Intervals.icu interval detection and training tool

[davidtinker]

David you got a mention on the TrainerRoad podcast this morning by the way.......

You mentioned you were going to look at customizing the min time used to calculate est FTP. Would it be possible to have an option where you can see the est FTP per ride as well. Could be useful to have this figure when for instance you do a FTP test on TrainerRoad to reference it with the est FTP for that session. Always good to have more than 1 estimate to work from

 

Wow I didn't know about the TR podcast. Tx! Do you have an idea how far in? I see the latest is nearly 2h long.

 

I will add an option to show the estimated FTP even if it is lower than the running estimate with decay.

 

This chat stuff has been sucking a lot of time. Once thats out the way its back to implementing analysis features. Still need to do group chat + sharing activities. Hopefully this weekend for at least group chat.

[Jaco-fiets]

Wow I didn't know about the TR podcast. Tx! Do you have an idea how far in? I see the latest is nearly 2h long.

 

I will add an option to show the estimated FTP even if it is lower than the running estimate with decay.

 

This chat stuff has been sucking a lot of time. Once thats out the way its back to implementing analysis features. Still need to do group chat + sharing activities. Hopefully this weekend for at least group chat.

David it is within the 1st 30min of the podcast if I remember correctly

 

Thanks for letting us share ideas and considering it

[madmarc]

My chart does not look like yours. I can see my model fit but there is no decoupling % displayed. Any settings I need to change?

 

Just to mention - decoupling is a measure done at Aerobic Endurance - I do 2hr + at 60 - 70% of FTP - My goal is to get a constant decoupling of below 5% before i go into build and specialty phase on TR.

 

I could never hold a constant FTP increase before, after a lot of research its because my Base was just not there, so it's back to basics of 12 week base training.

 

This is where Intervals.ICU is great as you don't have to get into a complicated calculation to work out your decoupling

[davidtinker]

You mentioned you were going to look at customizing the min time used to calculate est FTP. Would it be possible to have an option where you can see the est FTP per ride as well. Could be useful to have this figure when for instance you do a FTP test on TrainerRoad to reference it with the est FTP for that session. Always good to have more than 1 estimate to work from

 

I have added the estimate for the ride to the help info displayed when you click on the "Est. FTP" number. As you can see I have some work to do before the PPA Sportive #5 on 1st September! If I don't get my FTP up to 310w its not going to go well.

 

[madmarc]

Hi David

 

I see my weight doesn't update on each activity - I changed it in settings as well on STRAVA but it still reflects the old weight on new activities.

[davidtinker]

Hi David I see my weight doesn't update on each activity - I changed it in settings as well on STRAVA but it still reflects the old weight on new activities.

 

Hmm. I will have a look. If you change it on your most recent activity or in settings it should apply to new activities. It doesn't track Strava changes (yet .. thats planned for people with scales that update Strava),

[madmarc]

So it seems my Aerobic Endurance training plan is paying off with my decoupling becoming more and more consistently being under 5% - I don't worry about the model fit because i don't really understand what its reflecting or measuring.

 

Thanks to David and all the work you put into this add on, its one of the best analysis tools i've found

 

 

[Jaco-fiets]

David can you explain the "model fit" please?

[RyanAW]

So it seems my Aerobic Endurance training plan is paying off with my decoupling becoming more and more consistently being under 5% - I don't worry about the model fit because i don't really understand what its reflecting or measuring.

 

Thanks to David and all the work you put into this add on, its one of the best analysis tools i've found

 

 

David can you explain the "model fit" please?

 

Can you also please explain what negative decoupling means? I've seen this on a few occasions.

 

Thanks again!

[Jehosefat]

Can you also please explain what negative decoupling means? I've seen this on a few occasions.

 

Thanks again!

 

Negative decoupling would happen when you output more power at a given heart rate in the second half of the ride versus the first half.

 

For me, when this has happened, it has usually been because the first half of the race had some fast, sketchy sections (or skittish groups) that didn't require much power but got the heart rate up. Then things calmed down and/or required some proper power outputs in the second half.

[davidtinker]

The "model fit" is the R-squared value for the line running through the points. Its a measure of how far away from the line the points are. I might get rid of it because "poor" model fit doesn't mean you can't use the decoupling number and it causes confusion.

[Yyyy]

Just out of curiosity

 

I have been using trainingpeaks, Garmin connect and now intervals.icu to track my fitness progression over time and whether a session was effective or not.

 

Always had a difference between Garmin and trainingpeaks (whether its TSS, or measuring my Zones – power and hear rate)

 

Given I am a bit of a hybrid athlete that dabbles in Cycling/MTB, running, hiking, swimming, OCR’s and weighlifting. Not all measures or TSS scores are comparable. (also I only ever use a power meter on the wattbike)

 

However I was expecting intervals.icu to be very closely aligned to trainingpeaks.

 

Why would there be a difference between these two, specifically when it comes to TSB, ATL and CTL?

[davidtinker]

Training load (TSS) should be the same or very close for rides with power data. There might be some differences in how TP deals with drop outs, spikes and so on but I wouldn't expect much, unless the power data is very iffy. 

 

Intervals.icu uses previous rides with power and HR to estimate training load for activities with heart rate only. I don't know what TP does in that case.

 

If the training load numbers are close then ATL, CTL and TSB should also be close. Intervals.icu uses the standard 7 day and 42 day exponentially weighted moving averages for ATL and CTL.

Edited July 26, 2019 by davidtinker

[Vetplant]

Just out of curiosity

 

I have been using trainingpeaks, Garmin connect and now intervals.icu to track my fitness progression over time and whether a session was effective or not.

 

Always had a difference between Garmin and trainingpeaks (whether its TSS, or measuring my Zones – power and hear rate)

 

Given I am a bit of a hybrid athlete that dabbles in Cycling/MTB, running, hiking, swimming, OCR’s and weighlifting. Not all measures or TSS scores are comparable. (also I only ever use a power meter on the wattbike)

 

However I was expecting intervals.icu to be very closely aligned to trainingpeaks.

 

Why would there be a difference between these two, specifically when it comes to TSB, ATL and CTL?

 

 

Training load (TSS) should be the same or very close for rides with power data. There might be some differences in how TP deals with drop outs, spikes and so on but I wouldn't expect much, unless the power data is very iffy. 

 

Intervals.icu uses previous rides with power and HR to estimate training load for activities with heart rate only. I don't know what TP does in that case.

 

If the training load numbers are close then ATL, CTL and TSB should also be close. Intervals.icu uses the standard 7 day and 42 day exponentially weighted moving averages for ATL and CTL.

Yaseen, just check that the FTP settings per workout on the various sites are accurate, as that has an effect on the TSS per workout, which then snowballs into ATL,CTL and TSB.

 

Sequence of checks

Intervals.icu FTP vs Training peaks FTP? 

Intervals.icu LOAD vs Training peaks TSS?

 

You should find the source of discrepancy here...

[Jehosefat]

Intervals.icu uses previous rides with power and HR to estimate training load for activities with heart rate only. I don't know what TP does in that case.

 

As far as I can tell TP uses the heart rate zones that you set in your user settings (usually some model based on max HR or LTHR) to calculate what they call "hrTSS".

Edited July 26, 2019 by Jehosefat

[Yyyy]

Thanks guys 

 

Did some checks on my side

 

Last 3 outdoor rides, no power meter

TP TSS: 181, 51, 224

Intervals.icu training load: 179, 55, 252

 

Last 3 indoor wattbike sessions using power

TP TSS: 70, 56, 50

Intervals.icu training load: 61, 32, 49

 

When comparing CTL on TP only for Bike and MTB between vs intervals.icu there is an 11 point difference and a 37 point difference when compared to taking my total CTL for all activities on TP

 

FTP difference between the two apps is only 1watt so very close.

 

Had a look at other activities which is purely on heart rate data and notice big differences in the weightlifting side.

Last weightlifting session, difference was 27 points (intervals.icu giving less than half of what TP gives for lifting).

 

Heart rate zones on TP was set more than two years back but have set it to automatically adjust based on training.

 

Just looking at the HR based cycling data, the two apps are closely aligned whereas with the power based indoor cycling the differences are a bit bigger than expected.

 

I was trying to compare my runs and swims but anything before 28 June doesn’t have a load assigned to it…..could this be that intervals.icu was still getting to know me? (did sign up only about a month and a bit ago)