Aerobic Tracking

[Bogdan_Rylski]

I have an idea for a graph that would track W/HR ratio over time but also have the ability to include data based on the power zone. For example, include only Z2 efforts of 20min or longer. I find myself going back and forth between workouts and using the slider tool on the aerobic chart to get W/HR for my endurance rides, so it would be nice to have it all in one place.

There is a similar graph on Wko4 but the graphics are painful to look at, and I’m not sure if there is any criteria for data inclusion.

[david]

I am also keen to see W/HR over time as a measure of fitness. Interesting idea to do it only for Z2 efforts of more than a certain duration. I would need to pre-compute this data for each ride so need to know the criteria in advance. I could use the HR lag adjusted 1 minute data points used for the de-coupling chart. These are already cleaned up quite a bit with messy points tossed out.

Any other criteria other than the zone and duration? Maybe only consider work done in the first hour of the ride before getting hot and thirsty etc?

[Bogdan_Rylski]

@david

Any other criteria other than the zone and duration

Possibly tossing out intervals that exceed a certain amount of decoupling to account for fatigue? Not sure if this is a valid measure.

@david

Interesting idea to do it only for Z2

Z2 seems like the obvious choice but there might be some interesting information hidden in other zones, thus the idea to be able to include/exclude individual zones.

[david]

@Bogdan_Rylski

Possibly tossing out intervals that exceed a certain amount of decoupling to account for fatigue? Not sure if this is a valid measure.

Thats what I was hoping for with only using data from the 1st hour but your method will work better. I have the decoupling data.

[david]

@Bogdan_Rylski

There is a similar graph on Wko4

Please post a screenshot if you can. I don’t have WKO4/5.

[Bogdan_Rylski]

[david]

So all the blue dots are the power / HR ratios for each ride and the dotted line is the trend?

I am trying to figure out how to do something like this. Preferably by selecting and plotting “good” values for each ride so I don’t have to fill the screen with dots and plot a trend. I am starting with the 1 minute data points used to produce this plot:

These are already quite cleaned up and adjusted for HR lag. In addition I can keep only the ones with 60s of moving time. And then group them by zone. Initially I am looking at Z2 (169-230w). Here is a set of these points from one of my rides (a random hills and stuff effort). The index is the minute of the ride (zero start) and time is just that in secs:

index,time,watts,hr,ratio
5,360,201.0,125.0,1.61
7,480,175.0,137.0,1.28
10,660,176.0,140.0,1.26
11,720,181.0,136.0,1.33
12,780,219.0,141.0,1.55
13,840,173.0,136.0,1.27
14,900,171.0,135.0,1.27

https://pastebin.com/CbPTinkS

I want to try come up with an algorithm to select which points to use to create a “Power / HR ratio” value for Z2 for this ride.

• Exclude warmup (20m on this ride)
• Exclude any cool down

Other ideas:

• Only keep points from the 1st hour to avoid any decoupling effects?
• Only keep points that are part of consecutive sequence at least x minutes long (5 mins+?)
• Only use points around the middle of the zone (±5w, ratio at bottom is much lower than top)? What if ride doesn’t have any points near the middle? Don’t assign a value?
• Just average the whole lot? Or everything from the 1st hour?

[Bogdan_Rylski]

@david

So all the blue dots are the power / HR ratios for each ride and the dotted line is the trend?

Blue dots are efficiency factor which is “Normalized power (NP) or Normalized Graded Pace (NGP) divided by average heart rate for a workout or a selected portion of a workout”.
Red dots are PW/HR.
Dotted lines are trend.

[Bogdan_Rylski]

@david

Only keep points that are part of consecutive sequence at least x minutes long (5 mins+?)

My logic steps would go like this:

1. Find all intervals longer than (time) with average power in (zone) defined by (watt_range).

2. Calculate and define decoupling for given intervals.

3. Exclude intervals with decoupling exceeding (percent)

4. Calculate average PW/HR for intervals(1) that meet criteria(3)

5. Save the best PW/HR to the day it was achieved and plot [Date, PW/HR]

[Bogdan_Rylski]

A couple of things have come to mind on this topic as I have observed my cardiac response more closely with the Sieler tool. As I progress through the loading phase heart rate gets pushed down at the same effort level until it gets to the point where we might recognize that we are fatigued (cant get heart rate up). Now on the recovery week, I have noticed my HR climbing towards the top of the endurance zone and exceeding it while doing nothing more than a ride at 60-70% FTP.

These relatively short term variations undoubtedly have an effect on the display of long term aerobic tracking. They may have some value on their own (another thing I want to look into) but they also should be accounted for in a tool that aims to track aerobic efficiency over longer periods of time.

[Bogdan_Rylski]

With the new custom charts I have made an efficiency tracking chart:
Bars are CTL
Red is CIL
Blue is Efficiency on Fitness average

Next step is to filter the NP/HR

[Bogdan_Rylski]

@Bogdan_Rylski

My logic steps would go like this:

1. Find all intervals longer than (time) with average power in (zone) defined by (watt_range).

Just thought of this: An additional criteria should be variability index.

[david]

I have done some work on this:

Power/HR in Z2 metric

Intervals.icu now calculates power / HR for time spent in heart rate Z2 for each ride. This is the average watts / average heart rate for the parts of the ride done in HR zone 2 (aerobic or endurance). The ride is split up into 1 minute sections adjusted for HR lag. Only minutes in HR Z2 with cadence between 85 and 100 and little decoupling* are kept. The power/HR Z2 ratio is the average of these minutes and an indicator of aerobic fitness. [Screen Shot 2020-05-27 at 20.10.14] You can add this…