In this week’s edition of the number factory, I started out by comparing how each of the current top 25 teams are scoring after three weeks of competition this year versus last year, and a few interesting things came out of that. First, let’s look at the numbers.

Difference in average score after three weeks – 2013 vs. 2012
1. Eastern Michigan +3.425 (Ranking difference: +23)
2. Pittsburgh +2.912 (+20)
3. Michigan +2.617 (+18)
4. California +2.400 (+17)
5. Kentucky +1.925 (+14)
6. Minnesota +1.913 (+16)
7. Central Michigan +1.758 (+14)
8. LSU +1.541 (+10)
9. Maryland +1.480 (+10)
10. Kent State +1.425 (+9)
11. Denver +1.200 (+7)
12. Arizona +0.729 (+2)
13. Florida +0.659 (+5)
14. Oklahoma +0.392 (+3)
15. Stanford +0.312 (+2)
16. NC State +0.192 (-4)
17. Alabama +0.188 (+1)
18. Auburn +0.133 (-2) 
19. UCLA +0.013 (-3) 
20. Nebraska -0.075 (-2)
21. Georgia -0.383 (-2)
22. Boise State -0.413 (-4)
23. Utah -0.588 (-5)
24. Ohio State -0.641 (-8)
25. Oregon State -0.700 (-6)

First, I was surprised at how many new teams we currently have in the top 25 that almost never spend any time there, with seven new entries that were out of the top 25 last season. That’s a sizable number for a sport that traditionally sees little variation in the top teams. While I do expect the upstarts to fall, being the surprising early entrant in the top 25 was exactly how Kent State began the 2011 season. 

Also interesting to me was that just six of the top 25 currently have lower averages than they did at this point last season. Granted, some of that is because the teams that have fallen significantly from last year (Hello down there, Arkansas!) have dropped out of the top 25 completely, but another reflector of higher overall scores is that we have a few teams with higher averages that have actually dropped in the rankings. To illustrate the overall shift, last year a 194.000 average after three weeks of competition was good for #25. This year, it gets you #31. In fact, the current average score of all the top 25 teams is historically high. 

Average score of top 25 teams after three weeks
2013 – 195.639
2012 – 195.301
2011 – 194.998
2010 – 195.071
2009 – 195.121
2008 – 194.866
2007 – 194.710
2006 – 193.972
2005 – 194.570
2004 – 195.761

Note of slight inconsistency: In a couple of these years I actually took the numbers from the fourth week of competition instead of the third because so few teams competed in the first week (well under half) that it hardly could count as a week of competition. That ended up making the numbers higher than they might otherwise have been since the averages usually increase as the weeks go on. And yet, they were still well lower than the 2013 average.

It appeared last season that judging was starting to move concertedly back toward those 2004 levels where everyone got a 9.975 for a fall, and while the sample size is currently too small for this season to make any kind of sweeping statement, the increase we’re seeing is not insignificant. Happy 2004.

Is increased scoring a bad thing? Not necessarily, but it is notable. The separation is much more important than the total, and it’s something to watch as we go on. 

As teams move into February and March, the discussion will turn to finding 9.9s and how to turn 9.8s into 9.9s. The conventional wisdom is that it takes around twelve 9.9s (three per event) to have a strong shot to win a championship. At this point in the season, that is an unrealistic expectation. The expectation right now is that the top teams will begin to eliminate those scores under 9.800, the ones that have no place in a March or April lineup and that are unlikely to ever become those reliable postseason scores. 

Looking at the proportion of scores that top teams are recording under 9.800 gives a good sense right now of how much work needs to be done in terms of both eliminating falls and finding the competitors with the highest scoring potential. Whether it’s a 9.775 or a 9.300, that score cannot be accepted later in the season by the best teams. 

Because there is so much inconsistency in the first few weeks of a season, teams recording 9.750 after 9.750 can be ranked fairly well, but that team is unlikely to feature in the postseason and that will be reflected in these percentages as opposed to teams who counted disasters but are soon to improve.

% of routines scored under 9.800 for top 25 teams
1. UCLA [1] – 25.00%
2. Michigan [2] – 27.08%
2. LSU [5] – 27.08%
2. Florida [3] – 27.08%
5. Oklahoma [3] – 31.25%
6. Alabama [6] – 37.50%
6. Georgia [8]  – 37.50%
6. Utah [15] – 37.50%
9. Auburn [24] – 52.08%
10. Nebraska [7] – 54.16%
10. Arkansas [9] – 54.16%
10. Arizona [13] – 54.16%
13. Stanford [10] – 57.45%
14. Denver [11] – 58.33%
14. Minnesota [12] – 58.33%
16. California [17] – 60.42%
17. Maryland [17] – 62.50%
18. Kentucky [14] – 66.66%
18. Central Michigan [16] – 66.67%
18. Ohio State [21] – 66.67%
18. Oregon State [23] – 66.67%
22. NC State [19] – 70.83%
23. West Virginia [22] – 79.17%
24. BYU [25] – 81.25%
25. Southern Utah [20] – 83.33%

Utah and Auburn (in particular) are strong examples of teams that outperformed their rankings and are likely to move up soon based on the proportion of 9.800+ routines.  

We have no rankings this week because some of the teams are lazy or whatever and didn’t compete last weekend. This also gives the teams that performed well over the last couple of days a chance to screw it up before it counts. If there were rankings, our top 5 would have been as follows:

1. Michigan – 196.575
1. Florida – 196.575
3. Oklahoma – 196.450
4. UCLA – 196.150
5. LSU – 196.075

The first three teams should be pleased with the current course. Each had issues, but there was nothing of major concern for the first meet. The 4th and 5th teams both counted falls and still broke 196, which should seem like an accomplishment. However, the fall on bars for LSU doesn’t provide a ton of optimism for depth on their worst event, and UCLA looked depleted on every event, especially vault, where they don’t have a 4th routine right now, let alone the 7th and 8th options that good teams need.

Since there are no rankings, let’s take a deeper look at the scores for the teams from the preseason top 25 who were in action this week. Stats include all competition routines, not just the five counting scores.

National average by apparatus – Week 1
1. Vault – 9.784
2. Bars – 9.684
3. Beam – 9.649
4. Floor – 9.645

It’s no surprise that vault was by far the most successful event. Even though some high profile teams struggled or at least recorded their weakest scores there, vault sees far fewer falls than the other events, which primarily accounts for the significant advantage in score. In a bit of an upset, beam was actually higher scoring than floor for the top 25 teams, but the difference is too small to be significant.

Average routine score by team – Week 1
1. Michigan – 9.817
2. Florida – 9.775
3. Oklahoma – 9.747
4. UCLA – 9.742
5. LSU – 9.724

In spite of introducing the dropped scores into the numbers, the rankings remain identical to the top team scores. Even if we were to continue the rankings, Stanford and Georgia would be next, which also follows the results. Michigan comes out as the clear leader over Florida in these rankings because of much higher 6th scores, so I suppose Michigan wins the week in the tiebreak.

Average vault score by team – Week 1
1. Stanford – 9.850
2. Florida – 9.842
2. LSU – 9.842
4. Ohio State – 9.838
5. Georgia – 9.821

The team that vaulted only five people won the event. None of this dropped 9.725s slop.

Average bars score by team – Week 1
1. UCLA – 9.846
2. Florida – 9.829
3. Georgia – 9.821
4. Michigan – 9.796
5. Ohio State – 9.779

It’s positive news for Ohio State in the future that the team made two of the event leaderboards. They are going to be far better than that 194.600.

Average beam score by team – Week 1
1. Michigan – 9.838
2. Stanford – 9.804
3. LSU – 9.792
4. Denver – 9.725
5. Florida – 9.713

Denver had just one score over 9.800 and Florida had a fall, and yet both made the top 5. That’s the state of beam right now.

Average floor score by team – Week 1
1. Michigan – 9.846
2. Oklahoma – 9.829
3. LSU – 9.788
4. UCLA – 9.783
5. NC State – 9.754

LSU and UCLA both had falls from this group, but their 9.975s saved them. Plus, a lot of teams were in the 9.5s. Ouchie.

Worst rotation averages by team – Week 1
1. Ohio State floor – 9.404
2. Penn State beam – 9.458
3. LSU bars – 9.475
4. Oregon State bars – 9.479
5. Auburn beam – 9.483

Speaking of ouchie. 
 

The balance beam, she is a cruel mistress.

In 2011, I examined her cruelty by analyzing how teams fared when performing beam routines after a fall earlier in the rotation. The study yielded the following breakthrough: everything is horrible. Nearly every team had significant difficulty breaking 9.800 in those post-fall routines, often suffering a subsequent fall (or two). 

But what’s this? A ray of hope? For this past season, I analyzed the same data for the schools that qualified to Championships, averaging the scores for all routines performed on beam at any point after a fall or fall-equivalent performance (a score of 9.500 or lower), and found significantly stronger results across the country. In fact, beam was the lowest-scoring apparatus for a grand total of none of the teams during this year’s Super Six, and most teams put up respectable numbers over the course of the season.

Average beam score after a fall – 2012
1. Florida – 9.885
2. Alabama – 9.841
3. Oklahoma – 9.869
4. Utah – 9.814
5. Arkansas – 9.809
6. UCLA – 9.808
7. Ohio State – 9.780
8. Oregon State – 9.769
9. Georgia – 9.725
10. Nebraska – 9.722
11. Stanford – 9.700
12. LSU – 9.688

For comparison, last year’s leader, Utah, would have placed 7th in this year’s rankings. In fact, compared to last season, Georgia and Stanford are the only schools analyzed in both lists that regressed in their performances. Georgia actually did pretty well on beam through most of 2012, but the disaster from Championships brought the average down significantly. 

Florida managed a stellar 9.885 and did not record a post-fall beam score lower than 9.850 all year. Now, the Gators did perform only 5 routines after falls all year, so they did not have as many opportunities to ruin everything.

Number of beam routines after falls – 2012
1. Florida – 5
2. Oklahoma – 8
3. Stanford – 10
4. Georgia – 12
5. Alabama – 14
6. Utah – 20
7. Oregon State – 22
8. Arkansas – 25
9. UCLA – 26
10. LSU – 31
11. Ohio State – 32
12. Nebraska – 38

The difference in consistency is slightly amazing. Rare was the week that Nebraska or Ohio State was not fighting an early fall. Nebraska performed 78 beam routines as a team last year, and 38 (49%) of those routines took place after someone else had already fallen.

In theory, these first two lists should match up almost exactly because the best teams on beam will suffer the fewest falls and record the highest scores overall, but as we see that’s not always true. For instance, our regressed schools Stanford and Georgia performed relatively few routines after beam falls but fared rather poorly in those routines. Arkansas and UCLA, however, maintained relatively strong averages considering how many pressure-filled beam routines they were forced to perform.

Interestingly, our individual standouts did not always come from the strongest teams.
Usually, they were the select few who saved their teams from being unqualified disasters.

In creating these individual lists, I limited the numbers to those who performed at least 3 routines after falls because otherwise the sample size is just too small to make any kind of argument about quality.

Best average beam score after a fall – 2012 (minimum 3 routines)
1. Leslie Mak (Oregon State) – 9.900
2. Sam Peszek (UCLA) – 9.883
3. Katherine Grable (Arkansas) – 9.881
3. Vanessa Zamarripa (UCLA) – 9.881
5. Sarah Miller (Ohio State) – 9.875
6. Ashley Priess (Alabama) – 9.858
6. Geralen Stack-Eaton (Alabama) – 9.858
8. Kyndal Robarts (Utah) – 9.855
9. Cortni Beers (Utah) – 9.845
10. Jessie DeZiel (Nebraska) – 9.842
10. Jaime Pisani (Arkansas) – 9.842

Mak performed 5 routines after falls and scored 9.925 in 3 of them, including a vital Regionals performance where even a 9.800 would have seen the Beavers lose their Championships slot to Michigan. Interestingly, Priess and Stack-Eaton both performed 3 routines after falls, and each recorded a 9.925, a 9.875, and a 9.775 for the exact same average.

Worst average beam score after a fall – 2012 (minimum 3 routines)
1. Kaleigh Dickson (LSU) – 9.500
2. Brittany Skinner (Nebraska) – 9.555
3. Jamie Schleppenbach (Nebraska) – 9.588
4. Sarah Persinger (Georgia) – 9.512
5. Brittany Harris (Oregon State) – 9.613
6. Lloimincia Hall (LSU) – 9.663
7. Mary Beth Lofgren (Utah) – 9.675
8. Colleen Dean (Ohio State) – 9.696
9. Rheagan Courville (LSU) – 9.709
10. Lora Evenstad (Nebraska) – 9.738

Oh, Nebraska and LSU. Brittany Skinner won this ignominious title last year with a score more a tenth lower, so progress has been made. Overall, far fewer people were put up in high-leverage situations who were not comfortable in them. Courville and Evenstad made this list, but those averages aren’t even particularly worrisome.

Finally, special commendation for 4 mental giants whose teams subjected them to terror on a weekly basis. Each of these gymnasts performed 9 beam routines after falls during the season and did not fall once.

Team savior award – 2012
9 routines, 0 falls:
Sam Peszek (UCLA) – 9.883
Katherine Grable (Arkansas) – 9.881
Sarah Miller (Ohio State) – 9.875
Jessie DeZiel (Nebraska) – 9.842