Jumping into the workout designer and hoping to understand the principles is likely to generate more questions than answers.
As a baseline you need enough data to generate a valid fitness signature. Then you can play around all you want. Dialing in your signature will require a BT (breakthrough) ride or two whether it’s trainer based or outside following the workout guidelines, or a group ride that includes some hammering and a sprint.
As far as “useful numbers many hours into a ride”, IME the system just works. I assume because fatigue is factored in.
I know all that and I agree. However, there is doubt about how TP/MPA validates over time. I get that creating a 4 hour 100% workout is not your typical training, but I think you will agree that the fact that MPA doesn’t buckle one bit, is also strange. If you want to see a prediction on what your numbers will look like, 4 hours into an activity, it doesn’t help.
Excellent questions, and the discussion has been really good so far. I also just want to chime in with some thoughts & comments from our end at Xert.
By definition, threshold is the highest power that you can sustain that results in no net change in MPA (i.e. MPA will neither drop nor recover while riding at TP). As @Cyclopaat has already mentioned, for many athletes, the TTE at Threshold Power is typically around 40-70 min.
DS is an EWMA of XSS/hr with a time constant of 30 min. Since riding at threshold (with MPA = PP) is 100 XSS/hr, the difficulty will likewise top out at 100. If you try changing the workout to 4 hours at 101% TP, you’ll notice that this will change things a LOT.
As of right now, all XSS below TP is considered Low Strain (you can see that in this image, taken from our blog post on Work Allocation Ratios). Notice that below TP (250W for this athlete), it is 100% low strain:
However, its entirely possible that we haven’t fully modeled this correctly, and that there may actually be non-linear allocation of high (and even peak) strain when riding below TP, which would align with why some athletes find Sweet Spot training so effective.
It IS entirely possible to hit MPA & get a breakthrough 4, 5 hours into a ride (I’ve seen plenty of times), but it does require very conservative use of hard efforts earlier in the ride, as well as adequate hydration and nutrition throughout the ride.
Hope this provides some additional insight. Cheers!
How Is long term fatigue accounted for? The idea of LTP and energy decline makes a lot of sense to me, I’m just not seeing it actually applied to a workout.
I am fairly new to endurance riding and am pushing myself way outside what its accustomed to in these 4+ hour rides, so the fact that my performance is dropping drastically late in a long ride (more than the model says) is not a huge surprise. What is a surprise to me is the model seems to suggest that if I just break a few minutes to let my HIE fully recover I’m back to 100% which we all know is not true even if I was a pro.
I’ll also say that long term fatigue is piling on me more in one of these long rides than would normally be expected - that is the sheer amount of XSS (350 on the latest) for my level of training (40) leads to drastically more long term fatigue at the end of the ride than what I started with which is outside the realm of normal. Given a more typical workout for the level of training the long term fatigue is not so different between the start and the end of the ride so it is maybe insignificant?
As I train and improve my fitness and get more power readings and such I’m sure the numbers will start making more sense.
That makes sense, as you are stretching such an activity beyond your normal ones.
However, it’s intensity that puts a cap on how long you can maintain it, not your TL/fitness. Sure, better fitness implies a higher hardness level, but I’m at 110 and I can assure you, I still cannot ride endlessly at TP level.
I am interested to know about it more too. Maybe Scott can share his knowledge, ideas with us. I always felt impact of longer endurance ride (4 hours) with IF 0.65 more than 2 hours hilly ride with IF 0.9. Its not easy to calculate for low stress fatigue, because it can relate not only to fitness level, but bike fit position, hydration, food, pedaling efficiency, athlete type (fast/slow twitch) too.
I remember hearing in one of podcast, that LTP feels like TP when Endurance Energy is depleted.
In reality, we know that your ability to generate work above TP is compromised by glycogen depletion, calcium fatigue, and general fatigue of fast-twitch fibers themselves. To complicate the matter, you have all mentioned how nutrition and hydration will also play a role in your ability to perform after longer rides.
What LTP is meant to represent, is the lowest your TP will decline to, even after riding for several hours. What this means is that holding LTP near the end of a multi-hour ride might feel comparable to trying to hold TP when you’re fresh - not too hard, but certainly not easy. One of the many challenges that comes into play with implementing long-term fatigue is that TP, HIE, and PP do not deteriorate at the same rate, nor to the same amount. For example, since PP efforts are mostly fueled by PCr, your PP will actually be (relatively) unchanged, even after multiple hours - the only reason I can think of off the top of my head why it would decay would be fast-twich fiber fatigue. On the other hand, HIE is likely going to be affected the most by long term exercise, due to glycogen depletion and muscle fiber fatigue. The result is that your HIE capacity would decrease, indicating that MPA would drop more quickly for an effort that is XYZ watts above TP, when compared to being totally fresh. Finally, TP will decline, but it will reach a point where it could (in theory) be fueled aerobically by fat oxidation, so it will only decay so far. So, clearly many things to take into consideration (which is bad for modeling - generally keeping fewer parameters in the model is better) and I haven’t even mentioned the mental aspect of things…
I think the fact that we can model performance as accurately as we have so far is a step in the right direction. I think solving this would also be the holy grail of endurance performance.
Thank you for the explanation, I’d really love to have deep understanding of how all this works to fully optimize my training (to the best we understand). This actually is the place I’m not totally satisfied with xert so far - I think the pieces are here but it’s not all exposed or in the case of this discussion subject not fully modeled in (at least in a workout). I get it that most people don’t want to geek out on the details and just want straighforward advice, geeking is my nature however.
I’m looking at this now: http://baronbiosys.com/fat_and_carb_utlization/ - can you explain how that relates to this fatigue we’re talking about and the work contribution ratios? (feature request: I’d love to see these curves generated for my signature) I’m also interested how those change with fatigue.
I’m picturing MPA like a water tank with a float at the top (supply) - As the demand ramps so does the supply flow to match but it hits max flow at TP after which the level will start dropping. The lower the water level the less 'head pressure" so max flow out the bottom decreases as the level gets lower which eventually only matches what the float is pushing out when the tank is empty. As the time drags on the pump pushing water up gets tired so supply flow decreases eventually to the point it can only keep up at LTP, adding to the issue is the float also gets tired and sags to a lower level so the tank holds less max water.
Anyways, Interesting discussion. Thanks for the insight!
-Jeff
To be fair…is anyone even remotely close to having this modeled? The fact that we have users that are requesting accurate MPA hours into a ride just means that we’re on the right track. Further, if you were to get a BT 5 hours into a ride…which parameters change? TP? HIE? Endurance Energy? How much do each weigh in? Hopefully the discussion brings forward some of the many challenges that go into trying to model this.
Even if we started to model this out and projected MPA hours into a ride, users may likely consider the MPA value displayed too high or low (based on their
You would really find some of the papers written by Margaria & Hugh Morten in the 1970’s/80’s rather interesting. They were the first to start conceptualizing and developing a human bioenergetic model, which was comprised of 3 “tanks”, including Oxadative Metabolism (Fat & Carbs), Alactic Phosphoragens (i.e. ATP, PCr), and lactic (glycolytic - anaerobic carb) metabolism. They also defined the rate of water flowing out of the tank (W = total work being performed) as well as the metabolic rates of aerobic/anaerobic metabolism R1/R2 - there’s an exeerpt from Margaria’s work in the screenshot below.
Yes, Xert is certainly a step (or few) ahead of everyone else. I was more stating I find it somewhat difficult to figure out how to get what I’m looking for sometimes even though I know the information is here somewhere (maybe not exposed). For example I was just trying to figure out at what power level would be for my max fat burn as I have a theory that training up that system might be key (why base miles are so important). Ultimately I decided to do 20 min sessions at various power levels in the area I think it would be and reproduce my curve.
This projection many hours into a ride would certainly be a holly grail. The only reason for my disappointment is that I thought that was what Xert was advertising (almost but not quite). I think you have the understanding necessary to model the best case (upper limit) which would fix the glaring 4 hour at TP issue, but yes I agree truly accurate modeling would be near impossible as the variables would be beyond manageable.
Well, I’ve only had a power meter a little over a week and this whole lockdown thing has really messed up my schedule (I have kids now home while I try to work from home as well) so I haven’t really gotten the chance to do a XATA recommended workout or any of them in the player for that matter. I’m on a trial (which ends in a few hours) so can’t actually do a created workout, but it lets me play with the designer.
I did my little 20 min sessioning and found my burn max is about 185 watts (275TP/208LTP). Interestingly, I have a humon hex and 185 watts is about the threshold between a steady SpO2 and where it will start to dip (then stabilize, the downward trend starts at TP) . That’s another subject I’d love to explore more - muscle oxygen (and lactate).
Until this corona mess clears up I’m stuck indoors and time crunched so I think I’ll mostly spin there at 185 watts (which is a power level I can do things like watch serious shows and even read or work a little) with the occasional intervals to avoid degradation. Trial will expire for now, but I’m sure I’ll be back as a nerd needs his data I’m not sure if I’ll be locked out of this forum or not. I’ll find out soon.
All the more reasons to you start your subscription and gather weeks and weeks of power data following RECOMMENDED workouts based on your selected rider type and current IR (hours available per week).
I think you went off track as soon as you started to play with the workout designer. IMO that is an advanced feature for those familiar with structured intervals and power based training.
The core functionality of Xert is XATA using the library of curated (tried and true) workouts.
Try it and see. $10/mo; cancel at any time.
We’ll see how this Corona situation goes - time is extremely tight right now. That and my original lack of power meter did essentially ruin my trial. The reason I started tinkering was that my stats were way off due to the lack of good power data (235 TP 430 PP when I’m 275 TP and 1050 PP) so I wanted to figure out how to best correct it. I know this clashes with the no testing message, but it is necessary to test in situations like mine so I’d suggest including one in the trial.
Anyways, the current situation is (hopefully) temporary. Time to focus on what’s really important and get through it best we can. Stay safe and as best you can healthy!
Ok, I mostly understand this “3 tank model” now. Xert sounds roughly based on this in ideas but I’m thinking the model for a workout is actually the simpler 2 tank system (which would be doable to calculate on the fly in something like a garmin data field). As the lactic “tank” empties TP should trend towards LTP which doesn’t seem to happen in the projection, but does in real life so this makes a lot of sense to me.
For the 3 tank system I’d think more variables are needed. LTP (max from oxidative and constant assuming proper nutrition), TP (max from oxidative + lactic which will drop as lactic capacity is depleted), HIE (Phosphagen capacity), and a new one: EIE (Endurance Intensity Energy which is lactic capacity). I’m a software engineer by trade and thinking that modeling this would not be easy…
I’m not sure if I’ll be locked out of this forum or not. I’ll find out soon.