Garmin Fenix 5 : What is Firstbeat Training Effect 2.0 and Training Status – impact on XERT or GoMore?

firstbeat_logo_2013_wide[1]Garmin’s Fenix 5 advertises the use of Training Effect 2.0 and will also be including a new Training Status. This was developed by FIRSTBEAT and announced as long ago as December 2015. As we know FIRSTBEAT are a leading HR-analytics company who partner with Suunto, Huawei, Jabra, Samsung, Sony and others. Indeed many of you would have seen the following screenshot from the new Fenix 5 and thought, “Hmmm, that’s different”. Well at least I did 🙂

Garmin Fenix 5

Training Effect can be simply seen as the impact of the exercise you have just completed on your body and then the increase in the capacity/strength that will hopefully come with adaptation/recovery.

However there will be detail beneath the headline TRAINING EFFECT (TE) and one way that detail could be looked at is distinguishing between Anaerobic TE and Aerobic TE. My understanding/memory of talking to Firstbeat about this nearly a year ago was that this was initially designed with team sports in mind eg where there was a high degree of stop-start activity in a sport such as football. Initially, at least, the rigours of a 2 hour triathlon at some form of constant-ish level of exertion were not the focus.


Clearly buyers of a Fenix 5 will include team sports players. Whilst those players OUGHT not to wear a chunky watch whilst competing against others they could quite easily wear a caching HR strap such as Garmin’s HRM-RUN or Polar’s recently announced H10 and review the data later.

But what about endurance athletes? We go anaerobic sometimes, well often, don’t we?

A: Yes.

xert baronbiosys baron biosystems cadence optimiserI’m sure a few of you have been looking at XERT’s tools for cycling with power. If not, maybe you should. Simplistically if you have a 1 hour race and you average above your FTP for the first 30 minutes; then it necessarily follows that for the last 30 minutes you must average below your FTP. But precisely by how much is a tricky calculation. Sure it depends on the intensity of power-durations above FTP but it will also be impacted by how much recovery you have got in the first 30 minutes at times when you were below FTP; and of course in the last 30 minutes your recovery will increase but at varying rates. Hopefully this illustrates that it is a complex balance between your capacity and changing levels of expenditure and recovery – I don’t profess to know the detail of that but I hope you would agree that it is important in determining the level at which you can perform.

It requires a slight change in how you think. We are not quite talking about time-in-power-zones.

With Firstbeat, PERHAPS we are looking at similar abilities to XERT but with heart rate as a proxy AND with them integrated into the Fenix 5 firmware.


This then draws me back to a post I recently updated. GOMORE have a HR chest strap sensor that essentially performs a similar function but unlike XERT, it is based on heart rate and proprietary algorithms. Their STAMINA SENSOR (OK you can’t sense stamina but bear with me) can essentially tell you how you need to perform for the remainder of your exercise/race – similar areas potentially to what the Fenix 5 might be covering?

GoMore Stamina Sensor

Training Effect 2.0 is different because it tracks ANAEROBIC component training load. Training Effect 1 was geared towards looking at the OVERALL impact on VO2max (cardiorespiratory fitness) of an assessment of EPOC (Excess Post-exercise Oxygen Consumption).

You’ll have seen EPOC on the standard SUUNTO MOVESCOUNT charts for quite some time perhaps?. I think the following chart is a swimming EPOC chart of mine when I was doing some testing last year.

Suunto Movescount Swim EPOC

Anaerobic load is created when performing at high intensities, specifically looking at the level of intensity as well as its duration; but also considering the recovery state at the start of the interval. My understanding is that the result of Firstbeat’s algorithm is then compared to a sport profile to determine the likely adaptation.

I suggested that this sort of functionality needed to be built into high performance sports watches back in 2014 and in relation to an imagined 930XT in 2015 (here). Hopefully 2017 will see the Fenix 5 taking us a long way down the route towards these capabilities and the 930XT WILL follow suit.

IN GADGET-THEORY: It should therefore be possible, one day!, to say that you want to run for 60 minutes and have TE=4.2/5.0 with a 45% anaerobic component to the load. You then go off and do a few ‘random’ 1km reps. The gadget should be able to guide you AFTER those intervals towards achieving those quite complex BUT TRAINING SPECIFIC goals. One of the problems here is that the algorithms will rely on some form of indexing benchmark such as LTHR/VO2max…and I’m not always convinced those metrics are entirely correct. Nor do I think that the ‘readiness’ metrics, as you start an exercise, properly take into account the exercise that you did yesterday; that the device did not record; or that you had a bad night’s sleep.

So HRV may come into play here – HRV is used, in a simple case, to determine breathing rate but it also has much more complex use cases based on the state of the body’s nervous systems; here I mean the more complex uses in the context of estimating incoming fatigue levels when a session or interval is started.

Just thinking about it is giving me a headache. I’ll get back to the day job now.

Training Status – This is potentially MUCH more interesting and useful but I don’t yet know quite so much about it and it’s implementation into Garmin’s high-end 2017 devices (Fenix 5 and …??). If, like me, you’ve used some form of Training Load calculation (ie use TSB/CTL/ATL based on TRIMPs) then the new Training Status will be the same kind of thing but, instead, based on moving averages of EPOC-related data. This sort of stuff can be VERY useful in looking at your fitness TODAY and giving feedback on if your session was productive or unproductive and/or if you are maintaining/improving/peaking.


This post WILL BE UPDATED. I’m doing some related research in my spare time.

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.

14 thoughts on “Garmin Fenix 5 : What is Firstbeat Training Effect 2.0 and Training Status – impact on XERT or GoMore?

  1. I read some Firstbeat’s patents and the way you described the Training Status is quite good. An algorythm computes a readiness index based on your previous session (TE, Rest Time) and on few minutes of your current session to compare the 2 values and give you some feedback.. The patent doesn’t detail the metrics that is used for the index (TE, VO2max..).
    The only issue that I see today at this is the inaccuracy of those models (VO2, TE). They’re designee to fit for everyone and nobody.
    I wrote an article about it (in French) last month and I saw Firstbeat submit a new patent about a method to determine anaerobic threshold in December

    1. yes. there are also some white papers that firstbeat produce.
      and I did have a conversation with FB a while back about the first tranche of their metrics.
      as with all of these things you have to use them as an input into a decision that you make.

  2. Very interesting thanks but one thing I haven’t seen mentioned is that all these advanced analytics will likely require a chest strap as the optical HR sensors are not (yet?) accurate enough to produce r-r (HRV) data.

    Now the FR235 does give a TE, recovery and VO2Max so I suppose they’re just going by HR…the FR735XT gets the Lactate Threshold test but only with an external strap.

    1. supposedly I have an optical hr wrist sensor that is on retail sale already and HRV…we’ll see. (arriving this week apparently). I guess it all depends on the algorithms used for non-HRV estimation…I can easily work out my LTHR without HRV.

      1. yes. i’m wearing it. It’s only hrv at night 🙁 although that hardware could enable hrv in the day too. manufacturer needs to enable that and create functionality for that. it is NOT a HRM that broadcasts hr over standard protocols

  3. Been playing around with the F5 for a week and here’s what I found :

    – Apart from LTHR and the Stress Test (the one that always returns a 1 for me) that require a chest strap everything else is “compatible” with the oHR…which is convenient but also means they’re not HRV based so aren’t a lot more than variations on the MaxHR/Pace/HR ratios. At the very lease “Performance Condition” should be HRV based.

    – Training Effect 2.0 returns odd aerobic results. TE1 had always seemed “consistent” with the perceived since I started using it on my FR610 and the FR235 returned results that were close. Aerobic in TE2 is always higher than TE1, just a bit for very easy runs but then ramps up. Yesterday it told me I overreached (5.0) during a 16k MPace session at only 81% of MaxHR. Sure it’s a hard session but overreaching ? TE1 returned a 4.5 value that makes sense.

    – Training Status shows Fitness and Load trends (arrows), ok…

    – Training Load returns a number (752 in my case) on a blue/green/red scale and says “optimal” but there is no explanation for the value. I’m currently preparing for a marathon.

    So overall not very impressed and unless this improves I can’t see these new metrics as a very good reason to get the F5.

    1. from Fristbeat:
      Performance Condition is essentially a free-wheeling VO2max calculation. In other words, it’s looking at the relationship between heart rate and running speed – or watt output from a cycling power meter. So, the inputs of Performance Conditions are the same as VO2max, but the VO2max calculations introduces parameters that allow it to select those portions of the data that reflect a classic incremental exertion VO2max test. Pretty cool. For example, if you kept running on the treadmill after your Vo2max test was complete, eventually your ability to keep the current pace would decline. That’s effectively what Performance Condition is looking at.

      In Garmin devices, VO2max is a primary factor in scaling Training Effect scores to individual users. The higher your VO2max, the more fit you are, and the harder you have to work to achieve the same physiological impact (in terms of impact on future fitness levels). If something seems wonky with TE feedback on a new device, it’s likely that your VO2max hasn’t quite settled in properly yet. I’ve got a new Fenix 3HR for personal use, and it took a good couple weeks before it tuned up properly for me.

      That said, we are expecting a lot of initial confusion regarding how people make sense of the Anaerobic TE material. A big part of this is because people are familiar with the idea of Anaerobic as the label of one of the heart rate zones… but in reality training in that zone doesn’t necessarily mean that you are training a way that improves your anaerobic performance capability – which is what Anaerobic TE is looking at.

      1. Thanks for the insight. I’m not too concerned about the Anaerobic TE actually as I’ve only seen it kick in during alternating Anaerobic/Aerobic Threshold session on the track, which makes sense.

        The F5 homed in on the FR235’s VO2Max actually after maybe three runs so it seems that the rest of the metrics shouldn’t be thrown off like the aerobic TE is right now. I’ll see how it goes over the next few days.

        One thing I noticed today is that when the oHR is “out of it” when you start running the “Performance Condition” index gets thrown off. Maybe they should wait for it to stabilize before giving a value. And when the 630 was launched they clearly stated they were using HRV for PC…

      2. ‘m endeavouring to write something (with help from FB) to clarify what the various functions rely on. perhaps, for example, certain functions work differently on different devices where different source data (accuracy of) are available

Comments are closed.