Login required to started new threads

Login required to post replies

Live TSS calculation with altitude correction? Garmin App, TrainingPeaks, Xert?
Quote | Reply
Is anyone producing a TSS calculation that adjusts for elevation either live or post-hoc? It seems % changes in FTP could be initially based on tables, and then updated as the user spends more time at various elevations. And then, if the rider goes from 5k to 10k and back in a single ride, TSS is adjusted as the rider changes elevation. Thanks for any feedback.

A more general and naive question: if racing by power, how do folks adjust their goal power when going from, say, Boulder to sea level?

wovebike.com | Wove on instagram
Quote Reply
Re: Live TSS calculation with altitude correction? Garmin App, TrainingPeaks, Xert? [milesthedog] [ In reply to ]
Quote | Reply
Don't lose the forest for the trees. Unless you are oscillating between 12k and sea level it's not going to matter much. Individual physiology could come into play here, but that has more to do with going above 8-9k or higher.

Also note that TSS isn't a great point representation for intensity duration for desired adaptation. The point based intensity duration curve over values time in the tweener zones and not at the periphery. Get some lactate testing and thank me later.

Re: Boulder to sea level - The more fit you are the less the delta will be. When folks are super fit they can generally hold sea level watts up to a little over 7k.

36 kona qualifiers 2006-'23 - 3 Kona Podiums - 4 OA IM AG wins - 5 IM AG wins - 18 70.3 AG wins
I ka nana no a 'ike -- by observing, one learns | Kulia i ka nu'u -- strive for excellence
Garmin Glycogen Use App | Garmin Fat Use App
Quote Reply
Re: Live TSS calculation with altitude correction? Garmin App, TrainingPeaks, Xert? [MarkyV] [ In reply to ]
Quote | Reply
PPP: It is called training stress score and not training adaptation score for a reason.
Quote Reply
Re: Live TSS calculation with altitude correction? Garmin App, TrainingPeaks, Xert? [milesthedog] [ In reply to ]
Quote | Reply
You could do this using the elevation-corrected power function ("ecpower") and mFTP from the P-D model in WKO4.
Quote Reply
Re: Live TSS calculation with altitude correction? Garmin App, TrainingPeaks, Xert? [MarkyV] [ In reply to ]
Quote | Reply
MarkyV wrote:
The more fit you are the less the delta will be.

In fact, the opposite is true: the fitter you are (i.e., the higher your VO2max), the more you tend to be impacted by altitude:

https://www.ncbi.nlm.nih.gov/pubmed/8806931
Quote Reply
Re: Live TSS calculation with altitude correction? Garmin App, TrainingPeaks, Xert? [Andrew Coggan] [ In reply to ]
Quote | Reply
Would HR data help adjust the score? Or is this not calculated in the TSS? Either way, it would probably be beneficial to pay attention to HR and compare when at altitude.

USAT Level II- Ironman U Certified Coach
Quote Reply
Re: Live TSS calculation with altitude correction? Garmin App, TrainingPeaks, Xert? [MarkyV] [ In reply to ]
Quote | Reply
[quote MarkyVRe: Boulder to sea level - The more fit you are the less the delta will be. When folks are super fit they can generally hold sea level watts up to a little over 7k.[/quote]
Can you back this up with data? This is opposed to all I have read and the data I have from my athletes. In fact, I typically can see differences between 5000-7000 ft. in my athletes. There is a strong individual response to altitude, some are effected less than other but I don't know of anyone who can hold the same watts at sea level and 7,000 ft. I have National Champion rider that sees up to 10% drop from sea level to 5000 feet.

BoulderCyclingCoach.com
Quote Reply
Re: Live TSS calculation with altitude correction? Garmin App, TrainingPeaks, Xert? [Andrew Coggan] [ In reply to ]
Quote | Reply
Andrew Coggan wrote:
You could do this using the elevation-corrected power function ("ecpower") and mFTP from the P-D model in WKO4.

Will check that out

wovebike.com | Wove on instagram
Quote Reply
Re: Live TSS calculation with altitude correction? Garmin App, TrainingPeaks, Xert? [MarkyV] [ In reply to ]
Quote | Reply
If you live in Tucson at 3,000 feet and ride up Mt. Lemmon every day, how your power curve shifts as you ascend to 5k or 6k and then on to 9k is hugely dramatic, or at least it is for me, and a non-elevation weighted TSS assessment (or NP / IF assessment) would be very wrong.

So if you are living low and riding high like that, or living high and riding low (Leadville to Denver or Dallas) -- it would make a big impact on the power curve according to some known anticipated loss or gain as you perform/ride at different elevations, thus your pacing, your interval / your ride wattages, and your stress totals could be pretty significantly different.

I would guess the underyling inputs are individual and hard to quantify for practical use without some study, but I'm sure tour teams have a scientific perspective on it for managing performance of athletes and pacing in races when they are climbing up a mountain at high elevation but more or less living and riding otherwise at sea level. What my FT is in Leadville or or Denver or Dallas depends on where I live and set my baseline FT.

All that said, I think to the point about the forest and trees I agree, because TSS and training load numbers have much bigger sources of error, chief among them what the athlete inputs as his or her FT, which isn't exactly a knowable moving target, especially as you ramp fitness from zero, and also (to your point) the limitations in the way NP is derived from single-second data. So if the OP is using it to perfect IF and its derivatives (TSS, CTL, ATL) while living in Denver and doing some training at higher elevations and lower elevations here and there, that's getting a little anal for a figure that is by definition going to be imprecise. If >80% of my stress is happening between 5,000 and 6,500 feet, then my TSS figures are pretty good, assuming I have a great handle in my FT (which is a big assumption)
Quote Reply
Re: Live TSS calculation with altitude correction? Garmin App, TrainingPeaks, Xert? [Once-a-miler] [ In reply to ]
Quote | Reply
Once-a-miler wrote:
Would HR data help adjust the score? Or is this not calculated in the TSS?

TSS is based on power, and power only.
Quote Reply
Re: Live TSS calculation with altitude correction? Garmin App, TrainingPeaks, Xert? [kileyay] [ In reply to ]
Quote | Reply
kileyay wrote:
I would guess the underyling inputs are individual and hard to quantify for practical use

Not as hard as the effects of heat/humidity (which isn't to say that individual differences don't exist).

kileyay wrote:
TSS and training load numbers have much bigger sources of error, chief among them what the athlete inputs as his or her FT, which isn't exactly a knowable moving target

That's where the P-D model can come in handy. Sure, it can bounce around some, but it is one way of obtaining an objective estimate.
Quote Reply
Re: Live TSS calculation with altitude correction? Garmin App, TrainingPeaks, Xert? [Once-a-miler] [ In reply to ]
Quote | Reply
Once-a-miler wrote:
Would HR data help adjust the score? Or is this not calculated in the TSS? Either way, it would probably be beneficial to pay attention to HR and compare when at altitude.

Heart rate is not a variable in TSS. But I think you can get a gauge of your response to elevation increases and decreases using heart rate by analyzing the correlation between heart rate and power as you do a steady climb starting from sea level or 3,000 feet and going up to 9k (Tucson) or even up to 12k feet (Maui). But yeah, you'd have to test either in an altitude tent or somewhere like Maui, while controlling the conditions that you know impact the absolute and the variability within my heart rate for a given day (for me, total training stress across sports, caffeine intake, sleep, and temperature/hydration levels are known variables that impact heart rate) and on a specific level within a given ride (cardiac drift, temperature again).

Physiologists hate when plebes try to come up with a way to "adjust" TSS based on heart rate because heart rate is so impacted by those variables, and how those variables impact a given individual varies wildly from individual to individual. When I'm feeling my worst, my Garmin 935 thinks I'm in my best condition (10 minutes into a ride or run) because look how fast you're going at such a low heart rate when the reality is my heart rate is altogether anemic because I'm shelled from training stress. But if you spend enough time looking at your heart rate and power numbers, day after day, month after month, year after year, then you have a pretty good sense of what those things are for you.

When I lived in Denver I couldn't wait to go on a business trip to sea level and crush my tempo / threshold efforts 25 seconds per mile faster than I could back home. I felt like superman with an extra oxygen reservoir or something.

Going the other way is similar for me: a couple weeks ago, I was up from sea level in Tucson and climbed up Mt. Lemmon and Mt. Graham within a few days of each other but it reasonably similar weather conditions (~3,500 to ~9k feet). These are steady climbs and I made an effort at holding steady power within a tight band. My heart rate (netting out typical in-ride cardiac drift) was ~15 beats higher for the same wattage at 8,500+ than it was at 3,500 (which is basically sea level equivalent for me in my testing) and that's with an imperfect weighting to the significantly cooler weather at the top, which has a material impact on heart rate for me. All else equal, with respect to elevation, things start separating for me at 4k always -- by the time I'm in Reno at 4,500, how fast I can go is significantly different from how fast I can go at sea level, which I only learned, of course, when I showed up in Reno from Texas and blew up my race because I was trying to run what I could in Dallas -- and I can see this in my Lemmon/Graham data on a same day/ride basis (with the elevation as my axis).

To Coggan's point, when I ride with my friends up Lemmon or Graham, they seem to see fairly similar shifts with their heart rate/PE/power tracking as they climb. But when I do a big ride with my friends who are impervious to humidity on a very humid/hot day, well it's pretty obvious to me that I just shouldn't be riding with those friends on very hot and humid days because my numbers are much much more impacted on account of those conditions than theirs are, even if at 70 degrees with 20% humidity our FTs / FTs per kg are identical, and I end up dropped -- this is because for me, the impact of humidity and high heat (with the related variables like hydration) on my heart rate and effectively on the power I am capable of generating seem to be significantly greater than altitude.
Quote Reply
Re: Live TSS calculation with altitude correction? Garmin App, TrainingPeaks, Xert? [Andrew Coggan] [ In reply to ]
Quote | Reply
Andrew Coggan wrote:
kileyay wrote:
I would guess the underyling inputs are individual and hard to quantify for practical use


Not as hard as the effects of heat/humidity (which isn't to say that individual differences don't exist).

For once I think we basically agree.
Quote Reply
Re: Live TSS calculation with altitude correction? Garmin App, TrainingPeaks, Xert? [MarkyV] [ In reply to ]
Quote | Reply
MarkyV wrote:
Don't lose the forest for the trees. Unless you are oscillating between 12k and sea level it's not going to matter much. Individual physiology could come into play here, but that has more to do with going above 8-9k or higher.

Also note that TSS isn't a great point representation for intensity duration for desired adaptation. The point based intensity duration curve over values time in the tweener zones and not at the periphery. Get some lactate testing and thank me later.

Re: Boulder to sea level - The more fit you are the less the delta will be. When folks are super fit they can generally hold sea level watts up to a little over 7k.


You've kind of been beat up on the last sentence, so I'll pile on.

N=2 is that there is no way same/similar watts will be held on a time basis when comparing sea-levelish numbers to 7,000 feet numbers. I think that the reduction starts to become apparent at 4,000 feet or so. Again, N=2 comparing Austin to Sedona to Flagstaff.

Edit: Just read some of Kileyay's comments, he mentioned the 4k threshold as well.

Especially true for an altitude responder (me) vs non (Amy). I gradually got better at altitude, but was still hard pressed to equal my Austin numbers. She didn't. She mostly got more tired trying to hit the numbers...not including the last summer in Flag, which was essentially pre-leukemia.


Brandon Marsh - Website | @BrandonMarshTX | RokaSports | 1stEndurance | ATC Bikeshop |
Last edited by: -JBMarshTX: Sep 26, 18 8:00
Quote Reply
Re: Live TSS calculation with altitude correction? Garmin App, TrainingPeaks, Xert? [kileyay] [ In reply to ]
Quote | Reply
Speaking of power, heart rate, and power, here is a comparison I did for Dean Golich when developing the ecpower algorithm...note the closer correspondence of heart rate with ecpower vs. actual power beyond about 2000 s (corresponding to >1500 m):


Quote Reply