2024 US Nationals MAG Discussion

[MrSixkiller]

It would take 2 of the 4 AA'ers being unable to compete for Nedoroscik's limitation to become an issue, and at that point the team would be highly unlikely to medal anyway, regardless of whoever the 5th member is. The more pressing matter is the team needs to bring their highest scoring potential to the Olympics in order to compete with the rest of the world. With Nedorscik on the team they actually have an outside chance for Gold if they have the competition of their lives.

agreed. Putting a backup AA athlete in the 5th spot is a team that is playing for 4th or 5th place.

 

[OnoNoKomachi]

USA won't leave home the highest scoring team just to have a back up AA athlete instead.
That isn't how the selection criteria work.
It might come into play if the highest scoring team out of 4 events is different the highest out of of 3 out of 4.

The USA has been worried about not having the difficulty to compete, so I think there is zero chance of Nederocsik being left off the team because they want another AA person instead, as long as he gets the scores. I predict they will go with the highest total difficulty team if they have to make that decision.

He has to get the scores in the rest of the selection events.

 

[Denn]

It could actually end up being Hoopes.

You really think so?

 

[OnoNoKomachi]

You really think so?

No.

It's just not impossible.

Hoopes finished ahead of Nedorocsik at Winter Cup. Look how close these are.

I wouldn't bet on it (well I would if I was playing with money because I would get great odds), but it's not out of the realm of possibility.

The more recent results have Steve a bit further ahead of Hoopes, and Nedoroscik beat Hoopes handily in Baku. But they are still fairly close. One big mistake and Hoopes could be the one.

Brandon Dang was awfully close. Not sure they would have invited him as a second PH guy because he's not as experienced as Hoopes.

 

[rlayt]

Trials will be exciting! We will see who comes out ahead. Maybe Nedoroscik can hit twice in a row and reassure us all.

 

[OnoNoKomachi]

Steve has a good chance for the team if he does that. As pointed out by others, he has one of the highest potential scores.

 

[Aeris]

Hoopes would be a disappointing selection, his top score is .5 lower than Nedoroscik's and he has so much less experience with Olympic-level pressure. It is possible Hoopes can be the algorithmic selection though depending on what happens at trials, a 14.9 on Pommel is still adding more to the team than the Rings/Vault guys.

 

[OnoNoKomachi]

Some NBC style graphics for the first two nights

 

[OnoNoKomachi]

I combined ChatGPT with a spredsheet of the average scores from the first two nights.
What I got back was not crazy once the computer finally knew what to do. It used Python, and I'm no expert but the code did not look nuts and the answers made sense.

This is probably exactly how the team will be chosen...someone entering values into what is perhaps Python code - this is why it's so fast, NOT because "the only way they could pick the team so quickly is because they knew it in advance." It can be really quick.

Moving right along -
This code follows these steps:

1. Load the spreadsheet and the data from the 'Average' sheet.
2. Calculate the overall average score for each gymnast.
3. Define a function to calculate the total score for a given team of gymnasts.
4. Generate all possible combinations of 5 gymnasts.
5. Calculate the total score for each combination and track the highest score and the best team.
6. Display the highest-scoring team and their total score.

The result is that the highest-scoring team consists of Malone, Richard, Moldauer, Young, and Nedoroscik with a total score of 260.525.

Here are the next four highest-scoring combinations of gymnasts along with their total scores:

1. Combination 2
   • Gymnasts: Malone, Richard, Moldauer, Young, Hoopes
   • Total Score: 260.15
2. Combination 3
   • Gymnasts: Malone, Richard, Moldauer, Wiskus, Nedoroscik
   • Total Score: 259.775
3. Combination 4
   • Gymnasts: Malone, Richard, Young, Whittenburg, Nedoroscik
   • Total Score: 259.675
4. Combination 5
   • Gymnasts: Malone, Moldauer, Young, Whittenburg, Nedoroscik
   • Total Score: 259.675

All of these teams contain a PH specialist, which makes sense as these gymnasts can generate 15s or close to them, which is not possible for most of the others on any event.

Nedoroscik didn't even actually hit both nights. This isn't the most stable scenario, as it's tough to be consistent on PH. Even the best people fall.

Combination 3 isn't one I've seen anyone else come up with. Team 5 would suck on HB. 4 isn't great, either.

Combination 3 showed up because Young was replaced with Wiskus, and the only reason this happened is because Young bombed PH on night one, so his average is misleading. Young doesn't even show up on the PH lineup of the highest scoring team. Based on the numbers from 2 days, that makes sense. Based on our knowledge of the gymnasts involved.

Our robot overlords agree that Malone is a lock, with Richard, Young, and Moldauer not too far behind.

None of these are totally nutso.

 

[Denn]

An even better team selection model would run thousands of simulations of prelims and finals based on multiple past data points for each gymnasts for each apparatus to determine the best team — where the criteria is "score the highest total possible" in both phases.

It's not that hard actually, but I don't know if ChatGPT could do it for you, and I don't have the bandwidth to do it myself. (Sorry)

 

[QuietColours]

I saw someone on Reddit running Monte Carlo analysis of the WAG teams:

 

[OnoNoKomachi]

An even better team selection model would run thousands of simulations of prelims and finals based on multiple past data points for each gymnasts for each apparatus to determine the best team — where the criteria is "score the highest total possible" in both phases.

That would be better. But where are those other data points going to come from? I have a bunch it's USAG, NCAA, and international mixed together. These are sort of comparable, but not totally. Chat GPT could do it if it feels like it. I've seen it give up a few times.

USAG is only using data from Nationals and Trials. Everyone is in front of the same judges.

They will do calculations for prelims. I didn't bother with that.