De-Saturation Intervals: Moxy vs. NNOXX vs. Train.Red

A more standard view of intervals and training: HR and power

De-Saturation Intervals: Moxy vs. NNOXX vs. Train.Red

More: NNOXX Review

More: Train.Red Review

In some recent testing with NNOXX, Evan advised me that I needed to train to overcome my performance limiter which was deemed to be a “limitation to improving VO2max“. As well as increased endurance mileage, NNOXX recommended de-saturation intervals. Essentially intervals that deplete the muscle of oxygen in a protocol specific to the demands of your sport. For me, that’s achieved by a series of progressive efforts (ramps) that culminate somewhere over my threshold before repeating after a complete rest – for a soccer player it might be 30-second intervals.

NNOXX One analysed my half marathon effort – here is exactly what they found

The image at the top of the post shows what I did indoors on a Wahoo KICKR. Even though I was ‘prescribed’ these kinds of workouts for running, I tried them a few times and either had data capture problems or difficulties in controlling the increases in my efforts – in a nutshell, a controlled interval is WAY easier to perform in a controlled indoor bike environment than when running in my local park in the rain with a dog! I guess we all know that!!

This workout was physically hard (fairly hard) and I’d previously tried to do other tech comparisons with maximal tests and it’s just not possible for me to keep churning out those kinds of efforts day after day. As it happens I did a desaturation interval the previous day for running but was recovered enough to do this one on the bike.


NNOXX vs Moxy vs Train.Red

A comparison of SmO2 data between tools is incredibly difficult to do but on their own, each one is straightforward to use and interpret. It’s really time-consuming and there are quite a few bits of tech involved in the data capture. Getting every part of the ‘process’ to work has proved extremely challenging on every occasion. Perhaps I’ve encountered the most challenging series of issues I’ve ever had to overcome in all my time doing this site (10 years?). One small mistake (which happened a few times), one bug in some beta software (which happened), one of many batteries too low on charge (happened), or even the sensor moving (happened) or interfering with another one (happened) and the whole test-kaboodle is toast. I’ve devoted a few weeks of my time with little to show in terms of comparisons for you guys and girls.

Anyway, enough of my problems, here’s the kit I used

Then there are more problems with interpreting the data.

These SmO2 sensors don’t claim to be quite measuring the same thing. Train.Red measures relative changes and NNOXX claims to measure deeper, and Moxy has lower-resolution data. I can’t even use DCR’s tool to show you these as I can’t (easily?) get the data into a developer field in a FIT file for DCRAnalyzer to even read the data.

To compound all of this…very few of you are interested in Muscle Oxygen and won’t have got this far 🙁 Hey oxygen and fuel/glucose are quite important tho, right?

Here Are Some of the Views of today’s app data


The unique thing with NNOXX is that it shows NO data No other wearable does that.

The online platform has drill-down capabilities


You can see here that I have warmed up pretty well (1192). The subsequent efforts start at 1792 and the SmO2 performs pretty similarly for each interval falling to the same level at a similar rate.

the NO is a little harder to interpret. You see three aspects to NO in the intervals. Firstly it increases in line with progressive increases in power. Then it quickly drops once the work stops only then to bounce back to the highest levels in recovery before declining sharply as the next work period starts. an interesting takeout is the consistency of the peaks/troughs.

More technically: An increase in power results in an increased oxygen extraction fraction and decreased muscle oxygenation (SmO2). This means oxygen demand in the working muscles is greater than oxygen supply. Furthermore, NO increases, resulting in greater blood flow and oxygen delivery to tissues. The combination of these measurements means that I utilized an increasing fraction of an increased oxygen supply, which is why my mVO2 also goes up. These trends are more exaggerated in the first few sets.

Note: Recorded power levels were effectively identical for the 6 sets. however on the last two sets I used the big ring on the bike and this felt easier either because of a measuring error (higher resistance applied to a greater virtual speed) or because my cadence lowered slightly to a level I more usually train at. A more biomechanically efficient pedalling technique could explain lower NO for a given power output.

I felt that this workout was comfortably hard, or maybe a tad harder than that. I felt I had at least two more sets left in me.

The data backs up my feeling. I always was able to deoxygenate to the same point as the previous set.

Heart rate tells a similar story. It rose similarly in each interval to a progressively higher peak. But it didn’t tell me how many more progressively higher peaks I might be able to achieve as I will still be some way off my HRmax.



The Moxy data on the Wahoo tells a similar tale. Intervals 4 & 5 do show a slightly surpresed level of SmO2 recovery and the final effort does show the lowest point of SmO2.

Golden Cheetah

I’m unable to overlay the curves but they would be different.


Here is the TrainRed chart. Again a similar pattern although this data suggests that my recovery SmO2 levels are progressively declining.


Garmin Connect via FIT Developer Field

(I can’t show the app data as I have an NDA and am using the beta)

Take Out

It’s cool data. It’s also hard to compare!

These two new products (TR and NNOXX) use high-grade sensors and Moxy now has some serious competition. However, I’m pretty sure that data-driven runners and cyclists will want this data on a Garmin and in a Garmin FIT file rather than on a proprietary app when it comes to routine usage. Perhaps a smart app and rich ecosystem make sense if we want to dig deeper into more unusual data like this. Moxy follows the existing data standards and its data is easier to use after the workout.

I’ve seen some even cooler things that I’ve not written about yet. An example is Train.Red has such high-resolution data that you can EASILY see individual muscle contractions in detail. I’m not necessarily quite sure why I’d need to see that but I can and I have! But as an X-Fitter/Hyrox/lifter type person, maybe you would.

  • NNOXX – much cheaper, simpler app, measures deeper, also has NO and acceleration data
  • MOXY – lab standard and follows the standards. Lower resolution data
  • TRAIN.RED – high-frequency data, measures deeper, nice app, expensive and trying to move into the Garmin environment and labs.

Buy One


train-red fyer and plus
Use Code: the5krunner for a 10% discount – may not work with other discounts

Reader-Powered Content

This content is not sponsored. It’s mostly me behind the labour of love which is this site and I appreciate everyone who follows, subscribes or Buys Me A Coffee ❤️ Alternatively please buy the reviewed product from my partners. Thank you! FTC: Affiliate Disclosure: Links pay commission. As an Amazon Associate, I earn from qualifying purchases.

7 thoughts on “De-Saturation Intervals: Moxy vs. NNOXX vs. Train.Red

  1. I’m still rocking my BSX. I don’t know what to do with the data, but for kicks and giggles I do a ramp test every once in a while to find the inflection point.

  2. This is the review I wanted to see, the power of all these tool is how usable is the data. It still seems a bit like RunScribe that you get loads of data but trying to interpret to make data itself useful, and often the reference examples are too simple for what reality shows.

    I like data so the learning curve fine, but initially what’s needed is a guide saying this is what your data means and you should try specific xyz to improve.

    Still hope NNOXX gets released in the UK so I can compare with moxy.

    Two good points with SMO2 I find is seeing
    1. the impact of warm up before racing, or how much warm up is required
    2. What steady state looks for a specific duration before your strength/endurance begins to reduce

  3. Nice review – but, these products still have yet to show actual “clinical” benefit while using them long term. The cost/benefit is terrible IMO and data interpretation (including company claims) is very problematic. As mentioned above, the good old BSX is reasonable to use to locate the RCP (if you know what you are doing). Mine is still working as well and I wish they were still in business.

    1. You mean HEX or BSX?
      HEX is still sold by a Spanish company – . Apparently its CIQ app is quite nice. Maybe they’re just clearing out old stock.

      There are quite a few clinical validations of the Train.Red sensor (ie the sensor it uses not the brand’s product itself) I think? NNOXX I’m not sure about.

      when you say clinical do y ou mean in a sports lab sense or the sense of clinical use of NO? I’m probably less interested in either and, possibly like you, more interested in how it can be used by a self-training athlete.

      ty for the kind words. Hopefully I wil lbe able to do a more formal review that’s nicer 🙂 wish me luck !

      any readers: check out Brian’s site:

  4. For me as a professional exercise physiologist it’s very important that the data goes into a spreadsheet for analysis. Moxy connects to mobile gas analyzers (like vo2master) and full scale cpet machines. Does the nnox do the same?

    Also I don’t know how to use NO metrics in endurance engineering yet, besides “if levels are low, take citrulline and beetroot “, but I prescribe this in any case

    1. NO is more complex that than and not linked to dietary sources.
      NO vasodilates, as you probably know. this might be interesting:
      I have some more stuff coming out on NO in the coming weeks. as you can probabyl guess, it complicates SmO2 interpretation even more!!! actually refines it.

      i’ve been emailing Evan @nnox today as it happens. No I dont think there are links to any 3rd party apps. No doubt VO2master will be the first to link to! already links to it

Leave a Reply

Your email address will not be published. Required fields are marked *