Lake Placid Ironman 2023

Comparative metrics across Personal and Professional Swims

I was able to calculate stroke rate and body angles from videos of my own swims v a professionals: I used MediaPipe's (python package) pose module to extract wrist, hip, elbow, and shoulder positions via an x and y axixs and the associated coordinates can be tracked over time

From this we can see the stroke rate for the professional is much more efficent per stroke than myself - their body angle was roughly 40 degrees more perpendicular than myself (something I really need to work on)

Metric Swim 1 Swim 2 Pro Swim

Stroke Rate 94.0 84.4 74.6

Average Body Angle 129.7 130.0 162.8

Duration 13.6 13.3 30.0

The following graphs go more into detail of the pro's v my body angles

Examination of body angle changes during swim sessions

My body angle is consistently to low - you can see the pro is much more stable than I am at keeping hip's and shoulders consistently in line with the water

Examination of wrist angle changes during swim sessions

This shocked me the most - the wrist angle graph of the pro's looks almost like a heart beat (except for the camera angle change in the second half of the video)

Meanwhile my wrist angle is inconsistent - good to note for the future technique sessions to focus on this

Analysis of DPS (Distance Per Stroke) and SWOLF (Swim Golf) trends over time

All graphs from this point on were captured via FORM swimming goggles - These changed the game in terms of performance in my training swims

My Distance Per Stroke has a clear upward trend as my training progressed and I could feel myself getting more efficent as time went on

SWOLF = Time (seconds) + Stroke Count

Similar trend in my SWOLF where a downward trend indicates that its taking fewer strokes to cover the same distance

Notable peaks and valleys can be attriibuted to the fluctuations of each swim sessions focus (sprint sessions usually result in a higher SWOLF due to disproportionate increase in stroke count)

Correlation of calories burned with total swimming distance

Analysis of stroke count changes throughout training

Duration and Avg Stroke Count/Length exhibit inverse relationships in some areas. For example, longer durations (indicating longer swims) sometimes correspond to lower stroke efficiency

Breakdown of swim intervals by distance and stroke count

Key frame comparison for form analysis (session 1).

Key frame comparison for form analysis (session 2).

Key frame comparison for form analysis (session 3).

Key frame comparison for form analysis (session 4).

Overview of the distribution of biking activities throughout training

These are speciifc wahoo workouts that can be done via bike trainer - You can see I liked to test my power metrics often with the 4DP test

Specifically the 4DP test is meant to measure NeuroMuscular Power (NM), Aerobic Power (MAP), Functional Threshold Power (FTP), and Anaerobic Capacity (AC)

Analysis of average power output during biking sessions

Figure shows a steady increase that suggests improvements in power output, likely due to consistent training and adaptations

Correlation matrix showing relationships between various biking metrics

Being that there were so many metrics to stay on top of I created this correlation matrix to figure out which ones were relative enough to each other so that I didn't have to think of each one independently

This was essentially letting me know that normalized power is a little more relevant to track in comparison to average power

Cumulative Training Stress Score (TSS) over the training period

TSS was an intersting new metric to capture and be aware of - I normally feel like I have a good sense of how fatigued my body is but this level of detail helped in knowning when to ease back

Distribution of various biking metrics such as speed, power, and cadence

Analysis of session duration distributions across different biking workouts

Progression of intensity factor (IF) over time to assess training load

Bike IF and Run IF cluster around 0.7 to 0.9, suggesting most sessions were in the moderate-intensity zone

High-intensity spikes indicate effective inclusion of harder efforts for threshold development and or VO2 max improvement

Distribution of Training Stress Score (TSS) across different sessions

Trend of TSS and IF over the training period

Siimilar sort of trend here where a stable TSS over time is dicating most of the training period and then some higher days for harder training days

Weekly distribution of training load, indicating volume and intensity

Nice to see that I was increasing training load as the weeks progressed

Comparison of key metrics across different workout sessions

A portion of my workouts were captured via apple watch and another portion by wahoo - overall I think wahoo will be the best bet moving forward for tracking biking data

Comprehensive analysis of running metrics, including speed, endurance, and pacing

Being that I feel very comfortable with my ability to self-train my runs I wanted to take a deeper look at somethings that I haven't before

1. Calories burned per Mile vs. Pace: A general trend emerges: faster paces (lower min/mile) tend to correlate with lower calories burned per mile, while slower paces see higher calorie burn per mile & Outliers exist at very slow paces (>15 min/mile) with exceptionally high calorie burns, potentially due to walks, steep elevation, or strenuous conditions

2. Calorie Efficiency Over Time: Significant variability across sessions, with clusters of higher calorie burn during earlier periods (e.g., Jan 2023) and lower values in recent sessions - Larger bubbles (longer runs) generally align with moderate calorie burn per mile, suggesting standardized running efficiency during these sessions

3. Impact of Temperature on Calorie Burn: Calorie burn per mile seems relatively stable across most temperatures but shows slight increases in colder conditions (below ~50°F)

4. Impact of elevatioon Gain on Calorie Burn: Higher elevation gains (>400m) correspond to significantly higher calorie burn per mile, as expected due to the added strain of climbing

Tracking heart rate and cadence variations over time during runs

Something I learned was that during these very long distance events you want to have a faster cadence being that it means your foot is striking the ground for less duration of time and this is better for accumulation of fatigue/normal wear and tear of joints

However my cadence is usally a steady 130's-140's during these runs - I was adviced that 180 is a good number and tried it once and it just felt wrong

Analysis of pace trends and distance over the training period

Nice graph to show my pace given a distance - You can see the two outliers - one being a really long race where I kept a decent pace (~11min for 50k) and a very short run where I tried that 180 steps per minute cadence

Weekly mileage progression

Analysis of macronutrient intake distribution over the training period

I had a fairly similar diet during training as I did during normal day to day life - Granted I also ate much more after days that were specifically brutal (Chipotle was my go to)

Overall my diet served me well and was something I didn't have to think about much during training since I was already used to tracking macros

Correlation between weight changes and calorie intake over time

My weight decreased steadily over time and my calories remained pretty much the same - only thing to notehere was that the days leading up to the event I started to carbo load which was a massive uptick in carbohydrates

Detailed analysis of sleep patterns (going all the way back to 2020

1. Sleep Quality vs. Duration: Sleep quality peaks around 8 or 9 hours, but there is diminishing return beyond that - There are a few outliers with low sleep quality even at longer durations, possibly due to disturbances like snoring or movement

2. Sleep Patterns Over Time: Graph suggests steady improvement or consistency in sleep patterns since 2020 - some improvements I made were using black out curtains to get rid of excess light

3. Movements per Hour vs. Sleep Duration: Movement per hour remains low for most durations, with a few extreme outliers showing very high movement (>400 movements per hour) - likely just restless nights where I was actually awake

4. Snore Time vs. Sleep Duration: Snore time is highest in the 6 to 8 hour duration range, indicating a possible pattern of snoring during optimal sleep durations - didn't even know I snored this much, crazy.

5. Sleep Quality Distribution: There are some nights with low sleep quality (less than 0.5), which are less frequent but good to keep aware of

6. Bedtime vs. Sleep Quality: Sleep quality is highest for bedtimes between 10 PM and 12 AM - Bedtimes after 12 AM show a slight decline in quality, likely due to shorter sleep durations or disruptions to circadian rhythms