Author Topic: Performance Management Chart (PMC) - Fitness/Fatigue/Form tracking (CTL/ATL/TSB)  (Read 1355 times)

By request I'm starting a thread to discuss the example spreadsheet I use for tracking fitness, fatigue and form using TSS (Training Stress Score) values.

Here's my post from the Audax thread about fitness and recovery:-

The really scientific answer would lie in knowing the TSS values from every ride (and other strenuous activity you do) and then keeping track of your ATL, CTL and TSB values.

Knowing the TSS for each activity requires either a power meter and knowledge of your FTP, or an HRM and something that calculates HR-TSS.

The general principle of (cycling) TSS is that an hour at FTP = 100 TSS.

So 2 hours at 50% FTP also = 100TSS.

A typical 200km Audax I'm averaging about 50%-60% FTP, so if I'm moving for 10h of that then that's 500-600 TSS.

(As you get fitter you'll find that you need to go further/faster/etc to get the same TSS load as your fitness improves.)

With that you can work out CTL (Chronic Training Load = Fitness), ATL (Acute Training Load = Fatigue) and TSB (Training Stress Balance = Form).

(stolen from my post on LFGSS)

Quote
Most systems like PMC/F&F work on the same principles:-

    Fitness is based on the last 42 days data.
    Fatigue is based on the last 7 days data.
    Form is yesterday's fitness - yesterday's fatigue

I do all of mine in a spreadsheet now. The tricky bit is accurately calculating the various TSS scores. The CTL/ATL figures are relatively easy:-

    Today's CTL (fitness) is 41/42 of yesterday's fitness, plus 1/42 of today's total TSS score(s).
    Today's ATL (fatigue) is 6/7 of yesterday's fatigue, plus 1/7 of today's total TSS score(s).
    Today's TSB (form) is yesterday's CTL - yesterday's ATL

Here's a sample Google Docs spreadsheet I made to do it: https://docs.google.com/spreadsheets/d/1e4I1V7fdzyzkvbtfjiuDTbqbifBfJyFn6fnu4DzC848/edit#gid=0

The point of training is to ramp up CTL, but not too fast that your ATL also builds up and leaves you knackered (big -ve TSB). Then you get to tapering where you will lose a bit of fitness (CTL) but lose more fatigue (ATL) and so you hit your target race/ride/whatever with a big +ve form (TSB).

Finally, here's my data for 2014-2016 plotted on a chart (PMC = Performance Management Chart):-

http://www.greenbank.org/misc/pmc2014_2016.png



You can see me start to try and get fit from September 2014 as my daughter started school and I had a day off to beast myself. The blue (fitness) line starts to rise up nicely. It tailed off for Christmas, then a bit of an effort to kick it off again after the new year. The big fatigue spikes are Audaxes (Ditchling Devil in June I think). There's also another attempt to get fit in the first few months of 2016 but then it tails off and I lose a considerable chunk of my fitness before the big spike in late July which was the Mersey Roads 24h TT.

I had hoped to get my CTL up to >100, which means at least an hour at FTP every day (or 2 hours at 50% FTP every day). Never quite got there though.

(Mine is more complicated now as I'm mixing in running, swimming and cycling. Not all TSS are created equal!)

Anyway. This is more than the vast majority of Audaxers ever consider, but it's the scientific basis behind fitness, fatigue and form - which are the essential parts of recovery.

I'll update this post with more details as the discussion goes on...
"Yes please" said Squirrel "biscuits are our favourite things."

Feel free to wade in. I'm definitely no expert in this, this is all just what I've gleaned from various webpages, blogs, books, etc.
"Yes please" said Squirrel "biscuits are our favourite things."

Really appreciate this. I've previously used training peaks, today's plan, and lately stravistix. Having quit my coach, and no longer needing plans made for me (it's not rocket science), I'm loathe to pay the fees just for the ability to model the effects of weekly TSS. I have switched to the stravistix Chrome plugin (renamed elevate). Whilst stravistix is free, does cycling TSS with power, and calculates HRSS for running, it runs off Strava, so can only tell you CTL/ATL/TSB for what you've done, whereas I want to know the effects of weeks I'm planning.

I've plugged my numbers into Greenback's sheet, and the numbers are spot on. I've entered my planned weekly total TSS up to Christmas, just to make clear what I need to do to keep progressing.

Previously, I was doing 450/550/650tss cycles over three week loops. However, change of circumstances means if I can do >300tss a week, that's good enough for me.

Sent from my Lenovo P2a42 using Tapatalk

quixoticgeek

  • Mostly Harmless


Cheers for this, really useful stuff.

Power meter, Heart rate monitor, and knowing what your FTP is. Any other requirements to make this work?

How do you deal with FTP changing over time?

J
--
Beer, bikes, and backpacking
http://b.42q.eu/

Power meter, Heart rate monitor, and knowing what your FTP is. Any other requirements to make this work?

Theoretically you could do it just with an HRM and something that calculates HRSS. No need for a power meter.

I get my cycling TSS scores from Garmin Connect and/or Golden Cheetah. I should just write something to calculate it myself from the .fit file as a power spike in today's ride (2 seconds at 35kW) means it threw off the Normalised Power calculation, which throws off the Intensity Factor calc, which throws off the TSS calc. I don't think 1912.3 TSS for a 2.5h relatively gentle bimble is entirely accurate. I was able to get a more sensible figure from Golden Cheetah after fixing the erroneus power values (using fitfiletools.com) but it's a faff.

Running I calculate them based on perceived intensity (1.0 for a parkrun, 0.9 for a relatively hard training run, 0.8 for a steady run, 0.7 for a gentle run) multiplied by time (again, 1h at max I can do for an hour = 100 TSS). Most of the comparisons I do are based on VDOT (https://fellrnr.com/wiki/VDOT) which is great for comparing runs of differing length (as long as they're flat, it doesn't take into account elevation at all!). I've wibbled about VDOT a bit in the running thread.

Swimming TSS involves knowing your Swim CSS (http://www.swimsmooth.com/improve/intermediate/css-training) which is analogous to cycling FTP. I hate those tests so don't do them anywhere near often enough. I also need to do more work on my technique as that's free speed. Same HR effort and 20s/100m quicker, that kind of thing.

Eventually I want to automate all of this with my scripts reading my .fit files and detecting what type of run/cycle/swim I was doing and wibbling things automatically (i.e. detecting a 'race' like parkrun) or when I can be bothered to do a Swim CSS test (which I should do in either of the two 'easier' training weeks in a 4 week block).

How do you deal with FTP changing over time?

Since I rely on Garmin Connect and/or Golden Cheetah for my cycling TSS scores it's down to those systems. Garmin Connect automatically tells me my FTP is improved (based on 20 minute power bests from rides and, probably, from the associated HR) and so, if I accept the change, the subsequent TSS values will be lower as appropriate.

I've plugged my numbers into Greenback's sheet, and the numbers are spot on. I've entered my planned weekly total TSS up to Christmas, just to make clear what I need to do to keep progressing.

I put predicted numbers in manually so I can see what's likely to happen in the future (and avoid TSB going below -30 which leaves me dribbling on the sofa). In my version of the fitness tracking spreadsheet (now up to column CC) I've got a rough outline daily plan for my running all the way to London Marathon in late April 2019. I can then add in the twice weekly 5-a-side games and cycle commuting, and the Mondays that I try and do a long cycle. Then I've got swimming to throw in although that's sidelined for a while as I did Swim Serpentine 2 mile swim 4 weeks ago and have nothing else big planned until this time next year (might go for the "Super 6" six mile event at Swim Serpentine next year, and there's a 10k swim event in November that I'd like to do). Oh, I might try and do a half Ironman next summer but that's "only" 2000m swim.

I'm aiming to do 10+ hours a week, on a 4 week mesocycle[1] of hard, harder, hardest, easy. If I were to put TSS figures on it then they'd be similar to yours at 450, 550, 650, 300. There's a lot of give in my schedule though as it has to fit with family/school/holiday plans. This coming week is half term and I'll be lucky to scrape in with 200TSS I think. Once every 3 or 4 of those cycles (i.e. 12 or 16 weeks) I'll try and give blood too, and that has to correspond to an easy week!

My last 5 week totals:-
TSSTotal Time
346.154:29:00
233.344:03:00
466.446:16:00
326.655:12:00
625.6110:01:07

1. Guff here: https://www.trainingpeaks.com/blog/macrocycles-mesocycles-and-microcycles-understanding-the-3-cycles-of-periodization/
"Yes please" said Squirrel "biscuits are our favourite things."

I was running training peaks, which manages all of this pretty well, but stopped as I am not coaching my running friends at the moment. However, like the spreadsheet.

Meanwhile, I’ve been seeing how effective the Garmin metrics are - estimated FTP, TSS, training load and recovery time, plus live performance condition is quite interesting as well. This often starts positive and tails off during Ana ctovity!

Really, given ongoing knee issues I’m just riding a bit at the moment.

mattc

  • n.b. have grown beard since photo taken
    • Didcot Audaxes
I'm new to this formalization.

What does "form" actually mean?
(Is it something like what TSS score you can achieve that day? Or what your FTP will be that day? )

Or is it purely defined as the equation you posted?!
Has never ridden RAAM
---------
No.11  Because of the great host of those who dislike the least appearance of "swank " when they travel the roads and lanes. - From Kuklos' 39 Articles

I'm new to this formalization.

What does "form" actually mean?
(Is it something like what TSS score you can achieve that day? Or what your FTP will be that day? )

Or is it purely defined as the equation you posted?!

Effectively, how fit you are now offset by (minus!) how tired you are. So being fit and knackered is not a recipe for high performance. The idea is that you recover from tireness (fatigue) more quickly than you lose fitness, so resting after building fitness improves form allowing a peak performance.

The models are useful, but a bit of self learning alongside helps too.

mattc

  • n.b. have grown beard since photo taken
    • Didcot Audaxes
Thanks Mike.

But I get the intention (which I think is what you're describing); I'm asking for more specifics. If they exist!
Has never ridden RAAM
---------
No.11  Because of the great host of those who dislike the least appearance of "swank " when they travel the roads and lanes. - From Kuklos' 39 Articles

Googling: CTL ATL TSB would get you a whole bunch of pages that say it better (and worse) than I can.

TSB (form) isn't a specific metric of anything other than big -ve and you'll really struggle to get off the sofa to do anything. To know what it means in practice you've got to experience it  (you already have after long rides, only you didn't know what your TSB was that day).
"Yes please" said Squirrel "biscuits are our favourite things."

mattc

  • n.b. have grown beard since photo taken
    • Didcot Audaxes
Gotcha, thanks.

I've just realised there is a follow-up question: where do the 42 days and 7 days factors come from? Is there any science behind that?
Has never ridden RAAM
---------
No.11  Because of the great host of those who dislike the least appearance of "swank " when they travel the roads and lanes. - From Kuklos' 39 Articles

There seems an increasing interest, probably aligned to more available monitoring devices, in heart beat interval variance as a measure of fatigue.

There seems an increasing interest, probably aligned to more available monitoring devices, in heart beat interval variance as a measure of fatigue.

I think the Garmin 'Performance Condition' is based on this - apparently more variance is associated with lower fatigue. Also the 'Stress' measurement.

Gotcha, thanks.

I've just realised there is a follow-up question: where do the 42 days and 7 days factors come from? Is there any science behind that?

They are exponential decay periods used to measure the value/fatigue consequent of any particular exercise. I'm not sure how well established they are, but I suspect that they are at best approximations and that there is a range of appropriate periods across individuals and over time.

The idea is that undertaking training improves fitness and creates fatigue. The fitness gain is of less consequence over time and the fatigue fades over a shorter period also. For the model to be effective you need to measure training stress (hence TSS) and have a fair idea of the relevant time periods.

The model is a bit simplistic in that you (well I) don't seem to get an instant fitness benefit from any session as it takes a few days for muscle to be built or red blood cells to be produced, even ignoring the fatigue element. However, I think it's quite good as an approximation that you can learn well enough to use effectively.

Mike

mattc

  • n.b. have grown beard since photo taken
    • Didcot Audaxes
There used to be a saying - common in racing/training circles - that it took about 2 weeks to get the full benefits of any training. I think this massively predates power measurement and blood tests, but might have some truth in it!
Has never ridden RAAM
---------
No.11  Because of the great host of those who dislike the least appearance of "swank " when they travel the roads and lanes. - From Kuklos' 39 Articles

Yes, I think that's about right. When I was younger and running competitively at shorter distances (800m to 5k), if I had a long race to do then I'd do an overdistance run a couple of weeks before to help prepare. Usually seemed to do the trick.

Yep, and the whole point of pushing yourself a few weeks out from an event (if you plug numbers into a spreadsheet like this) is that you're trying to maximise form (TSB = CTL-ATL), not just fitness (CTL).

To maximise fitness you just go out and smash yourself, the problem is you'll be knackered the next day (and subsequent days if you do something big enough). The maths in this just reinforces that idea.

There's nothing in human physiology that says that fitness accrues over 42 days and fatigue dissipates over 7 days, but I'm guessing that's the best numbers you could pick that fitted the model at the time.

Much like HRmax = 220 - age is, in most cases, utterly rubbish (because it's only meant to be an estimation in the absence of other data, if you have data that contradicts it then use that, not the dodgy formula). It's that any other number than 220 gives even an even worse fit population wide. 220 is the number that is the least worst fit.
"Yes please" said Squirrel "biscuits are our favourite things."

jiberjaber

  • ... Fancy Pants \o/ ...
  • ACME S&M^2
Hopefully this link will work.  It's the spreadsheet I used to get me to LEL, an over engineered version of Greenbanks one with lots of other tabs.  I think I originally got it from a google search of PMC and TSS planning worksheets.  I found it quite invaluable but eventually opted to spend on Trainingpeaks as I could drag and drop plus it did all the hard yards for me of calculating HRTss and Power based TSS and mixing them all up in to my TSB etc.

https://drive.google.com/file/d/1S4At-ba7UgFdUpuBrPtcfGSS3fPCpscO/view?usp=sharing

There's also a whole load of other superfluous data in some tabs where I was trying to estimate from historic data the training impact of various rides - I've essentially now distilled that down to the usual Audax distances and 3 intensities (pub, tour, audax) and now just drag these templates around on TP to build my forward plan...  the good thing about TP is I can adjust fairly quickly when something gets in the way of the plan (weather, work, real life etc).  You can do that with a spreadsheet also, but if you are balancing a training plan from something like trainerroad around planned events, then it takes a bit more management to juggle the ripple effect on ramp rates etc when stuff moves.


Here's what my current TP looks like for the next couple of weeks - this week is mostly out the window already I am away with work and then elsewhere for the end of the week but it gives you the idea I guess... (I googled around for a discount code and used that).  The end of each row gives a summary of the weeks impacts planned vrs completed and where your form is expected at the end of the week.  As said upthread, you start to get a feel for what -20, -45 and -75 feels like and I can now correlate it quite well with how I feel like getting on teh bike or not and even mid ride on some Audaxs - though the latter might be psychological.



Other systems online are available with similar functionality, I think there is a British Cycling version of Today'sPlan with one of their subscription options for example.  Not all the options can do HR & Power TSS however, but almost all of them have a trial period to play with them over.  There's quite a lot of discussion though may be hard to find in the Base Training thread.
Regards,

Joergen

I used to used Training Peaks with a coach, and Today's Plan on my own. For me, I didn't get much out of training peaks (compared to a spreadsheet and stravistix), outside of it being ready for my coach. Today's Plan auto-generated sessions to fit goals, and auto synced with Zwift, which was nice. Then they hugely upped the price. However, I stored all the sessions for the year I had it, and can look back at my previous plans and reuse them.

Having stopped paying for these and Strava, all I'm missing is the ability to look at changes in my power curve. still, a year of battering myself made barely any change, so maybe I'm best off not looking!

I have started running with a club, and I'm my group are some serious ultra runners. Asking them about their weekly TSS, hr lactate threshold etc drew blank looks. I'm going to share this sheet + stravistix with them, and see what they make of it. They make it sound like the training loads they are doing are reasonably modest.

Sent from my Lenovo P2a42 using Tapatalk


zigzag

  • unfuckwithable
i find it quite fascinating and amusing to look at the data from the long* rides. while training following a structured plan is the optimal way, i wonder what is the best way to ease back into the programme after the inevitable "crazy" rides, training camps, holidays etc.? do the easy rides until fatigue goes away (this can take few weeks in some cases)?

*highest tss in a week so far was on giro delle republiche marinare audax - 2413 8)