Category Archives: Fun with Numbers

First Meet History

The first meets of the season are in the books (for almost all the teams), and since everyone has decided based on only that who the Super Six teams will be, we should probably just fast forward to regionals, right? We have all we need to know.

Last night, Stanford performed a catastrophic floor rotation, a clunky beam rotation, and a surprisingly OK vault rotation to score the traditional first-meet 1.100 in a show of true compassion for Georgia. Stanford’s 193.250 is its lowest first-meet score in over a decade, and yet my general impression was, “Could have been worse.” So there’s that. Stanford would like us to know that starting slowly is all part of a cunning master plan that works almost a third of the time. UNSTOPPABLE.

Considering various teams and their general trends of starting slowly/quickly and whether that’s a good thing or a bad thing or even matters, I plotted the first-meet scores of the eventual champions for each of the last 10 years.


What this tells us is that there’s not a rule. You could be Oklahoma in 2014 and start high, stay high, and end high (team motto?), or you could be Alabama in 2011 and start in the garbage before pulling it back. The good news for teams that struggled in the first meet this year is that apart from that Oklahoma 2014 number, these aren’t overwhelming scores. (But they’re also not 193s, Georgia and Stanford.) Champions don’t have to be champions in the first week and often aren’t.

That’s also reflected in the average first-meet scores for the six ultimate Super Six qualifiers.


In 2014, the good teams all started well and remained excellent right through to the end, but that’s sort of an outlier. In 2015, Stanford and Auburn were trash in the first meet and came back to make Super Six. In 2011, none of the six final qualifiers scored higher than 195.700 in the opening meet. Scores were lower as a whole in 2011, but not that much lower. So it can be done. Continue reading First Meet History

Does It Pay to 1.5?

The start value of the Yurchenko full may have been changed to promote more variety on vault, but the most delicious byproduct of the move has been the creation of a big, fat dilemma. The new strategic twist for coaches to grapple with: does an extra .05 actually make the Yurchenko 1.5 worth it, or is it smarter to stay with the trusty full?

Preseason training videos reveal that this dilemma is even more widespread this year than last year, and over the next month or so, coaches will have to make major decisions about whether their gymnasts should actually compete that wonky 1.5 they’ve been training.

So……should they?

Thankfully, we now have a whole season of evidence to use in making that decision for them, so let’s take a look at whether competing the Yurchenko 1.5 actually ended up being worth it in 2016.

Item 1: The average vault scores from the national championship (semifinals and Super Six), separated by type of vault. The first table includes all vaults, while the second table removes the falls.

2016 Nationals – Average Scores (with falls)
Yurchenko 1.5 9.829
Other vaults 9.821
Yurchenko full 9.803
2016 Nationals – Average Scores (no falls)
Yurchenko 1.5 9.867
Other vaults 9.863
Yurchenko full 9.803

The story these tables tell is a relatively optimistic one for Team 1.5. Even with a few falls at nationals, the 1.5 still ended up being more valuable than the full on average, a good argument for its being worth the risk. When falls are not included, the margin between the vaults balloons to 0.064, greater than the 0.050 difference in start value. Continue reading Does It Pay to 1.5?

Returning Routine Rankings 2017

Yes, that is a picture of D-D Breaux in a pink hardhat. Because there doesn’t need to be a reason.

Now that the NCAA schedules are finally coming together-ish, it’s probably important for us to start remembering who the people are and what the things do. It’s a really tough job. We need three months.

Before beginning to evaluate this year’s incoming freshmen, I decided to check out where the teams stand without them, how they rank using only scores from 2016’s returning gymnasts. It’s a totally scientific and unimpeachable way of quantifying just how much work the freshmen and new transfers will need to do for teams to return to (or improve on) last year’s level.

When available, I used RQS for each gymnast, but when not, I used full season average.

Most teams do return at least five people who competed on each apparatus last year, but when they don’t, I filled out the remaining scores with punishment 9.700s (I told you, totally scientific). It’s a way of making sure each team has a comparable total, operating under the belief that for these top 15 teams, the backup gymnast who wasn’t good enough to compete probably would have scored a replacement-level 9.700. That is, unless the returning scores were already lower than that (*cough* Utah’s beam *cough*).

1. LSU – 197.726
Gnat – 9.965
Ewing – 9.905
Hambrick – 9.880
Finnegan – 9.835
Cannamela – 9.835
Macadaeg – 9.790
Priessman – 9.750
Finnegan 9.915
Hambrick – 9.905
Zamardi – 9.875
Priessman – 9.869
Gnat – 9.727
Cannamela – 9.663
Finnegan – 9.915
Gnat – 9.895
Macadaeg – 9.890
Hambrick – 9.885
Ewing – 9.870
Priessman – 9.725
Cannamela – 9.603
Gnat – 9.980
Macadaeg – 9.950
Kelley – 9.885
Hambrick – 9.880
Ewing – 9.865
Finnegan – 9.692
Zamardi – 9.517
Cannamela – 9.050

Losing only Savona and a not-100% Wyrick from last year’s Super Six team, LSU is sailing smoothly on most events. Continue reading Returning Routine Rankings 2017

United States v. Russia v. China

Following up on my exploration of which skills are the most popular in US elite routines, I decided it would be interesting (to me exclusively) to compare US composition to routine composition in Russia and China to illustrate the very different approaches taken by the three countries and where they can learn from each other.

And by each other, I mean the US. Because, let’s be real, the US won by 8 points.

The US numbers are based on routines from the three major domestic competitions this summer, Russia’s are based on Russian Cup (and occasionally Russian Champs if the gymnast didn’t compete at Russian Cup), and China’s are based on the Chinese Championships. I did not include all of the seniors from Russia and China at those meets because…well, they’re not on Youtube. But also because many gymnasts attend those meets to compete for their region/province but aren’t international elites and don’t have a comparable skill level.**

So, here we go. The “winner” for each skill is highlighted.



  • Russia and China have been much more diligent about getting rid of those trash-shoots that do nothing to boost the D score than the US has, though the toe shoot does remain popular among the bad Chinese bars workers—the non-L-grippy ones who aren’t allowed to be seen in public and have long since been given up for dead because they’re not Fan Yilin. (Yeah, I’m talking about you, Liu Jinru.)
  • The toe-on Shaposh 1/2 is Russia’s compulsory bars skill, while China is more comfortable with Stalder Shaposhes than either the US or Russia. It is interesting to note how few Chinese gymnasts do any variety of Shaposh 1/2 considering how valuable it is and how high their D scores are nonetheless. The toe-on Shaposh 1/2 is absolutely essential to Russia’s high bars Ds.


  • And we thought the lack of transition variety in the US was bad. While the Pak is “compulsory” in the US, it is LITERALLY COMPULSORY in Russia and China. (Which also explains the lack of shoots since the Russians and Chinese never even face that direction on the low bar.) The US is the only nation crass enough to still use outdated and obsolete bail handstands.


  • It’s worth noting not only how few same-bar releases are being done, especially by Russia and China, but also how different each country’s choices are. Every US gymnast does a straddled Jaeger, every Russian does a piked Jaeger (which will suit them very well in the next code), and every Chinese gymnast does a Gienger.

Continue reading United States v. Russia v. China

American Elements 2016: The Hottest (and Frostiest) Skills in the US

It’s that time of year. Put on your gymnerd hat, your spreadsheet suspenders, and your white orthopedic percentage-comparison socks because oops, I made some charts again.

As is annual and traditional, I have taken all the skills performed by US senior gymnasts this summer and ranked them based on the percentage of total routines in which they appeared, comparing that data to the same information from the previous four years. Which skills are wildly popular? Which skills are horrible losers? How has it changed over time? Let’s find out.

I have highlighted a few of the significant trends in bluish and reddish as a way of COLORS.

Also the usual disclaimer that I didn’t include skills like giants and back handsprings because meh. Everyone, obviously.

So, like an NCAA away team, we’ll begin on uneven bars and let everything go downhill from there.



  • The rise of original-recipe Tkatchev. I didn’t expect this. Compared to 2015, the regular Tkatchev is more popular this year, while E releases like the piked Tkatchev and Stalder Tkatchev are less popular. It should be the other way around considering the extreme value of E releases for connection bonus. Then again, the US gymnasts racking up the D score on bars right now like Locklear and Kocian are not same-bar releasies by any means.

Continue reading American Elements 2016: The Hottest (and Frostiest) Skills in the US

Bars Scores: Pretty Cracky, Right?

Right. Let’s get into it.

If, upon subjecting yourself to some of the bars execution scores from the Olympics, you began formulating questions like, “Huh?” “What?” and “How’s your crack addiction?” you were not alone.

To me, the most unexpected scoring-related development at the Olympics was those massive bars scores (high vault scores and beam taking forever were both way too predictable), especially compared to previous years.

This table lists the average execution scores awarded this quad during all world/Olympic finals (team, AA, and event) on each apparatus.

Year VT Execution UB Execution BB Execution FX Execution
2016 9.027 8.549 8.267 8.324
2015 9.097 8.259 7.906 8.338
2014 8.933 8.150 8.179 8.058
2013 8.962 7.848 7.716 8.117

We have some degree of Land of the Rising Scores happening on all the events compared to 2013, which is consistent with the 2012 quad when the execution scores were alarmingly low in 2009 and rose progressively from there. Continue reading Bars Scores: Pretty Cracky, Right?

*UPDATED* US Women’s Olympic Team Calculator

NOTE: This has been updated to include the results from both days of P&G Championships.

Following up on the men’s Olympic team calculator, I’ve also done one for the women so that we can compare the scoring potentials of all the various team options we could possibly imagine, and some we couldn’t imagine.

We don’t have as much to go on yet for the women, but this calculator takes into account all scores recorded so far in competition in 2016 to make up for it.

Once again, here’s how it works. In the cell next to “Team Member #X,” write only the surnames of the five gymnasts on the prospective Olympic team you’d like to test. Once you’ve listed at least three gymnasts, you’ll see a three-up, three-count total for that team on each event along with a full team total. There are two calculation options here. The first is based on the highest score each gymnast has achieved in competition this year on each event, and the second is based on the average score each gymnast has achieved in competition this year on each event. The optimism option and the realism option.

As a handy reminder of who the people are and how their last names are spelled (be sure they’re spelled correctly), here are the 24 women included in this calculator: Biles, Raisman, Douglas, Hernandez, Nichols, Smith, Skinner, Baumann, Dowell, Locklear, Kocian, Gowey, Schild, Hundley, Dennis, Desiderio, Gaskins, Navarro, Frazier, Trautman, Paulson, Ramler, Musselman, DeGuzman.

*FINAL* US Men’s Olympic Team Calculator

NOTE: This has been updated to include the final results from Olympic Trials.

Because we’ve all (and by “we’ve all,” I obviously mean just me) been agonizing over various US men’s team options ever since nationals made everything harder, I’ve put together a handy-dandy little calculator so we can assess and compare all the possible groups of five Olympians (both realistic and five-Paul-Ruggeris-type). That way, everyone can join in the fun! This is what fun is, right?

Go ahead and give it a whirl!

Here’s how it works: In the cell next to “Team Member #X,” write the surname only of the gymnasts on a prospective Olympic team. Once you’ve listed at least three gymnasts, you’ll see a three-up, three-count total for that team on each event—as well as a full team total—based on the scores from nationals.

There are two options here, the first one is based on the highest score received by each gymnast at nationals (if you’re a glass-half-full kind of person), and the second one is based on the average score received by each gymnast at nationals (if you’re a glass-half-pommel-horse kind of person).

As a handy reminder of who the people are and how their last names are spelled (be sure they’re spelled correctly), here are the 18 remaining options for the US men’s Olympic team: Mikulak, Brooks, Dalton, Modi, Melton, Whittenburg, Moldauer, Kimble, Ruggeri, Orozco, Maestas, Naddour, Oyama, Bailey, Penev, Leyva, Legendre, Wynn.

She’ll Never Get Those Scores Internationally

Secret Classic, the most important gymnastics competition of the year until whatever’s next week, will be upon us as fast as you can say, “That connection is stupid, honey.” With it will come a heap of Olympic team predictions and proclamations about how those scores will or will not translate to the Olympics, burying us under a pile of our dear old friend, “She’ll never get those scores internationally.”

But will she?

The answer is…mostly. Sometimes.

Let’s begin in 2012 and work forward from there.

Here, I’ve taken the average execution score each US Olympic team member received at the four major 2012 competitions (Classic, Nationals, Trials, and the Olympics) and plotted them by event to compare the scores received domestically to the scores received internationally. I’ve excluded team members who did not ultimately compete that event at the Olympics—because then there’s no point of international comparison—so vault does not include Ross’s domestic scores and the other events don’t include Maroney’s.


Domestically, most events saw a slight increase in execution scores toward Trials, with Trials featuring the most enthusiastic judging (or, a nice person could argue, the most perfected routines). That’s something to keep an eye on this year as well, Classic as the most conservative of the US meets.

Once we arrive at the Olympics, the execution scores decrease on some events, but not all events and not for every gymnast. As is well documented, vault execution scores had a prescription drug problem at the 2012 Olympics and were largely off the chain, higher than at any point in the US season. Beam also remained quite constant, falling just slightly for the Olympics (a number which includes mistake routines from Douglas in EF and Raisman in the AA). The execution scores for hit beam routines between Trials and the Olympics were similar.

Of course, D score on beam was a different story, mostly because of Wieber’s walkover hell sandwich that the Olympics judges scraped off the bottom of their shoes with a stick and wiped on the curb. That’s where we can point to US judges doing a disservice by propping up unrealistic D expectations, but in execution, what we saw early was what we saw later. Continue reading She’ll Never Get Those Scores Internationally

American Skills – 2015 Edition

It’s that time of year again. A brand new batch of fresh summer routines has once again been bestowed upon us by Mrs. Karolyi’s Traveling Circus, which means it’s now my turn to break these routines down into their constituent skills to see what trends in routine composition emerge. Which skills have become the coolest kids in school and are totally dating Brett Bretterson? Which loser skills are eating lunch by themselves in a bathroom stall like Stephanie Tanner before Gia teaches her about smoking and Ace of Base (and meth, probably)? And does any of it make sense? Or are all these routines stupid?

Let’s get into it. On each event, the skills are broken down by category, with the corresponding numbers indicating the percentage of US senior elites who performed that skill at the national championship in the given year. I have included all skills of C value or greater (so none of this bhs or giant nonsense), as well as the A dance elements on beam as a way of keeping tabs on how people are choosing to fulfill the dance combination requirement. As always, I counted the skill attempted rather than the skill that would actually receive credit because this is about evaluating intended composition choices. Though let me tell you, that was a rough game this year on floor. We’ll get there in a second.

Some of the notable rises, falls, and year-to-year comparisons are highlighted. Because people like things with colors on them. Apparently. 


  • The tkatchev made a nice little comeback this year in most of its flavors (stalder, piked, and plain). Only the toe-on variation saw a fall in 2015 as more people have started performing a greater variety of entries, which is always a good thing. It also makes sense to Shayla up these bars routine since tkatchev variations are so valuable for CV right now. Last year, I was a little surprised by how few we saw, but they’re coming into line now. Overall, the gymnasts are stepping up the difficulty with their bars releases. Although some of that is just Brenna making the whole group look like daredevils.   
  • The straddled jaeger remains the gold standard of non-tkatchev releases. As in, the only one. No piked versions this year, and no giengers again. Poor gienger. The gienger is a leg-separation deduction trap (hi Sophina!) while the straddled jaeger isn’t, so if you’re choosing one, it makes sense to choose the jaeger. But seriously, you’re telling me no one out there can throw us a nice little Peszek-level gienger?

  • Everyone remains all about the toe-on. And by everyone, I mean 52.63% of people. I’m still not really clear on why this trend has come on so forcefully in the last couple years. Sure, many people do need an 8th skill to count and the toe-on is the easiest C element on bars, but that was also true in the last quad, when relatively few people were doing toe-ons with no pirouetting (14% in 2012).  
  • We also saw a bit of an upswing in the toe-on 1/2s this year, though not a terribly significant one. This is understandable as a result of the popularity of jaegers. Got to do something to get facing the right direction.  
  • The stalder full continues to be the big loser among bars turns after having enjoyed greater popularity in the last quad when D pirouettes were more valuable for CV. Now, not so much. The relatively strict deductions for late pirouettes (compared to releases) have also contributed to this decline since the value of the skill just isn’t worth the potential for a large late-finish deduction. Now, the gymnasts shove their one D pirouette into the routine (toe-on full) for a necessary D skill, but they’re loath to put in another one if they don’t have to.  
  • Sadly, the weiler 1/2 (aka, the wolf turn of bars) has returned to its previous levels after a refreshing dip last year.

  • More than anything else, this is the quad of the pak salto, which was clear to even more extreme degree this year than in the previous two years. The main (only) reason for the emergence of the Pak Posse is the pak’s ability to get gymnasts facing the right direction for shaposh variations. Couple that with the elimination of the bail+stalder shoot CV, and there’s no longer much reason at all to do a bail. In fact, pretty much everyone still doing a bail is also doing a pak and just needs another countable high-to-low element.
  • The lone ranger not doing a pak is Aly Raisman. Every other senior has a pak (or Bhardwaj, which is a variation on the same theme) in her routine. Raisman also has one of the lowest D scores. So that’s not a coincidence.
  • But! We have four different high-to-low transitions this year instead of two. So that’s something. Everyone light a candle for Brenna and Kyla.

  • One of the stranger blips in 2015 is the decrease in shap 1/2s compared to last year. It’s not large enough to be a thing, but I would have expected an increase this year since it’s so valuable for D score. You’d think more people would be learning it. I especially expected to see an increase show up in the numbers because these bars routines are getting much harder. Last year, 42% of US senior competitors had a D score in the 6.0+ range. This year, it was 74%.
  • But note that the stalder shap is getting increasingly popular, mostly with people who already have the toe-on version in their routines and are now including two shaps. That’s one of the places this increased difficulty is coming from.
  • Unsurprisingly, shoots to the high bar continue to be an endangered species as a result of the bail issue discussed above, but the toe shoot in particular in an interesting case of composition change in a very short period. 
    In the 2012 quad, few people performed the toe shoot because it’s a B and couldn’t get CV out of a bail, while the C-rated stalder shoot would. In 2013, we saw a jump in the frequency of toe shoots compared to stalders as many people retained their 2012 composition but realized they could do the toe shoot instead of the stalder since neither would receive CV anyway. But lately, we’ve seen the toe shoot start to decline once again as people adapt to the code of this quad, emphasizing more pak-shap action without needing a shoot at all. 
  • Yay! This year, the disappointing trend of repetitive bars dismounts reversed a tiny bit as we saw four whole dismounts (four!) compared to last year’s two. A few of those half in-half outs and DLOs have been replaced by DLO fulls and double fronts (however misguided) for some more variety to get us back closer to previous levels. It’s a start. 


    • The queen is dead, long live the queen. After a rough year of being exiled to the wilderness by that pretender to the throne, the layout stepout, the aerial walkover mustered an army and forced a revolution to regain her rightful place atop the beam acro standings.
    • The acro landscape this season is actually fairly different than it was last year and reflects a reversion to the mean, more similar to previous years. The overarching trend here is increased difficulty as we see a drop in D acro elements like the side aerial, punch front, and side somi and a corresponding increase in harder elements like the layout (2 feet), front pike, and barani. D acro elements had been on the rise because of this quad’s emphasis on acro + dance combinations, but why do D acro + A dance when you can do E acro + A dance? (Because you’re Aly Raisman and it makes you fall on a split jump?) 
    • Back pikes and tucks are also up this year with several more people electing to do switch split + back pike/tuck/puck to squeeze out that 0.1 CV.  

      • 100% switch split. Oh boy. Since I started doing this in 2012, this is the first skill that has appeared in 100% of US senior routines (not counting round offs and back handsprings). If it was ever going to happen, this was the skill. The switch just makes too much sense because of how many birds can be killed with this one stone. (I just now realized what a horrible expression that is.) 
        Not only is the switch split the easiest C dance element (and since pretty much everyone will end up counting at least one C dance element, everyone wants to do a switch split), it’s also valuable for that connection into a C acro skill for 0.1 CV or into a split jump/sissone to fulfill the dance combo. But at the same time, do we approve of any code that will encourage a skill to appear in 100% of a nation’s routines?
      • Another interesting development is the slight decline of the split jump. A few more people this year exercised different options in completing the dance combo, either through other A elements (note the uptick in wolf jumps and pike jumps) or the Baumann strategy of doing switch to switch half and daring the judges not to award the connection. 
        There may also be a concern that not hitting 180 is getting deducted more strictly on split jumps, since it’s the most obvious of the dance element errors, and perhaps people feel they can get away with a little bit more suckiness on a wolf, pike, or sissone than on a split. We also need to talk about the dire state of these sissones. They’re basically just split jumps with a low front leg. That’s not really the deal.
      • I included a total wolf turns category this year because wolf turns have dominated the conversation lately. While much more prevalent for the seniors in 2015 than in 2014, wolf turns weren’t performed by everyone the way it sort of seemed. We did see quite a few in the juniors, though, which indicates that it will be a growing trend in coming years. 2013 also saw a high percentage of wolf turns, though there were far fewer total competitors in 2013 so it may not have seemed like as many.  

        • A little bit of a comeback for the double pike this year. The proportion of people doing dismounts more difficult than Ds has remained relatively steady over the last four years. It’s usually just the brave few. Twisting dismounts were not in fashion this year, part of an overall trend toward double saltos and away from twisting skills. 

          FLOOR EXERCISE: 

          • Here, we see more of the trend away from twisting skills. Look at the numbers for the triple full, 2.5, front double full, and even 1.5 (which should still be pretty useful for that indirect connection bonus). They’re all way down from 2012. Much of the reason is that the highest tariff elements (what, am I British now?) are double saltos, not twists, so if you need to push the difficulty to try to stay somewhere in sight of Simone, you’re going to learn a DLO, not a 3/1.
          • In keeping with that, the overall adjustment this year was toward difficulty as everyone competes with each other to show off the biggest D. More piked full ins, double arabians, double fronts, etc. 
          • On a positive note, the proportion of people fulfilling the forward tumbling requirement with some random aerial in the middle of nowhere was down to 37.5% this year from 50% last year. So yay?

            • Now here, Wolf Turn Nation is really in full bloom. The inexplicably D-rated double wolf is skyrocketing over all the other floor turns because it’s so much easier to complete and get credit for than the double L or double Y. The double wolf turn on floor really needs to become a C in the next code. 
            • Fewer single full turns reared their heads this year since the value of those D turning skills is too irresistible to pass up. Nearly no one is willing to thrown in a full turn just to fulfill the requirement before calling it a day anymore. We did, however, have a lot more people utilizing the turn combination for CV this year, which accounts for the increase in single L turns.  

              • We’ve seen a really huge leap in the number of switch fulls being attempted since the beginning of the quad, which was particularly evident this year and is directly linked to the new rule that allows gymnasts to leap out of another leap and receive credit for the first leap, even if it’s (somewhat) under-rotated. That means that nearly everyone who used to play it safe and do a switch 1/2 now feels the freedom to try the switch full, even if she can’t actually do it. So really, the most popular dance element this year wasn’t the switch full, it was the switch 9/16ths directly into a BS sissone. What was actually being attempted was really hard to say most of the time. There were a lot of “Was that a try?” moments.
              • This year, 87.50% of gymnasts elected to perform an A dance element directly out of another dance element, while just 56.25% elected to perform an A dance element directly out of a tumbling pass in an attempt to get that CV. 
              • Also, L hops. No.