratio with difficulty score. A question arises, if the decreasing influence of execution on final score (and therefore ranking of competitors) is the desired outcome of the new CoP and future evaluation of gymnastic routines. It was also questioned if the decrease in execution score should be solely explained by the increase of difficulty (which probably means more deductions) and some minor changes in CoP after year 2006, or it showed (possibly unjustified) changes in applying the CoP. Keywords: Artistic gymnastics, European Championship, Males, Judging, E-score, D-score INTRODUCTION As in several other "esthetic" sports, e.g. figure skating, diving, rhythmic gymnastics, and synchronized swimming, a score given to a competitor in artistic gymnastics also depends both on difficulty and the execution (flawlessness) of a routine. What part of the final score is given to each of these two components depends on sport rules (Čuk, Fink & Leskošek, 2012). In artistic gymnastics a major revision of Federation Internationale de Gymnastique (FIG) Code of Points (CoP) occurred in year 2006. Old CoP (FIG, 1997; 2001) follow a "perfect 10" format, where a start value (SV) of 10 was the highest possible score of a routine, given to a gymnast with a "hard enough" routine with all required elements and no deductions in execution. A new open-ended format was introduced in year 2006 and revised in year 2009 CoP (FIG, 2009). The final jury score is a sum of difficulty (D) score and execution (E) score, each of which is given by different panel of judges. The D score starts at 0 points and has no upper limit. Except for vault, where each element has it's own difficulty value (until 2011 the highest value given to a gymnast at EC was 7.0), the 10 most difficult elements are counted, each from 0.1 (A elements) to 0.7 (G elements). Each apparatus has four element groups designated as I, II, III, IV, and, except for floor exercise, a dismount Science of Gymnastics Journal 29 Science of Gymnastics Journal Leskošek B., Čuk I., Bučar Pajek M. TRENDS IN E AND D SCORES AND THEIR INFLUENCE... Vol. 5 Issue 1: 29 - 38 group designated as V. A gymnast is awarded 0.5 for each element group, if he performs at least one element from that group, no matter of its difficulty (except for the dismount where an element must be at least D to receive a 0.5 points). Additional points are awarded for connections of high valued elements. The E score starts from 10 (if at least 7 elements are performed) and has a lower limit of zero, but is usually around 9 points at major competitions (Olympic Games, World and continental championships, World cup). Several other changes were introduced in the new (2006) CoP. Perhaps most influental for E score, deductions for small, medium and large errors and a fall changed from 0.1, 0.2, 0.4, and 0.5 (FIG, 1997) through 0.1, 0.2, 0.3, and 0.5 (FIG, 2001), 0.1, 0.3, 0.5, and 0.8 (FIG, 2006), to 0.1, 0.3, 0.5, and 1.0 (FIG, 2009), respectively. From year 1997 the upper limit of scale of difficulty was extended from super E (F) to G in year 2009 (FIG, 1997; 2009). Several other minor changes in the CoP were introduced in year 2009, e.g. E jury score deductions for too short exercises. In concordance with Boen, Van Hoye, Auweele, Feys, & Smits (2008), who found open feedback causes conformity bias in judges' scores, after year 2009 only final E jury scores (i.e. average E jury score of middle four judges) are displayed to the public and later published in the official results. Several gymnasts, coaches and officials opposed CoP changes in year 2006, especially omitting the traditional 10.0 points format. Although analysis of officiating some major gymnastics events held after 2006 showed metric characteristics (reliability, validity) are generally acceptable (Bučar, Čuk, Pajek, Karacsony & Leskošek 2011; Leskošek, Čuk, Pajek, Forbes & Bučar Pajek, 2012), some other problems arose, which may or may not originate from CoP changes. Thornton (2010) noted, the highest execution score given in the both men and women event final contested at the World Championships or Olympic Games between 2006 and 2009 shows a clear trend of declining and 2009 scores seem to be "trapped" between an 8.5 and a 9.0 points regardless of the performance. Thornton found little justification of this trend in rules changes (apart from women's floor) and speculates the problem lies in application of those rules. The purpose of this paper is to quantify trends in E and D scores in the last year before rule change (2005) and in period after that (2006-2011) at male European Championship as one of the most important gymnastic competitions and possibly find reasons for those changes and their consequences. METHODS Data: All E and D scores from Competition I (qualification) and III (apparatus finals) were retrieved from FIG officials and Internet (Longines timing 2008, 2010, 2011). 2007 scores are missing as they are not published on the FIG Internet site and authors were not able to obtain them from UEG officials. Additionally, 4 routines with zero final scores (all from vault, 3 from qualification and one from apparatus finals) were excluded. Data analysis: The E score was computed as (10-deductions-penalties) for period after 2005. In 2005, E deductions were computed as (Start value-Final score). D score in 2005 was computed as (Start value-4) to make this score more comparable with D scores in 2006-2011. From so defined E and D scores, the gap behind the best E and D scores for each apparatus, session and year (competition) were computed. Basic distributional parameters of E and D scores and their ratios and gaps behind the best score were computed and plotted. RESULTS The number of competitors in the qualification round (Table 1) is different between years 2005 and 2011 and on different apparatuses, but it mostly ranges Science of Gymnastics Journal 29 Science of Gymnastics Journal Leskošek B., Čuk I., Bučar Pajek M. TRENDS IN E AND D SCORES AND THEIR INFLUENCE... Vol. 5 Issue 1: 29 - 38 between 80 and 100, except for 2010 where it is slightly lower. The number of competitors that perform also the second vault varies from 13 (2010) to 35 (2005). The scores from both vaults were joined before the further analysis. In apparatus final, on all apparatus eight gymnasts competed and on vault all perform two vaults except for 2007 where one of the competitors performed only one vault (for the second he received a score of zero, which was excluded from analysis). Table 1. Number of competitors in qualification round by year and apparatus. Year Apparatus 2005 2006 2008 2009 2010 2011 Total Floor 87 87 96 87 75 94 526 Pommel horse 89 92 100 84 73 94 532 Rings 95 88 96 83 77 93 532 Vault 1 86 85 96 80 73 86 506 Vault 2 35 23 26 33 13 27 157 Parallel bars 86 93 99 79 76 92 525 High bar 94 88 97 80 76 89 524 Total 572 556 610 526 463 575 3302 Although some irregularities exist, there is a general tendency of an increase in difficulty score and a decrease in execution scores between 2006 and 2011 (Table 2, Figure 1). This is not only true for the central tendency (median, mean) of scores, but also for maxima. In the year 2011, in both sessions and on all apparatus (except for a tie on parallel bars in qualification), the E score was lower than in the first year after the major rules change (2006); in the D score, in only 2 out of 12 cases, lower scores were observed in 2011 than in 2006. In 2005, the last year of closed-ended system, many routines were awarded the highest possible score for difficulty, i.e. a start value of 10; e.g. in 2005 apparatus finals, three out of 16 vaults and 22 out of 40 routines on other apparatus were given this value. In most cases the decrease in E score was higher than the increase in D score, resulting in a weak trend of decreasing final scores (Figure 1). In most cases differences (variability) between competitors in both E and D score increased in the period 2006-2011 compared to 2005, while there is no clear trend in variability change within the period 20062011 (Table 2, Figure 1). Although E scores remained higher than D scores, the ratio between them has decreased between 2005 and 2011. Although the ratio is much higher in qualification sessions than in apparatus finals, this decrease is similar on all apparatus, i.e. from around 1.5 to 1.8 in the qualification round and from around 1.3 to 1.6 in apparatus finals (Figure 2). Extremely high variability in same cases is mostly due to single extreme outliers (most notably in rings qualifications in 2006, where one of the competitors received a D score of 1.0 and E score of 8.925), which probably arose from too short of an exercise that was penalized only after the revision of the CoP in year 2009. At the beginning and at the end of the observed period, i.e. in years 2005 and 2011, in qualification sessions, competitors are much more heterogeneous in the D score than in the E score, while in apparatus finals in 2005 and 2011 the situation is the opposite, most notably on rings in 2005 where all but one finalists had the same D score, i.e. start value of 10 (Table 3). In intermediate years (2006-10) the variability in the D score is, especially in apparatus finals, only slightly higher than in the E score. Science of Gymnastics Journal 29 Science of Gymnastics Journal Leskošek B., Čuk I., Bučar Pajek M. TRENDS IN E AND D SCORES AND THEIR INFLUENCE... Vol. 5 Issue 1: 29 - 38 Table 2. Medians, interquartile ranges and maxima of D and E score by year of competition, session and apparatus. Year of competition Median Interquartile range Maximum 05 06 08 09 10 11 05 06 08 09 10 11 05 06 08 09 10 11 Difficulty score Floor 5.00 5.40 5.40 5.50 5.40 5.50 .80 .90 .98 .80 1.00 .90 6 6.6 6.5 6.4 6.7 6.7 n o P.horse 5.30 5.10 5.15 5.30 5.20 5.20 .90 .88 1.20 .90 1.10 1.13 6 6.3 6.5 6.8 6.7 6.9 IS Rings 5.10 5.20 5.30 5.50 5.30 5.40 1.20 1.38 1.30 1.10 1.05 1.20 6 7 7.3 6.9 6.8 6.8 S Vault 5.70 6.60 6.20 6.20 6.20 6.20 .50 .60 1.20 .80 .80 .80 6 7 7 7 7 7 u P.bars 5.00 5.10 5.50 5.50 5.20 5.55 .83 1.10 1.20 1.00 1.10 1.00 6 6.5 6.7 6.5 6.5 6.7 o H.bar 5.10 5.50 5.40 5.40 5.40 5.70 .90 .98 .85 1.00 1.08 1.20 6 6.4 7 6.9 7 7.2 pparatu s finals Floor 5.85 6.15 6.15 6.25 6.40 6.40 .17 .35 .45 .25 .43 .37 6 6.4 6.6 6.5 6.7 6.7 P.horse 6.00 5.70 6.20 6.50 6.30 6.40 .15 .83 .30 .27 .43 .68 6 6 6.7 6.7 6.7 6.7 Rings 6.00 6.55 6.70 6.70 6.50 6.70 .00 .85 .75 .18 .78 .18 6 7.1 7.3 7 6.8 6.8 ,( co to o —1 ai ct> co cd o ui ctj t» to o oi oi oo