Correlation Analysis of OPEN 24.1 and 24.2

[hmseej]

Hey everyone,

I've conducted a correlation analysis on fitness training data, examining performance metrics from two workouts (24.1 and 24.2), along with personal characteristics like height, weight, and age. Here are the correlation coefficients:

https://preview.redd.it/608qy6rqkco...bp&s=9cd37c1f3caababdd7912bf45ed05f32f52cf073

1. Time 24.1 vs. Reps 24.2: There's a moderate negative correlation (-0.46), suggesting an inverse relationship between the time of the first WOD and the reps completed in the second WOD. This implies that individuals who complete the first WOD faster tend to perform more reps in the second one.
   [*]Time 24.1 vs. Height [in]: The correlation is low (0.17), indicating a weak or nearly nonexistent relationship between the time of the first WOD and the height of the individual.
   [*]Time 24.1 vs. Weight [lb]: The correlation is very low (0.09), suggesting there's no significant relationship between the time of the first WOD and the weight of the person.
   [*]Time 24.1 vs. Age: The correlation is low (0.15), suggesting a weak relationship between the time of the first WOD and the age of the person.
   [*]Reps 24.2 vs. Height [in]: The correlation is low (0.18), suggesting a weak relationship between the reps completed in the second WOD and the height of the individual.
   [*]Reps 24.2 vs. Weight [lb]: There's a moderate positive correlation (0.25), indicating a positive relationship between the reps completed in the second WOD and the weight of the person. This implies that heavier individuals tend to perform more reps in the second WOD.
   [*]Reps 24.2 vs. Age: There's a moderate negative correlation (-0.17), suggesting an inverse relationship between the reps completed in the second WOD and the age of the person. This implies that younger individuals tend to perform more reps in the second WOD.
   [*]Height [in] vs. Weight [lb]: There's a moderate positive correlation (0.64), indicating a positive relationship between the height and weight of the person. This is expected as taller individuals tend to weigh more.
   [*]Height [in] vs. Age: The correlation is practically null (-0.01), suggesting there's no significant relationship between the height and age of the individual.
   [*]Weight [lb] vs. Age: The correlation is very low (-0.08), suggesting there's no significant relationship between the weight and age of the person.
   

 

[2die]

@hmseej Put these in a regression model, standardize the variables, and tell us which matters the most.

 

[hmseej]

@2die yeah, Monday when we have the results of 24.3

 

[2die]

@hmseej I suggest converting each WOD outcome to reps / 60 seconds.

For 24.2 it's super easy: Reps / 20

For 24.1 it's Reps / 15 for time-capped individuals, 60*180/time in seconds for those who finished the workout.

For 24.3 it's Reps / 14 for those who got through 5 rounds, Reps / 14 for those who did not get through 5 rounds, and 60*70/(time in seconds - 60) for those who complete the WOD. There will be a tiny group who complete the 5 rounds in the 14:00-15:00 minutes range ... but it should be so small that you can treat them the same as Reps / 14 without affecting the model much.

 

[hmseej]

@2die thanks, I will keep in mind

 

[kebo1957]

@hmseej Throw into XGBoost and show us the shapley value bee swarm plots, feature importance plots please

 

[fayemarie91]

@hmseej Wod science posted his results showing how important height was for 24.1. Are you refuting his claim that 24.1 favoured shorter athletes?

 

[hmseej]

@fayemarie91 based on analysis of the first 10,000 results in 24.1, WOD actually favors shorter people. I saw the same result in my box. In any case, the correlation is low, 0.17, but at that competitive level small changes can make a big difference. Could you send me the link of Wod science?

 

[fayemarie91]

@hmseej From the 10-20 minute mark of this video he talks about it.

 

[hmseej]

@fayemarie91 I reviewed the video and honestly I don't know why it shows an opposite result, despite indicating that shorter people may benefit in the burpees, and then it says that the top 25% are shorter. He considers 100,000 athletes and I only considered the top 10,000, which could be the reason, although I don't believe it. And I don't know how he cleaned the data, and he also doesn't indicate the value of the correlation.

In the result I'm presenting, something curious happens: in 24.1, shorter people benefit by 0.17, but in 24.2, taller people benefit by 0.18, and I believe it's due to the rowing.