Hi,

I was just wondering who else pays attention to Strava Fitness and Freshness and what they make of the scores?

How useful is it?

For me I probably spend too long looking at it, but it doesn’t seem to show anything useful other than that if I have done a hard ride my fitness goes up as does my fatigue. My scores hover in the low 50’s for both at the moment.

Cheers

Andy

I use it but like you am unsure, mrblobby and others on here seem to think its not worth much.

I think it drops fitness too quickly. I really look at it to confirm what I’m doing isnt too much/too little and to assess tapering. I’ve never deliberately ridden/not ridden because of the graph.
Currently I’m fitness 62, fatigue 75 and form -14

I use it, but mostly as a comparison against how i’m feeling to guard against over-training.
My premium runs out in 2 weeks and not planning to renew – don’t think i’ll miss the chart.

It’s useless for anyone who doesn’t just cycle ’cause it doesn’t include running etc. in the calculations.

There’s a great article on the Strava blog about what the numbers mean. Most people perform better when “slightly fatigued” apparently. I use it as a kick in the butt to try and get the fitness number up and I notice the more I do that the more the fatigue seems to die off quicker. I’m not using a power meter though just HRM.

Just taken a look, as I skipped using a HRM since I long one last August until this June it’s not much use… I’ll keep checking as my riding is stepping up at the moment.

Its useful as an approximation to my fitness year on year and I can see the peaks and troughs of my training cycles, if you use HR & power.

However “distance” appears to be a big factor in its calculations, so hammering away on a turbo appears to have minimal impact on the graphic and therefore doesn’t represent fitness that well.

However “distance” appears to be a big factor in its calculations, so hammering away on a turbo appears to have minimal impact on the graphic and therefore doesn’t represent fitness that well.

Distance doesn’t have any impact on the calculation at all. Its purely a function of intensity x time. The effect you are likely seeing is that it undervalues the effect of short hard efforts (VO2 Max for example).

Bit silly this stuff isn’t also in the mobile apps – not everyone uses the website all the time.

Just found it on the website under the training menu across the top. No idea what the numbers mean (yet) ” In general the overall numbers aren’t as important as general trends. If” doesn’t help..

Going on HRM alone. (which i’ve only been using 2-3 months tops)

Fitness 8, Fatigue 12, Form -4.

Well my legs are still a bit stiff from the weekends exertions and i’m still tired.. The general trend on the fitness graph is ‘up’ which is good

Kryton57: If it’s basing its stress scores on heart rate then it will under estimate the intensity of interval sessions due to heart rate lagging behind effort so much.

I think the numbers are only useful for trending an individual – not sure you can compare from one person to another. Mine is showing fitness 18, fatigue 25 and form -7. No idea how relevent those numbers really are though – it’s hard to see how form could be dropping given I’m setting lots of Strava PB’s on each of my rides at the moment.

I didn’t even know it existed
<logs on to Strava>
Ah it is a premium only feature, I’m not paying for it.

I use it but like you am unsure, mrblobby and others on here seem to think its not worth much.

I don’t really know about the Strava one but I use the PMC in Training Peaks a lot, which is where it comes from. In Training Peaks you have CTL (Chronic Training Load => Fitness) and ATL (Acute Training Load => Fatigue). The difference between them being TSB (Form). You get a graph of it over time (the Performance Management Chart) and projected into the future based on planned workouts.

I base a lot of my winter training around it. I want to try and steadily build up my Fitness week on week while not letting the Fatigue get out of hand. I try and avoid any big jumps so as not to over train. For example, this was last winter…

The pink line jumping all over the place is Fatigue, and the steadily increasing blue one is Fitness. The orangey one is Form. The dots are workouts based on TSS.

The gap between the two values is more useful when trying to come into racing form. A big negative or positive value generally isn’t great. With a bit of experience you’ll know what values you tend to race best at and you can use them to plan a peak.

Thing is with Strava at the moment, it’s quite handy that it’s got the values to show you were you are, but you can’t use it to plan sessions and to work out where you’re going to be. Though the equations are well known and it’s easy enough to knock up a planning spreadsheet.

anything useful other than that if I have done a hard ride my fitness goes up as does my fatigue

Fitness roughly works over the past month and a half, and Fatigue tends to work over the past few days, so one tracks a lot slower than the other. If you put in a hard few days of training your fatigue will jump up and you fitness will slowly rise. Same with a few days of rest, fatigue will go down rapidly, fitness less so. As you’d expect really.

There’s more here…

http://home.trainingpeaks.com/blog/article/what-is-the-performance-management-chart

There’s loads of stuff on it over at TrainingPeaks if you do a bit of digging.

Kryton57: If it’s basing its stress scores on heart rate then it will under estimate the intensity of interval sessions due to heart rate lagging behind effort so much.

Kryton is probably using power, so thats most likely not it for him. However yes, this is one of several reasons the HR metric on the chart is flawed. Other examples include HR being suppressed when heavily fatigued, and HR being grossly overinflated on rough descents on the MTB.

No idea how relevent those numbers really are though – it’s hard to see how form could be dropping given I’m setting lots of Strava PB’s on each of my rides at the moment.

If you are gaining fitness (trending over several months) you can go faster even with poor form. However -7 isnt poor form, its very close to peak condition. -50 would represent poor form, and is not at all uncommon after a particularly hard race or training block.

Edit: @mrblobby – I believe the strava one is almost a direct copy. But you cant do manual entries or see the effect of planned workouts.

However “distance” appears to be a big factor in its calculations, so hammering away on a turbo appears to have minimal impact on the graphic and therefore doesn’t represent fitness that well.

Curious that. It should be based on TSS (or the Strava Stress Score.) How does that value look for turbo sessions? Is it unusually low? Are you using a PM or HR for those sessions?

Also if you’re using Trainer Road, how does the TSS reported compare to the Strava Stress Score?

It’s useless for anyone who doesn’t just cycle ’cause it doesn’t include running etc. in the calculations.

Training Peaks is good for that as you can enter in TSS values for different types of workout. It’ll do a HR based TSS for runs.

I think it drops fitness too quickly. I really look at it to confirm what I’m doing isnt too much/too little and to assess tapering. I’ve never deliberately ridden/not ridden because of the graph.

I think that’s really where it’s limited in Strava. In training peaks (or even if you knocked up your own spreadsheet) you could enter in your plan for the next couple of weeks and see how the values change and tweak your plan based on those profiles, instead of just looking at the day by day values.

I went premium about this time last year so started using an HR for all my rides (dont have a power meter). I commute a lumpy 10m each way at least 4 times a week and only occasionally get out on weekends, but if I do its usually a long one.

It took 5-6 months before my f&f scores plateaued and it did seem to get it roughly right, but not enough to really do anything with, certainly no more than learning how you feel riding in each day. For me it missed due to not considering weather (wind) because it didn’t know my power, the fact I was swapping between bikes a lot (fixie, mtb, commuter with big panniers), and how fatigued I was from working.

My score peaked at 49 after a good commute week and a big weekend ride, but interestingly it said my form was -15 before that big ride and I felt great on it. Generally fitness floated between 40-45 though once settled.

Ive given up wearing the HR for commutes now though, its an extra thing to remember in the morning and too hot and sweaty in this current hear for what is little benefit. So my ‘scores’ are effectively rubbish and not worth looking at.

Having said that I wore it for a tour on the bank holiday, 400 miles easy terrain, few hills if any, gentle pace with average HR 100-120 and never above 150, long days though to make up the miles. That jumped my fitness from 23 to 41 and form from 0 to -71 LOL! I believed that one the next week, ouch.

and how fatigued I was from working

🙂 I often think it should allow you to enter a TSS value for how hard a days work you’ve had or how little rest you’ve had.

On that subject of rest and stress I’ve been tracking HRV (Heart Rate Variability) for the past few months, another interesting metric to add to the mix and compare to the Freshness/Fitness values, and can be a good indicator that I’m overtraining / not getting enough quality sleep / coming down with something.

I am using power. Re distance / time fair point I did actually mean time on he bike. E.g I can do four hard Turbo seasons a weeks and see a point jump of 3-4 in strava per session, but then a few days later do a 100k z2 and see that ride jump the score by 10pts or more. Vis a Vis it would appear that I got loads fitter doing a months worth of weekly 100k z2’s than it would if I spent a month on Turbo intervals, which isn’t necessarily the case.

Re distance / time fair point I did actually mean time on he bike.

Well it is basically time x intensity (formula being TSS = (sec x NP x IF)/(FTP x 3600) x 100). That’s partly why if you’re time limited it’s actually not possible to build fitness beyond a certain point as you need to be slowly increasing your TSS week by week and you can only bump the intensity up so much. You can’t really do much more than an hour at an intensity of 1.0 (as that’s an hour at your 60MP/FTP) and you certainly can’t do that day after day.

I am using power. Re distance / time fair point I did actually mean time on he bike. E.g I can do four hard Turbo seasons a weeks and see a point jump of 3-4 in strava per session, but then a few days later do a 100k z2 and see that ride jump the score by 10pts or more. Vis a Vis it would appear that I got loads fitter doing a months worth of weekly 100k z2’s than it would if I spent a month on Turbo intervals, which isn’t necessarily the case.

Not my experience at all.
Just did a bit of digging through training calendar for recent rides.
3hrs30 Z2 produced 84 Sufferscore, 1hr25 inc 2×20 @LT produced 100 (the 2×20 will account for 66.67.
Only really short hard efforts seem to underscore imo

Edit: As blobby pointed out – IF of the whole ride is used in the the calculation, and as shorter harder intervals require longer easier rest periods you end up with sessions that are both short and don’t have a particularly high IF.

3hrs30 Z2 produced 84 Sufferscore, 1hr25 inc 2×20 @LT produced 100

The TSS for that Z2 ride seems really low. I do about 30 TSS an hour Z1 and usually somewhere around 45 to 55 TSS an hour Z2. Based on power.

Though maybe suffer score is worked out differently to TSS. Or maybe your method of determining zones are different to mine.

The TSS for that Z2 ride seems really low. I do about 30 TSS an hour Z1 and usually somewhere around 50 TSS an hour Z2. Based on power.

Though maybe suffer score is worked out differently to TSS.

Agreed, there are clearly some minor differences in the calculation. Without power data they have to make some assumptions about intensity.
Still, in terms of how a ride ‘feels’ i don’t feel its too far off. Certainly felt more tired/sore after the 2×20’s than the Z2 ride.

My heart rate at a given intensity varies absolutely loads. Because of the nonlinear relationship between heart rate and suffer score, I’ve had massive suffer scores that distort the amount of work I’ve done just because the zones were simply wrong for that day. I imagine it could easily go the other way on some days. Using a powermeter the whole time is obviously the answer…

indeed.
just as soon as someone makes a power meter that either easily swaps between 4-5 bikes or is cheap enough to buy 4-5 meters.

It’s probably not too far off. A few cleat based ones coming out. The 4iiii crank based is 350 quid for 105. SRAM making it easy to swap the spider on new bikes for a Quarq for about an extra 400 quid.

Certainly worth having on the turbo bike and whichever bike you ride outdoors the most.

Though for steady rides, if I look at the value HR based TSS would generate in Training Peaks, it’s not actually too far off the Power based TSS.

The gap between the two values is more useful when trying to come into racing form. A big negative or positive value generally isn’t great. With a bit of experience you’ll know what values you tend to race best at and you can use them to plan a peak

Interesting, I didnt know that, I assumed that the more positive form the better – not that my form ever gets more than about plus 10, and only then when injured!

It is frustrating it doesnt combine running too. I don’t normally run but now its cx time a bit of running is necessary.

I’ve only really used it to track where I am relative to previous years. My worst form was -59 after the Salzkammergut Trophy A course. It took me a while to recover from that.

Interesting, I didnt know that, I assumed that the more positive form the better – not that my form ever gets more than about plus 10, and only then when injured!

It really depends on the context. If your strategy is to do a big block of work say for a week a week before a big race, then really taper down for a week, you’d likely end up with a big +ve value. And that might be fine if you’re doing a big target endurance event. But if it’s for, say, a weekly CX series race (a short intense thing), then it may not be ideal. If you’re racing weekly then you really want to maintain your fitness over the season so it’s a bit of a balance.

You can always look back over races where you feel like you’ve gone well and see what the values looked like building up to that then try and reproduce a similar profile.

I’ve only really used it to track where I am relative to previous years. My worst form was -59 after the Salzkammergut Trophy A course. It took me a while to recover from that.

So that’d be a big dose of TSS causing a huge fatigue number compared to the fitness number. After an event like that you’d want to take it easy for a while to bring down the Fatigue number, though you’d see a gradual rise in the Fitness number as it starts to reflect the high TSS of that effort.

I monitor mine but think it over-predicts rate of decline for fitness. It’s a le impulse indirect response model where cycling activities provide an impulse and there is a characteristic time (think half life) for fitness improvement and fatigue. These two half lives are different, hence windows for performance.

The rates are fixed for everyone, and this is nonsense. Fitter riders have slower declines in fitness and faster recover (it’s really a non linear model). Even so, with a year of powercal data, it’s ok. I struggle to keep mine over 60, and don’t worry about a week off training – called a taper.

The rates are fixed for everyone, and this is nonsense. Fitter riders have slower declines in fitness and faster recover (it’s really a non linear model).

You can adjust the constants in TrainingPeaks somewhat to take this into account if the defaults don’t work that well for you.

It is a pretty crude model I guess, but it does seem to work quite well for me. It helps me a lot in planning and in maintaining consistency of training out of season, and also taking a bit of the guesswork out of it and avoiding over training.

At quite a basic level too, if I have a fitness of 60 then I know a moderate load is roughly 420 TSS a week (60*7). So if I’d know, for example, for a hard week I’d need to add a chunk of TSS to that, and for an easy week I’d want to take a chunk off.

Though as someone somewhere once said, if being awesome was all about maintaining a high CTL (Fitness) then audax riders would be wiping the floor with everyone! It’s mostly about what you do with your chunk of TSS.

@mrblobby, if im looking to go timetrialing with access to a turbo with HR and speed/cadence and a wattbike that will provide data for my garmin, is Training Peaks the best programme to follow?

breninbeener, Training Peaks isn’t a programme to follow, it’s just a tool to analyse performance and plan with. Have a look at the website. Though there are a hundreds of training plans available to load into it, and they are quite nicely integrated into the tool, they’re usually around 100 dollars each. I’ve not actually looked at any of those though. If you did use it you could load all your turbo and wattbike work into it and use it to analyse the data and track track progress.