Strava vs. Elevate fitness/fatigue/form differences

Triathlon Forum
1

I’m trying to train as smartly as possible because I want to be better fit. I’ve used Strava and Elevate fit/fatigue/form for about a year and have tried to read articles/watch vids on what it all means - and frankly since I’m pretty much a beginner they don’t make much sense to me, especially because the two systems give two different sets of “answers” from the same data. My reading on the calculations makes me think the calculations are the calculations and therefore the numbers should be the same - but again, I’m a beginner. For example, today Strava has 74/75/-2 and Elevate has 52.9/55.7/-10.1.

Do both/either of these measurements use variable HR method? Why are they different? Should I always go easier when my form is less than -10? Which one should I use? Basically, without too much technical information, how do I use the three numbers and which one should I use?

If I use elevate, today I should go easy (form -10.1)? If I use Strava, I can go hard today (form -2)? If I want to use VHR training, what would I use?

I just don’t know which one to use and how to best use it.

I appreciate all the wisdom here.

Thanks

2

Not totally sure since I don’t use Elevate or Strava much, but Strava doesn’t go into much detail on their adaptations of TSS/CTL/ATL. They do use forms that use HR if you don’t have power, but I think uses power if you have it.

Per my understanding Elevate is more of a pure Trainingpeaks-style PMC with classic Coggan-based calculations.

Shooting from the hip, I’d say just pick one. And I’d pick Elevate as it’s more transparent than Strava, and goes deeper into analytics if you want to go there.

And as an aside, use the #'s as an informational tool, but not as specific instructions on when to go easy, hard, etc. It’s one of the things to consider when tracking training, but not a be-all, end-all. I think Coggan used the quote to use it as a type of description of your training, but not prescription on what to do. (Well I think he was talking more about the W/kg-by-rider-class table, but I think it also applies to the PMC chart.)

3

And as an aside, use the #'s as an informational tool, but not as specific instructions on when to go easy, hard, etc. It’s one of the things to consider when tracking training, but not a be-all, end-all. I think Coggan used the quote to use it as a type of description of your training, but not prescription on what to do. (Well I think he was talking more about the W/kg-by-rider-class table, but I think it also applies to the PMC chart.)

Thanks for the information. What is the PMC chart?

4

And as an aside, use the #'s as an informational tool, but not as specific instructions on when to go easy, hard, etc. It’s one of the things to consider when tracking training, but not a be-all, end-all. I think Coggan used the quote to use it as a type of description of your training, but not prescription on what to do. (Well I think he was talking more about the W/kg-by-rider-class table, but I think it also applies to the PMC chart.)

Thanks for the information. What is the PMC chart?

The origins of the stress/fitness/fatigue style graph was something called the “Performance Management Chart” in software related to TrainingPeaks called “WKO”. And then copied/adapted in dozens of different forms, including TrainingPeaks’ main tool. There are also tons of variations on the formulas used. Some to try to improve on others…some to try to get around giving credit to the original inventors (Banister, Coggan, et al.).

I think the Elevate one tends to adhere to the original WKO-style pretty closely. What Strava does is a bit of mystery. The numbers from one should not be compared to the numbers on the other. (but in theory they should track each other to some degree in general “shape” of the resulting curves). If they don’t, that’d be interesting.

5

Thanks for the help. I think I understand it better now.

6

Here’s some more info on Strava’s metrics. They use Dr. Skiba’s presentation of cycling analytics instead of Dr. Coggan’s. However, I believe the PMC concept is Dr. Coggan. Training Peaks trademarked Coggan’s terms, Skiba has allowed for open-source use.
https://science4performance.com/tag/bikescore/

More information: Analysis of Power Output and Training Stress in Cyclists: The Development of the BikeScoreTM Algorithm.
http://citeseerx.ist.psu.edu/...ep=rep1&type=pdf

PMC Concept from the author. You can use just about any metric to populate the PMC. Oftentimes, 6 PMCs based on 6 metrics will yield exceptionally similar charts. There are a number of spreadsheets available online for download that will allow you to play with the PMC.
https://www.trainingpeaks.com/blog/the-science-of-the-performance-manager/

7

And as an aside, use the #'s as an informational tool, but not as specific instructions on when to go easy, hard, etc. It’s one of the things to consider when tracking training, but not a be-all, end-all. I think Coggan used the quote to use it as a type of description of your training, but not prescription on what to do. (Well I think he was talking more about the W/kg-by-rider-class table, but I think it also applies to the PMC chart.)

Thanks for the information. What is the PMC chart?

The origins of the stress/fitness/fatigue style graph was something called the “Performance Management Chart” in software related to TrainingPeaks called “WKO”. And then copied/adapted in dozens of different forms, including TrainingPeaks’ main tool. There are also tons of variations on the formulas used. Some to try to improve on others…some to try to get around giving credit to the original inventors (Banister, Coggan, et al.).

I think the Elevate one tends to adhere to the original WKO-style pretty closely. What Strava does is a bit of mystery. The numbers from one should not be compared to the numbers on the other. (but in theory they should track each other to some degree in general “shape” of the resulting curves). If they don’t, that’d be interesting.

I tried to reverse-engineer the Strava algorithm at one point – results below:

Fitness = CTL or Chronic Training Load = \Sigma_i=0^\infty (1/42)1.3RE_i*exp(-i/42)

Or the sum of 1/42*1.3 times your relative effort (RE) times e to the i/42, over all days from today (i=0) backwards indefinitely (i=infinity) where i is the number of days ago that you did the activity.

Fatigue = ATL or Acute Training Load = \Sigma_i (1/7)1.3RE_i*exp(-i/7)

Form = TSB or training stress balance = Fitness = Fatigue

Broadly speaking, this means Fitness has about a 6-week memory, fatigue has a one-week memory, and if you exercise monotonously everyday with the same relative effort RE=X, you will end up converging (over a span of a few months) towards a state with Fitness=Fatigue=1.3*X.

Strava’s relative effort (RE) is akin to Training Stress Score or TSS, and is based on either power zones or heart rate zones. I only have experience using it with HR zones, and do not know the formula, nor can I find anything clear about the nuts and bolts of how it is calculated. For me, relative efforts for running races are roughly: 5k ~60, half marathon ~250, marathon ~400. An article about relative effort suggests median users have a 10k run relative effort of ~120 and a olympic bike leg split relative effort of ~190.

8

I’m not an expert.

My ride at Midsouth 100 seemed to be way underestimated in Elevate and way over estimated in Strava (at first). I was on a bike with no PM and didn’t wear a HRM, so it’s my fault more than the software.

In Strava, you can rate the effort. It was a hard race in mud that lasted about 7 hours. If I rate it as “hard” then the TSS value for that workout is WAY high. It showed 826, which is over 100tss/hr, so not possible. I edited it and called it moderate (right in the middle of the scale), and it came up with 467. That’s probably somewhat close, but at no point in the day did it feel like a “moderate” workout. It was downright hard. But in strava’s terms, I guess “hard” would mean at or above FTP the whole time, which for 7 hours isn’t possible.

Elevate gave it a stress score of 186 on that ride. That’s based on the estimation that i did the ride at 100w. It must have estimated watts off of speed (makes sense) but obviously it doesn’t know whether I was in thick mud all day or primo tarmac; Or if i was going 12mph on technical single track (hard) or 12mph coasting around my neighborhood (easy). So in this case, Elevate estimated way low.

In the end, I guess both are only as accurate as the data you put in. I shouldn’t expect either one to accurately guess the actual TSS of a workout when i don’t have HR or power data. For me, it seems a little easier to update the effort slider in Strava after a workout in order to get a close estimation. For me, Elevate says 74 for CTL and Strava says 81. Not a huge difference, but still significant.

9

Thanks folks for more information.

I read somewhere that form shouldn’t be below -35 (not sure if that means Elevate Form or Strava Form). Is this because I’d be too Fatigued to get much if any benefit because I’m already spent or is it because of the increased possibility of injury?

I try to keep my RE on Strava between 600-700 on my regular training weeks (not a recovery week and not a week when my schedule keeps me from training as much). Is this “good”, “bad” or a personal thing in that everyone’s number is different (assuming profile data is set correctly)?

10

Thanks folks for more information.

I read somewhere that form shouldn’t be below -35 (not sure if that means Elevate Form or Strava Form). Is this because I’d be too Fatigued to get much if any benefit because I’m already spent or is it because of the increased possibility of injury?

I try to keep my RE on Strava between 600-700 on my regular training weeks (not a recovery week and not a week when my schedule keeps me from training as much). Is this “good”, “bad” or a personal thing in that everyone’s number is different (assuming profile data is set correctly)?

Again, I’m not an expert. I’m sure somebody will chime in and tell me how wrong i am.

-35 is pretty low, but i wouldn’t say your form should never be below that. During big training blocks I tend to stay below that level for a week or more at a time. And certain big events will put me back under that. I wouldn’t necessarily plan on being at -35 going into a big race i want to perform well at, but i don’t think it’s bad to dig yourself a hole like that, followed up by some recovery.

My personal limit is about a CTL of 100. Maybe without a full time job, marriage, home, etc I’d have the motivation and durability to get higher. But averaging 700 TSS points per week is about as good as i can do. Above that, I start to lose drive/motivation, OR my patellar tendonitis starts to flare up. I can only imagine that the pro ironman athletes are way above that.

11

So is RE the same as TSS (I’m just a beginner)?

12

So is RE the same as TSS (I’m just a beginner)?

Similar in intent, but different in implementation. RE primarly (exclusively?) uses heartrate while pure TSS uses power. hrTSS uses heartrate instead of power. But I don’t think RE is pure hrTSS either.

13

Similar in intent, but different in implementation. RE primarly (exclusively?) uses heartrate while pure TSS uses power. hrTSS uses heartrate instead of power. But I don’t think RE is pure hrTSS either.

Okay thanks. I only have HR monitor and no power meter, so RE is good for me.