Lactate threshold in Suunto Run
-
Hi,
In the Suunto Run the CTL widget shows me a very different lactate threshold (152bpm) than ZoneSense does (166 bpm).
Which one should I use to set my HR training zones?
-
@Pep-Carbonell only accurate method is a professional lactate measurement on a treadmill.
If you can not do that or you don’t want to spend the money, a field test like a 10km Time-Trial/Race or 1h all-out is the most accurate you can go.
The average HR for the last 20 min is in the ball park of ±3 BPM of your lactate threshold. If you do that 2 or 3 times over the next months to reduce daily influences you will be able to quite accurately narrow down your HR at the lactate threshold.Formulas like 90% of HRmax are way to inaccurate and can be ±20 BPM off. ZoneSense is already more accurate but can still have around ±10 BPM inaccuracy as scientific works could show.
A deviation of 10 BPMs is quite good compared to the other calculation methods but it’s still huge if you want to train correctly in the zones. -
@brave_dave thank you so much, I see…
My last 20 minutes of a 10k race and my Zonesense anaerobic thershold are quite similar (+/-2bpm).
But then, there’s the widget on Suunto Run (the watch) that tells me that my lactate thershold is 12 or 14 bpm lower than the other methods and this confuses me, I guess it’s wrong.
Anyway, thanks again.
-
@Pep-Carbonell yes, the values measured by SZ can be correct and can be closer or further away from time to time but it is hard to know when it is accurate and when not. Scientific papers could show that DDFA is correlated with the lactate threshold but even under perfect conditions/set-up on a treadmill, it is hard to detect the correct value. A publication from May 2025 (it’s actually from the scientific partners from Suunto from Tampere University) showed that the DDFA method indeed performs better than the estimations from HFmax. DDFA showed on average a deviation of ±10 BPMs from the actual value in 58 subjects…±10 though is usually bigger than the difference in HR between 10k and Marathon Pace. 10k is normally well above threshold and Marathon well below. At +10 BPM above my threshold I’m almost dead why -10 feels really easy :D. So it works in principle but is it accurate enough? A good race-based estimation will probably be on average a lot closer than ±10 BPM.
(Here is the publication of you are interested https://physoc.onlinelibrary.wiley.com/doi/10.14814/phy2.70241)
Of note: the accuracy of detecting the aerobic threshold was with ±12 BPMs even worse. -
@brave_dave great answer! thank you so much, especially this final note about the often forgotten aerobic thershold.
In the end, I guess RPE is more important than it seems.
Regards!
-
@Pep-Carbonell you are welcome!
Yes, if you are really experienced RPE can be way more accurate than wrong results from formulas etc. as you also mentioned in your initial post. But unfortunately RPE is also no alternative as it is so subjective that not even professional can accurately sense their training zones.
Pace zones based on race efforts are still best for training control when lactate testing is not possible and that’s what a lot of pros also still use. -
@Pep-Carbonell said in Lactate threshold in Suunto Run:
In the Suunto Run the CTL widget shows me a very different lactate threshold (152bpm) than ZoneSense does (166 bpm).
In my experience, the value in the widget is at least very stable. It doesn’t constantly fluctuate up and down like the value from the Zonesense calculation. And even though I don’t have a lab measurement, I think it’s at least close to the truth.
Maybe I’ll do a test sometime. I have a lab just a few blocks away, but I can get some really good running shoes for the money
. For now, I’ll stick with it like I do with VO2 max. I don’t know if it’s accurate, but the trends seem to be right. -
@Yves_ Yes, the widget is very stable, but according to my feelings, the widget seems close to LT1 and Zonesense seems close to LT2.
In the end, the more I run the more I let myself be guided by my feelings glancing sideways at the data
-
@brave_dave Hi, I read the article some weeks ago and AFAIR I am not sure I agree with your conclusion (which, by the way, aligns with the authors’): “the DDFA method indeed performs better than the estimations from HFmax”.
The article states:
“With the thresholds derived from the measured HRmes/max, on the other hand, the correlations are the largest (0.55, 0.78) from the compared methods, but the mean differences μdiff (10.65, 7.24) are only slightly better compared to those of the HRtheo/max Ts.”
For the HRmes/max method, the Ts are estimated based on a ratio of HRmes/max and, according to the article, this ratio is “arbitrary” chosen. If you decide to use the available data to define this ratio, the estimation from HRmes/max would outperform the ZS method for the population tested since the standard deviation is better for the HRmes/max method as stated in the article excerpt above. Defining those LT / HRmes/max ratio valid for the entire population would need probably more that 58 values but I don’t see any reason to choose arbitrary values for the ratio as long as data is available to estimate this ratio.
If I am not wrong, this is a kind of a definitive death for the actual version of ZS but…
I think that ZS could still be useful for some people. Side effects of heart medication can lower the maximum heart rate and the lowering depends on the time elapsed since the medication has been taken. The lowering can be higher than 20bpm so higher than the ZS method uncertainty estimated in the article. And it seems that ZS is highly correlated to this lowering. But right now, this is only an anecdotal experience without any scientific validation. To some extent, even without medication, I think that the maximum heart rate can also vary but probably not as much as with medication. -
@LGoSo hey, thanks for reading the paper and your nice input and willingness to share your information. But I cannot follow you to 100%. Maybe I do not completely understand your thoughts…
The point I don’t get is with the ratio.
What the authors do is, they define the first threshold
(T1 = 70% of HRmax) and the second (T2 = 85% of HRmax).
Then they have two HRmax values they can base their estimation on.
HRtheomax = calculated by the age formular 220-Age
HRmesmax = measured by the all-out test on the treadmillThen they compare both estimated values T1 & T2 for both methods to the actual values calculated by the lactate testing. That’s why they get for each method two correlation and deviation values as well as for the DDFAT method. (E.g. HRmesmax T1 = 0.55 and T2 = 0.78)
The overall correlation of HRmesmax is indeed better than the DDFAT method for both threshold but the deviation from the actual value is higher. So therefore even that DDFAT has a lower correlation, the accuracy is higher because of the lower deviation.You are definitely right in your point, that if there is a systemic over- or underestimation of the HRmesmax method and compensation for that error would lower the absolute deviation from the actual values, the method might become more accurate than DDFAT because of its higher correlation. But I would think that this has already been tried.
-
@LGoSo what I just realized is that there might be indeed a systemic error for the HRmeasmax T2 . When you look at Fig. 4 almost all values are below the 0 value. The individual points are more compact/have less deviation from the measured average. Therefore, always compensating for that error would shift all values slightly up. Due to the lower data-spread this might even lead to a higher accuracy compared to DDFAT, at least for the second threshold T2…
FYI: the publication/ZS is in parallel also discussed in this thread, if you haven’t already seen it
-
B brave_dave referenced this topic
-
@brave_dave Yes, that’s exactly what I tried to explain.
