How Vault is Scored

[KristyJ]

I'm not a judge... But I think judging more holistically from perfection to absolute slop per phase makes sense. If preflight, block, post flight, amplitude, and landing all start out of 1.0 execution per vault, you just take each phase in turn and judge how good it was out of 10 tenths, max E score is still a ten for both vaults combined. And you can't lose more than a point per phase. I'd think of it as "from 0-10, how good was that phase?" and call it a day. Use the full range from 0-1.0 off per phase, (so you can't lose more than a point per phase).

Examples. Somebody like paseka should be losing .5+ for her pre flight alone. The one armed chengs could get the full blocking point off for not blocking appropriately (and now you don't have an arbitrary 2 point deduction), whereas people like Andrade or Simone are close to deduction free.. twister vaults could be .3-.5, or worse if you're bending your legs ... Little shuffle or lean on landing? .1 off. Step? .2. Chest down and step? .5ish. Roberson drilling a twist into crunched knees? .7. Hand / knee down? .8 off. Can't land it at all? Full point. Just, use the full range of execution as it relates to each phase.

There are absolutely vaults with well more than a point of deductions in the post-flight (leg separation/knee bend/toe point/hip angle/amplitude/height/distance etc) and it would be a shame if those couldn't be sufficiently differentiated from others that are clean.

 

[dweiss324]

^Well in my hypothetical example, I separated amplitude (ie height and distance, maybe even direction) for its own point worth of deductions. Plus if you had to do two vaults, and both had garbage post flight, you'd essentially be counting those deductions twice. Is anybody getting anywhere close to 2 points of deductions just in post flight form? Not even counting amplitude or landing? I would be surprised if yes. If gymnasts were counting a half point of deductions on post flight and another .5 on amplitude for each vault, and then accruing landing deductions for each, you'd be down to around a 7.0 E without even considering preflight and block. Who gets 7.0 E scores on vault? I think it would be fairly accurate in differentiating Paseka vs Andrade or Simone throwing their usual program if applied well.

Paseka: 5 tenths pre , 2 tenths block, ~3 tenths post flight, 3-5 tenths landing, 2 tenths amplitude... Girl loses 2 points in E per vault, being generous. I don't know. Give her 6.0 E, add the D and vault differentiation bonuses.... I don't see why that wouldn't be fairly effective in judging against a fairly clean DTY + Lopez combo. Or if Andrade or Simone were accruing less than a point total in E deductions over 2 vaults.

 

[Denn]

That's related to what I was trying to get at earlier with the idea of breaking a vault into a few elements for the D-score, too. They more or less do this already with execution, and it's already true that most phases can't be deducted more than 1.0, though it's a little more in a couple cases. So you could be onto something there.

Here's the full list of deductions with some mark-up. The First-flight, body position (in the air), and landing match 3 of your categories, so I didn't label them. The pink is what you called "amplitude" I think, yes? The dark purple is probably your "block" (I included the snap).

I do think that having two vaults would allow for a little bit simpler of a breakdown with the deductions. And the body position deductions could be the same as everywhere else.

 

[Denn]

MAG uses this diagram. Notice height and distance aren't listed, so that would be your 5th "element" @dweiss324 .

BTW, I actually think that the "preparation for landing" deduction is unnecessary. If you land upright, not twisted, and stuck, you prepared enough. I'd be okay if it said "no kickout / flare-out" as a 0.1 deduction to reward people who DO this, but 0.5 off for not preparing to land is always redundant to other deductions.

I also think that, consistent with what many of us say about preferring .1, .2, .3, .5 deductions instead .1, .3, .5, that would work just fine on vault with 2 vaults.

 

[FrolovasDoubleLayout]

E-jury also is supposed to enter deductions per phase, though depending on the event and the scoring system in use, this is not always the case. It is the same for artistry and execution - at major events, they are entered separately on the tablet - hence why artistry can be seen in some results books, At some events, judges are instructed to enter them as a final group (art. 0,4 + ex. 2,1 would be entered under execution as 2,5).

 

[Denn]

Hi all — 25 posts in, it's clear this is no longer just about apparatus finals Day 1, so I spun it out into a new thread.

 

[dweiss324]

I ran my proposed system through chat gpt and fed it a YouTube video of event finals for 2025 worlds, which it can analyze frame by frame, with instruction to take holistic phase deductions, corroborated by textual commentary found in tweets and bbs and (something else). I cannot believe that's possible.

Anyway, I havent rewatched the vault final but it has vaelen coming in first slightly ahead of melnikova because this hypothetical system prioritizes sound technique for her rudi and tsuk full and relatively clean execution, over chucked difficulty. Vaelen - melnikova - roberson - LMF in that order

I haven't rewatched the final yet to agree or disagree, but it's giving me a full list of it's interpreted breakdowns for E for each vault and then compiling them with the halved D scores. It's an interesting alternative take to be sure, and I'm very curious what it would do with past EFs.

 

[KristyJ]

I've always judged vaults by phase, as well, I just felt it was unnecessarily prescriptive to be like "each phase is one point."

 

[Aeris]

That AI result sure is something. It might be helpful to have a computer provide height/distance numbers, but I wouldn't want to use it for much else.

 

[KristyJ]

I don't mind the computer providing some data (i.e. height in comparison to gymnast foot height at the point of repulsion from the table or somesuch), but I wouldn't want it calling the shots. The out-of-bounds debacle in Paris showed that unreviewed data is as bad as no data.