Training effect broken does not match heart rate zones anymore with firmware 12.10

Yesterday the watch updated to 13.0 and didn't do a hard restart (now I did). Last months I ran a marathon in April and after that took some weeks to rest my body so had little activity. I guess that's where the problem kickstarted thinking back. I saw at a certain point after some irritation about the wrong benefit logging that the max HR was way too low so I adjusted it to the figure it was before and also adjusted the threshold heart rate which was also way lower than before. 

Last week I did a factory reset without sync of past activities so the watch could build up it's algorithm again. Looked good at first with correct benefit logging (base for base runs). And after a strong interval training it put the max HR up to 188. Untill the run of today where it seemed totally off again (and with update to 13.0 yesterday). It logged threshold which is a zone 4 run for my run which was in zone 2.

Ah well I'll try the lactate threshold test run and see what that brings.

R R said:

And after a strong interval training it put the max HR up to 188.

Turn the auto-detection for HR Max off and do a field test to determine your HR Max. You can also complete a maximal but steady pace anaerobic workout (operative word is maximal effort) and continue the last interval until exhaustion. You can add 4bpm to your peak HR. Another tip is that for most individuals the LTHR is about 90-92% of Max HR.

Thanks. Based on this my max hr probably could be a bit higher. But in the past I've never seen it higher than the 188 it's currently in now. The lthr is on 173 which is also in bounds based on your indications. Maybe both could be a bit higher.

But with current indications that doesn't account for the errors in primary benefit where I ran a basic run in zone 2 with average of 148 heart rate and the Garmin labels it as a threshold run. So since about 2/3 months something changed in the Garmin algorithm where it might look at threshold speed instead of time ran in heart rate zones like the others stated. I find that still very weird because you can't check it during running (what speed to run for basic run or what speed for tempo run etc) or even afterwards like you can with running in heart rate zones which was always very clear. So it's an adjustment that is completely ridiculous.

R R said:

So since about 2/3 months something changed in the Garmin algorithm where it might look at threshold speed instead of time ran in heart rate zones

It definitely doesn’t look at heart rate zones, and never has. I seriously doubt there was a major change in the algorithm. Over the last few releases, Garmin added some dampening in the VO2 Max updates, so it now changes much more slowly after each run.

As I said earlier, the LTHR values don’t impact the training effect calculations either.

i think you have some erratic data that threw the model off. This will go away with time if you continue wearing a chest strap and use good pace data.

Many months later.... The problem still persists. Tried many different things like the test run. But somehow Garmin keeps adjusting my zones lower and lower. Today was a great example again ran fully in zone 2 with an average heart rate of 139 and it labeled it as tempo instead of base. Before all this started to happen my base ribs were with an average of 148 heart rate. If we can't trust Garmin anymore why ever buy new models.

R R said:

But somehow Garmin keeps adjusting my zones lower and lower

What are your zones based on?

When you say they are adjusted lower and lower, what does it mean? Do you see the boundaries for each zone getting lower?

Yeah, there are some options here.

With regards to HR Max detection, Garmin updated the algorithm recently, and it seems to be working well for some users. I am testing it again. Otherwise, as described above the best method for me is to run my best 5k, take the peak HR and add 5bpm for HR Max. I used this method for the 6 months or so and got 162bpm despite measuring 165bpm when I picked up training after an injury. HR Max tends to go down if you train a lot, research shows. As I turned on HR Max detection again, the watch proposed 161bpm, so pretty good. Thaking my HR Max from 165 to 161 resulted in a decrease of VO2 Max as expected by a point for running and 2 points for biking. Now my VO2 Max for biking is just 2-3 points higher than my running VO2 Max which is within the 5% of error for this algorithm.

With regards to the HR Max value, it is the cornestone of the Garmin metrics. It influences everything from Training Effect, Load, Recovery, Lactate Threshold, etc. Unfortunatey, it is not easy to measure your HR Max and users are often confusing it with the highest HR they have reached, which might be an HR Max effort or not, or might be a fluke in the data.

With regards to training zones, if your training zones are not based on HRR or HR Max but on LTHR, you have a good method for training, but there is additional risk of a misalignment with % of (VO2 Max-VO2 Rest) that the watch uses de facto for EPOC, Training Effect, etc. At least with % of HR Max, when the watch, right or wrong, changes your HR Max, the zones are aligned with the new value. With % of LTHR, you will have to wait for a new LT detection (right or wrong) to realign things. % of HRR aligns the best with the way the watch calculates its metrics, but is less user friendly if your train with HR maybe. If you train with pace or power, I suggest you use %HRR for the heart rate zones. Then  you might see a misalignment between pace, power or RPE and heart rate zones that would point to an erroneous basis for HRR, which might be HR Max or rest HR. This misalignment shows through the DSW, the stamina graphs and/or the race predictions.

With regards to alignment between training zones and training effect, Firstbeat tells us that the TE can be over- (resp. under-) estimated when:

- the athlete is over training with too much high intensity training, respectively with too much lower&longer intensity training

- the HR Max is too low, respectively too high

In your case, I would start with setting your HR Max as well as you, do a new LT test (if your use LTHR for zones) and run across a range of efforts (ie balanced focus).

Microsoft Word - EPOC_based_training_effect_assesment_REVISED 2.doc (firstbeat.com)

Thanks very much for your extensive answer, very insightfull. My guess is that my max HR declined due to not running very intensive runs for a long time. I avoid sprint interval training runs and never realy do maximum 5 or 10K  event runs. That way I limit risks of getting injured in my marathon training but I guess the max HR over time would decline because the watch over time only sees lower max HR values in the runs.

I guess changing the max heartrate and cancelling the auto detection was a good first step. I'll also run a new LT detection run in the coming week and see what happens. I really miss the predictability I had the last years of running in Zone 2 as a base run so you also immediately feel the progress when the pace gets higher when running on the threshold of zone 2 to 3. But seeing tempo run indicated even when running lower in zone 2 is really annoying and actually skips the use for this expensive watch for a major part. I would love to buy the epix pro but I will not because the indications are not logical and predictable at the moment (well already for more than half a year). Let's see what happens and I'll update here if it works.

R R said:

marathon training but I guess the max HR over time would decline because the watch over time only sees lower max HR values in the runs

What could happen, if over training, during a marathon focused training plan, would in fact be some under-estimation of your training effect. IE, a tempo run would be seen as base. That said, to feed the watch with a variety of efforts (for all metrics), you don't need to run 5k PBs or crazy 400m repeats. You just need to keep your training focus balanced. Unfortunately, to get a fair training focus you need good HR Max estimates.

When I first go the Epix 2, I spent about 6 months modifying the HR Max value around the 220-age result and I was using the adequacy of the pace targets vis a vis the HR Max zones to gauge whether I was close enough, in particular the lactate threshold DSW pace target. This actually worked out very well. I ended up with an HR Max 10bpm lower than the formula suggested. Then I switched to using anaerobic workouts peaks, then to 5k races.