Related
This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

Low Exercise Load for long rides

sioni91
over 1 year ago

I've been using the performance metrics on the Edge 530 and Garmin Connect, using exercise load and load focus to structure training with decent results so far. I train for long brevet rides and I find the exercise load reported for these long rides specifically to be confusingly low. I ride with a power meter and Wahoo tickr fit optical wrist HR.

Here's a recent example that highlights the issue over two steady endurance rides of different durations, one week apart:

Ride Time hh:mm Average Power (W) Normalized Power (W) Average HR Intensity Factor TSS Exercise Load
1 3:45 136 147 145 0.65 157.5 167
2 12:23 126 145 150 0.63 489.1 189

Both rides were of similar intensity, and ride 2 being roughly 3 times longer than ride 1, it makes sense that the TSS is roughly 3 times higher. However the Exercise Load is only slightly higher, which seems way off. It becomes frustrating then when taking a day or two of recovery after a long event and the training load goes to "low", making it difficult to use the training load and load focus for the start of the next training block.

Last year I rode an event with 5 days back to back of this type of duration, which obviously left me feeling destroyed, but on the performance stats it looked like a recovery week!

It may or may not be related, but I have the setting enabled to detect FTP changes, and after Ride 2 I got a notification that my new FTP detected was ~3500 W (it's not). I "skipped" it, but maybe it already skews the metrics for that activity? There are no spikes in the power data so I don't know how it arrived at that number either.

Any thoughts anyone? Cheers!

  • 0 over 1 year ago

    I don't have any direct insight into your question, but I will say for any serious training I wouldn't trust any optical wrist heartrate sensor.  In my experience they read much lower than the actual indicated by a proper chest strap heartrate monitor during any type of vigorous activity.  Perhaps artificially low heartrate readings are part of the problem you are seeing.  It's very little money to get a chest strap HRM, I would start with that.

  • 0 over 1 year ago

    I think there must be a problem with the training load metric. I use a heart strap. When I did a 6k run 2 days ago, I was given a training load of 154 and I felt good.  Then yesterday I cycled over 80k, I was given a training load of 36 and I was completely exhausted.

    I should add I don't have a power meter for my bike and so if this is a prerequisite then this would ecoair the problem I have experienced.

    If training load is used to make recommendations for my next activity, I fear it won't currently be good advise for me.

    I will be of course listening to my body. But it's a shame this is clearly wrong because I have often benefited from lfollowing garmin/firstbeat recommendations

  • 0 over 1 year ago

    Hi, your results seem reasonable based on what I've seen in my testing and everyday usage. TSS increases linearly with time, however "Load" does not -- it has diminishing returns as time goes on.

    Load comes from FirstBeat's proprietary secret sauce algorithms (now owned by Garmin), so there is limited public knowledge of their algorithms, and no scientific peer reviews (last I checked). In FirstBeat lingo, Load = "Excess Post-Exercise Oxygen Consumption" (EPOC). Although the Edge 530 will not record EPOC, you can display it in a data screen and then monitor it in real time during your ride. You will see that as time goes by, the EPOC will increase more slowly (assuming steady effort).

    I did quite a bit of testing where I manually recorded the EPOC data during steady activities at different intensities and compared it to FirstBeat's sparse "white papers" {cough} about their algorithms, to gain insight into how it is calculated. I don't have access to my work/data at the moment however I recall it behaving relatively predictably based on their description, at least for aerobic efforts. The contribution of ANaerobic efforts during the activity was more unknown (they cause sudden increases). So overall their model appears to place greater emphasis on shorter aerobic activities than longer ones, with longer activities having diminishing benefits. (I'm no fitness scientist, just a data geek, but my limited understanding is this is generally accepted fitness theory.)

    Gamin's "Training Effect" also appears to be modelled similarly. Aerobic Training Effect (and anaerobic) can also be added to a data screen and monitored in realtime during an activity. Plus if you use an addon from Connect IQ called AppBuilder5+ you can record aerobic TE data to view it in Garmin Connect (but not ANaerobic TE). You will see it initially increases quickly (appears to be a hard-coded "warmup") up to about TE 2.0, then it goes at a slower rate to TE 3.0, and etc. in a linear connect-the-dots trend (wish I had some of my charts handy to post). Anyhow point is it uses the same diminishing returns model, and it tops out at 5.0 regardless of how long the activity continues.

    BTW you can also use AppBuilder5+ to calculate/record NP, IF, and TSS in realtime to view them in Garmin Connect (since those Coggan / Training Peaks formulae are publicly known).

    Hope that helps! Some of this was mentioned in past posts here, those might have more info.

    An aside about the Wahoo Tickr Fit HRM, I use it as well. Note the arm bard is worn on the forearm just below the elbow, not on the wrist.  In my testing with direct comparison to a Garmin chest strap, their data agreed very well so I was happy with its accuracy. The arm band does lag however (by about 2 seconds if I recall correctly), so the chest strap shows an increase in HR before the armband does, and similarly for HR decreases. But that isn't much of a problem for me either.

    Cheers!

  • 0 over 1 year ago in reply to Kyle J

    Thanks Kyle.

    I suppose another means of verification would be to compare rides using a power meter. They are soo expensive.  Altough an indoor trainer may be a sufficient test but doing 80k on an indoor trainer might be more than I could take

  • 0 over 1 year ago in reply to Kyle J

    Thanks for this comprehensive response. I will certainly try to replicate some of these observations like monitoring EPOC in realtime. I have also started to wonder whether pauses can contribute, if EPOC is falling during paused periods. I might do rides up to 3-4 hours without much paused time, but in a 12 hour ride can accumulate a lot more time stopped. 

  • 0 over 1 year ago in reply to sioni91

    The Load / EPOC should never drop during an activity, only increase. It represents the effort we've already expended, so it can't be "undone".

    Pauses/rests in activities do have an affect tho. If I recall correctly, it takes some time after a rest before the Load / EPOC begins to increase again. I think part of this is simply getting HR etc back into a working zone after resting. The rest I think is in how FirstBeat models it, with a sort of recovery occurring during rest / very low intensities, so Load / EPOC doesn't begin to accumulate again until at higher intensities again for a period of time.

    In any case, based on what I've learned about it, I suspect that these metrics / models simply aren't well tailored for long activities. They appear to be aimed at 2 hrs-ish or less, and maybe biased towards higher intensity activities. They seem to "top out" with diminishing returns at longer durations and low/moderate intensities, and things like Training Effect simply cap at 5.0 max.

    Recording a new separate activity after each rest stop might be an interesting experiment. Garmin does claim to track some cumulative effects between activities (for things like recovery, etc). But it might be gamey too, I really don't know. I believe there is still a lot of handwaviness in these products (by any brand), and the science isn't yet where some the marketing hype claims it to be.

  • 0 over 1 year ago in reply to imduk

    Yes the test sessions I did were on a trainer with power meter so I could maintain steady effort as much as possible. I forget now whether I was holding constant power or constant HR. I think I tried both. (It matters because as time goes by HR will naturally increase at steady power, or power needs to reduce to maintain steady HR.) It was definitely a lot of work to do those sessions and collect consistent data... A pandemic-era boredom project lol!