Sunday, 23 April 2023 19:45

The NCAA Scoring Conundrum

Florida's Riley McCuskter Florida's Riley McCuskter (c) 2022 Lloyd Smith, All Rights Reserved, Used by Permission

Leading up to the NCAA finals, a number of articles hit the mainstream press regarding a potential scoring problem in the NCAA.  Scoring during the 2023 season became quite uneven and at times, blatantly generous, and it became evident that deductions were not being taken.  This was not news to hardcore fans, who had been noticing the wide variation in scoring throughout the season, and the obvious (and not so obvious) missed deductions.  So how did we get here and how do we move forward?

Since 2004....

A very popular refrain we've echoed in headlines and meet summaries is, <team name> hit a score "not seen since 2004".  What makes the years between 2005 and 2020 so special?  Well in 2005, following several seasons of record high scores, two changes occurred.  First, the Judge Assignment System (JAS) was deployed and second, in 2005 and then in 2006, the NCAA Code Modifications were made stricter.  The JAS was set up as an option to assign judges to locations, including rules that limited the number of times a judge could see a team, both home and away.  This was done to help even out scoring variation between regions and reduce bias, intentional or unintentional, from certain judge/gymnast/school pairings.  The NCAA Code Modifications are NCAA specific modifications made to the USA Gymnastics Code of Points for Level 10 competition.  These code modifications included new special requirements, new compositional requirements, new vault values, and new skill values.  In 2005, the level of difficulty required was set higher than ever before, and in some cases, as high as the current code.  As a result, in 2006, the Regional Qualifying Score from the #36 Team was 193.21, down dramatically from the 195.675 posted by the #36 team in 2004. 

A Weakening Set of Code Modifications

Over the last 15 years, the NCAA Code Modifications have become, in some ways, progressively easier while other requirements have become slightly tougher.  A few years back, there was an attempt to toughen the code by reducing the start value of routines on UB, BB and FX before any bonus, from 9.5 to 9.4.  However, at the same time, the NCAA Committee added in additional bonus opportunities and raised the value of certain skills, largely mitigating the impact of the lower starting value.  Since that change, additional skill upgrades and further rule relaxations have made the Code Modifications progressively easier.  For example, today's minimum requirements to start from a 10 on FX are much weaker than the requirements in 2006.      

Rising Levels of Talent

While the Olympians and World Championships Medalists grab a lot of attention among the incoming recruits, the overall level of athlete competing within the NCAA has also risen.  20 to 30 years ago, it was not unheard of for a Level 9 to earn a scholarship and achieve All-American honors at a Top 25 program.  Gradually, over time, the opportunity for a Level 9 gymnast to compete on a Division I team has diminished.  Today, Level 9s are increasingly rare at all levels, from Division III to I.  At the same time, the level of talent competing at Level 10 has increased, with more E skills and more complex combinations attempted, encouraged by the latest USA Gymnastics Development Program code of points.   

Balancing Injury Risk, Wear and Tear

The rising tide of injuries and the wear and tear on gymnasts has been top of mind for decades.  As a result, the NCAA has adopted Code Modifications that are intended to emphasize execution and not overly reward difficult skills, especially dismounts.  The code also was designed to span the range of athletic ability from Division III to I.   This has had the additional bonus of increasing parity among the teams, creating a more level playing field and helping the score grow in popularity.  However, as the level of talent in the NCAA has risen, the ability of gymnasts at all levels has outpaced the NCAA Code Modifications.  While there are available deductions to take, the execution level of today's NCAA gymnast leaves the judges with only a few small obvious errors and the rest in more esoteric deductions to take.  

Resetting the Start Value, and Unintended Consequences

A few years back, the NCAA attempted to strengthen the code and provide more opportunities for separation of the athletes by lowering the base start value from 9.5 to 9.4.  Gymnasts normally earn bonus points above the base start value through difficult skills and bonus combinations.  This means a gymnast needs to earn six tenths of bonus versus the five tenths under the previous set of rules.  However, when the NCAA implemented this reduction in the start value, they also added a series of additional bonus opportunities and raised certain element values that made it easier to earn back the missing tenth.  After a short period of adaption, the gymnasts started posting scores even higher than before.  Gymnasts were able to continue to use the same routines or make simple modifications to existing ones that added little artistically or athletically to their sets.  We see certain skill combinations overused, simple beat jumps added in combinations on the BB or at times, awkward leaps on the FX.  What's more, these bonus opportunities, when combined with changes in the underlying Development Program code, have had the unintended effect of encouraging certain common routine constructions, skill choices and favoring of certain combinations that gymnasts can execute cleanly and consistently.  

The outcome of these changes has heavily impacted the floor exercise.  The popularity of two-pass (tumbling) has exploded in the last two seasons.  The bonus opportunities added to the code and the relatively weak minimum tumbling requirements has diluted the content of floor exercise routines across the nation.  While this has been a positive move for athlete health, it has reduced the variety and dynamics of the tumbling in the exercises and has often been replaced with long passages of choreography with few dance skills of value and too often, marginal artistic value.  It also diminishes the reward for athletes that continue to pursue three tumbling passes and/or tougher tumbling.  In the end, it becomes difficult for both fans and judges to properly separate the good and the great. 

Tougher Evaluation Possible, but with Downsides

Deductions in the NCAA largely follow the USA Gymnastics Development Program Code of Points.  There are plenty of available deductions specified, in most cases, to provide separation of the athletes.  We saw that in practice during the season where teams would have their team totals drop by a point or more, even with no counting falls, just due to variation in how the existing deductions were taken on event to event or meet to meet.  Routines at the club level are often judged more harshly than the same sets in the NCAA (and of course, the code is different).  However, after you take away the clear and obvious deductions, you are left with deductions that are increasingly more difficult for a typical fan to understand, and criteria that is increasingly subjective.  For example, fans and judges alike can consistently rate steps, hops and short handstands (usually, there are exceptions).  But trying to understand the differences in swing quality on UB, body posture and alignment on non-value parts on BB or the precision of footwork on FX begins to get very difficult and opens up a wider range of subjectivity. This can be difficult for fans to understand and can leave many fans disappointed at perceived bias and unfair treatment. 

Reducing Opportunities for Bias, Real or Perceived

The judging assignment system and the limits placed the number of times a judge can view of a team have helped improve consistency in scoring.  Under the system, judges are assigned to judge meets and are restricted from seeing teams more than twice at home and twice on the road.  The schools no longer have direct control over their assignment.  This has helped reduce variation in scoring strictness by exposing judges to more variety of teams, more exposure to a wider range of judges on the panels and has reduced opportunities for real or unintentional bias.  However, it has its limits.  School budgets are always under pressure, especially among smaller programs, and thus it is impractical to increase the judging panel size (from two to four) or increase the amount of travel seen by the judges (by reducing the number of judges from the home area from 50% to 25%). 

Exacerbating these problems is the fact is that limited travel budgets (or travel restrictions due to COVID) and conference obligations mean that some teams see the same opponents four or five times in a season.  Some teams may end up traveling to meets within a relatively small geographical region or state.  This intensifies regional variances in scoring strictness and can lead to surprising drops in team scoring potential when the team travels outside of this geographical area.

Solutions to these problems all take money, something that is often in short supply.   However, improvements are possible in both regular judge training (which occurs today) and changes in accountability.  Today, judges are evaluated by the coaches.  This serves to be a powerful but flawed feedback loop, potentially encouraging high scores that are good for rankings, fan and media attention, and ultimately yield for the coaches, job security.  The simplest, and most cost-effective solution would be to shift evaluations to a mix of random and targeted, video-based audits.  Audit candidates could be identified through data analysis of score variation from panel averages, to check for inconsistency in scoring behavior across teams, gymnast home state and location.     

So What Should Change in the Code?

What are some ways to address the problems under the current code and to provide for more separation between the athletes?  The key is implementing changes that minimize subjectivity as well as encourage variety, to test the athletics skill and ability, but without increasing injury risk.  The current set of NCAA Modifications make it too easy to reach a 10 and to perform a minimum of skills in order avoid deduction opportunities.  This results in a lack of separation between athletes and an overemphasis on small, obvious errors like hops, steps and bobbles.  Ultimately, routines begin to look largely alike, trust in the system is impacted and ultimately, fan interest is impacted. 

Here are some proposed changes on each event that could provide additional opportunities for showcasing skill and ability without resorting to excessive difficulty, while at the same time closing out favored "loopholes" in the current code.  (The below assumes the reader has some knowledge of the existing modifications and rules.  Check here for the 2023 NCAA Modifications.)

Vault 

On vault, the current code is working the best of all four events, for the most part.  One change that has not been successful is the use of direction lines in the landing area of the mat.  It was widely cheered when first introduced, and during the season was helpful in showing vaults that had problems with direction, often due to flaws that are difficult to see in real time, like early turning on the table.  Unfortunately, the resulting impact on the scores awarded has been mixed.  This is because these lines were introduced but no clear guidelines were given to the judges on how they should be used.  Thus, vaults with clear direction issues have often appeared to not suffer any form of consistent deduction.  The underlying WDP Code of Points provides for up to 0.3 of deduction for direction.  The NCAA should implement a consistent standard of evaluation like:

  • Gymnast lands with both feet on or outside the lines:  0.2 to 0.3
  • Gymnast lands with one foot on or over the line:  0.1 flat
  • Gymnast lands with both feet between the center of the two lines and the outside line:  0.05 flat

Uneven Bars

The current uneven bars code does not separate the athletes effectively and has, in effect, favored certain "stock" routine compositions.  The NCAA has wisely decided to not push too hard for dismount difficulty, as it can be a source of injury.  However, things can be done to encourage more variety and provide for better separation of athletes.  Here are a few things to consider:

  •   Deduct 0.05 for the absence of a Group 3, 6 or 7 element (underswings/clear hips, stalders, circle swings/hechts) used in a bonus combination (in addition to the existing 0.05 for lack of variety as specified in 3.2.a)
  •   Restore the deduction for "Up to the Level" in choice of releases to a full 0.2, instead of flat 0.05 today. 
    • No deduction:  Two D+ releases, one of which is single bar release OR one D and E directly connected
    • 0.05 deduction:  Two D releases OR one single bar D directly connected to a C release
    • 0.1 deduction:  One D and one C release
    • 0.15 deduction:  Two C releases directly connected OR a D plus a B
    • 0.2 deduction:  Two C releases
  •   Change the 0.1 release bonus to:  0.1 for any E element (except dismounts) or D single bar release in bonus combination

Balance Beam

The current balance beam code is working out fairly well but does suffer from "minimum skill" routines and an overweight on certain skills and combinations.  Here are a few things they should consider changing:

  • The current "Up to the Level" (UTL) provision provides for additional difficulty when the athlete's acro series does not provide combination bonus.  This can be in the form of an additional D acro (including mounts or dismounts) or an E dance skill (including mounts).  The added encouragement for E dance provides excess weight to certain skills, especially when combined with the upgrade in bonus value to various E dance skills in the NCAA and USAG WDP code (from a D in FIG).  it also contributes to the proliferation of minimum value routines, whereby routines have become stripped down to the minimum possible number of skills to avoid opportunities for deductions.  The additional encouragement for D dismounts as a method of satisfying ULT runs counter to the long-standing principle of not encouraging dismount risk (due to the higher incidence of injury).  This UTL provision should be simplified to an additional D acro (including mounts), like it was in prior versions of the code.
  • The acro bonus combination rules have been modified to treat a front aerial as a C for bonus combinations with back handspring variants (4.4.c in the Code Modifications).  This should also include the kickover front variations (one and two foot landing variations).
  • The beat jump, currently an "A", should be non-valued for the sake of bonus combinations.   

Floor Exercise

The latest Code Modifications for the floor exercise have created a flood of two pass routines, with skills trimmed back to a minimum so sparse that some gymnasts have suffered deductions for accidentally lacking the minimum number of B value skills (5 skills of B or higher are needed).  While this change has the very positive benefit of reducing wear and tear, we've been left with less dynamic routines and little separation between the good and the great.  While changes were added to reward dance and dance skills, these skills have been pared back to the minimum on the event.  Here are some suggestions to restore the balance:

  • Require the 0.1 bonus for a double flipping salto (Rule 5.5) to be in the last 15 seconds of a routine.  Some gymnasts are completing their last pass in the first 30 seconds of the routine, and enjoying a benefit that was intended to reward strong dismounts.  
  • Expand the "progressive distribution of elements" deduction to "up to 0.1".  Require progressive distribution of elements of bonus value in the first third, middle third and last third of a routine, as well as considering the distribution of the difficulty throughout (all in one section of a routine, like closely distributed acro passes).  Lack of acro bonus in the last 20 seconds of a routine should be an automatic flat 0.05 deduction.    
  • Deduct 0.05 if two pass routines do not include at least three C+ dance skills, and not in the same dance passage.  Athletes that choose two passes should still have to show elements of difficulty throughout the routine.  
  • Change the "Up to the Level" requirement to 0.2, but with stronger criteria that sets the previous minimum requirement to a 0.1 deduction.  This may seem harsh, but it will allow more room for objective separation between gymnasts.  The prior criteria allowed a gymnast to start from a ten with a Rudi to split jump, back layout 1 1/2 to punch layout and switch leap to switch side leap with an extra 1/2 turn.  This resets the max score of the same routine to a maximum of 9.9.
  • Deduction Three+ Pass Two Pass
    None

    E salto, 2nd D salto plus min C+A or C+B indirect combination bonus in third OR

    D salto in combination, a 2nd D salto, a D+ dance plus minimum C+B direct or C+C indirect combination bonus OR

    D salto in combination, plus two other D saltos 

    E salto in each OR

    E salto plus 2nd pass with minimum D+B direct or D+C indirect OR

    Minimum D+B direct or D+C indirect in each

    0.05

    D salto in two passes, plus minimum C+B direct or C+C indirect combination bonus in 3rd OR

    D salto, D dance plus two passes with min C+B direct or C+C indirect combination bonus

    E salto plus 2nd pass with D in bonus combination OR

    D salto in bonus combination in one, D dance plus minimum 2nd D salto OR

    D salto in bonus combination in each

    0.1

    One E-valued element (Acro or Dance) OR
    Two different D elements (one may be D dance) PLUS

    Last pass with C salto in bonus combination or minimum D salto

    E salto plus minimum C+B direct or C+C indirect in 2nd OR

    D salto in bonus combination in one, plus minimum D OR

    D salto, D dance plus minimum C+B direct or C+C indirect in 2nd

    0.15 D salto and last pass with C salto in bonus combination

    E salto plus C salto in bonus combination OR

    D salto, D dance plus C salto in bonus combination

    0.2 All other constructions
    All other constructions
     

This UTL change essentially sets the former level required for no deduction to a 0.1 deduction.  It provides for additional separation of routines without excessively burdening all divisions of competition. 

Is the Simplest Solution the Best?

The changes above attempt to strengthen the existing NCAA code modifications, at the risk of increasing complexity.  Perhaps the simplest solution is for the NCAA to simply adopt the USA Gymnastics Women's Development Program (WDP) code of points for Level 10 in its entirety, with minimal modifications.  As noted earlier, nearly all Division I and the majority of Division II and III athletes come from the Level 10 ranks.  Adopting the same code would also help with judging consistency, as the code would be well-known and well-practiced by the judges.  The WDP code forms the basis for the existing NCAA code modifications.  Instead of further strengthening the existing code modifications, a simpler solution would be to reset the NCAA code modifications to the existing WDP code.    

Some simple modifications to the existing WDP code might be required, but the resulting NCAA code modifications could be kept at a minimum.  For example, the landing lines could be retained, and deductions added as we specified above.  The bonus system that provides for an extra tenth of start value at the NCAA level could be eliminated, without too much impact, to retain the integrity of the perfect 10.  Under the current WDP system, a gymnast only maxes out at a 10 but can achieve it with an obvious small deduction, like a step.  She must, however, fulfill a certain set of requirements, including an E skill.  As we noted above, the bonus rules and the elevated skill values on select skills in the NCAA have largely negated the impact of lowering the start value from 9.5 (as used in the WDP code) to 9.4.  In almost all cases, the differences between the NCAA code modifications and the WDP are rooted in extra skills and varieties of skills that test the ability of the athlete and allow for more separation, rather than purely focusing on top level skills.  

What would be different?  Less than you might think.  On UB, the WDP Code adds additional compositional deductions, like the need to perform two of three classes of skills:  a longitudinal turn, a group 3, 6 or 7 element or a forward circling element.  Performing just one is a 0.1 deduction, performing none is a 0.2 deduction.  The WDP Code also provides a full 0.2 for "up to the level" on releases, a set that is very similar to our proposal above.

On BB and FX, the WDP Code provides for a full 0.2 in up to the level (UTL) deductions for Acro and separately for Dance.  While there are considerable differences in these deductions compared to the NCAA, the biggest difference is on the BB.  On BB, a gymnast who wants to avoid starting from a 9.9 due to built-in compositional deduction may have to add an additional C or D dance skill and an additional C salto or D acro skill.  On FX, a gymnast may need to add a dance skill or upgrade their existing dance skills.  To reach the higher scores of 9.9 and above, a gymnast may need to upgrade their tumbling, especially in instances where they are just meeting the minimum NCAA requirements.  With perhaps the exception of added dance difficulty for all gymnasts, these changes would not be unique to the NCAA.  In fact, many are very similar to prior versions of the NCAA code modifications.

Looking Ahead

The NCAA may choose to make changes to the existing code modifications and increase accountability, or they may minimal tweaks.  Last season, they added a number of bonus opportunities and effectively weakened the code.  The direction they choose will be determined by June.  We can only hope for some change, as the outcry for change reached new heights this past season.     

 

Login to post comments