The analysis was conducted on the results of the 2023 men's quarterfinals, as published on crossfit.com's leaderboard. Webscraping techniques were employed to gather the data, with a total of 7556 participants; however, only 827 were analyzed. These were the ones who managed to complete all the WODs, and their data was recorded correctly.
The WODs performed were as follows:
3.1. The following graphs show the relationship between age and the WOD results, showing no correlation, a similar result is found with the athlete's height and weight.
https://preview.redd.it/y3h2nxs7wvx...ed&s=6e22b096af233cec5ba2e8b5740b05885ea85bfd
Coefficients: [[3.00, 17.86, 1.49, 85.58, -205.70 , 37.56, -226.58, 177.47]]
Intercept: [659.62]
The WODs performed were as follows:
- Wod 1 (W1T) Time cap: 15 minutes 9 front squats, weight 1 225lb (heaviest) 9 handstand walks, 25 feet 15 front squats, weight 2 185lb 15 muscle-ups 21 front squats, weight 3 135lb (lightest) 21 chest-to-wall handstand push-ups
- Wod 2 (W2) 12-minute AMRAP 70lb: 8 dumbbell snatches, arm 1 8 overhead walking-lunge steps, arm 1 8 dumbbell snatches, arm 2 8 overhead walking-lunge steps, arm 2 40 crossovers
- Wod 3 (W3T) Time cap: 10 minutes 5 rounds for time: 5 burpee box jump-overs 1 clean and jerk *Add 1 clean and jerk after each round. ♂ 275lb clean and jerks, 30-in box
- Wod 4 (W4) 20-minute AMRAP 1,000-meter row 50 GHD sit-ups 500-meter row 25 V-ups
- Wod 5 (W5T) Time cap: 15 minutes 21 deadlifts, weight 1 225lb (lightest) 21 chest-to-bar pull-ups 15 deadlifts, weight 2 275lb 15 bar muscle-ups 9 deadlifts, weight 3 315lb (heaviest)
- Histograms for numerical variables: Histograms were created for the main features, showing that the participants' characteristics are normally distributed, with the most common age being 28, from East North America, and weight between 87 and 89 kg, with height 177 – 179 cm.
- Box plots for categorical variables: It could be inferred that 75% of the athletes are 31 years old or younger, with a height of 180.3 cm and weight of 92 kg. Regarding WOD performance, the median times for WOD 1, 2, 3, 4, 5 were 10.9 minutes, 376 reps, 7.78 minutes, 510 reps, and 6.6 minutes, respectively.
- Analyzing relationships between variables (scatterplots, correlation): A correlation matrix was created, highlighting that the athletes' ranking position does not show a significant level of relationship with age, height, and weight. The most significant correlation, at 69%, is between weight and height.
3.1. The following graphs show the relationship between age and the WOD results, showing no correlation, a similar result is found with the athlete's height and weight.
https://preview.redd.it/y3h2nxs7wvx...ed&s=6e22b096af233cec5ba2e8b5740b05885ea85bfd
- Analysis of athletes by region:
- North America East: 13.5
- North America West: 16.5
- Europe: 20.9
- Oceania: 121.2
- South America: 121.9
- Asia: 195.1
- Africa: 564.8
- Europe: 547
- Oceania: 642
- North America East: 774
- South America: 808
- North America West: 831
- Africa: 858
- Asia: 814
- Bonus modeling: Using a Support Vector Machine model with a 99% score, it is possible to predict the ranking using the following independent variables: age, height, weight, and WOD results. To make the model more practical and obtain coefficients at the expense of the score, a linear model was employed with an 87% score. The following coefficients and intercept were obtained, which can be used to calculate an athlete's ranking if they haven't competed or to review their current ranking based on their latest WOD results:
Coefficients: [[3.00, 17.86, 1.49, 85.58, -205.70 , 37.56, -226.58, 177.47]]
Intercept: [659.62]