This. Include the zeros.

Gotcha - so the reason the avg is higher in this case is that I'm pulling the zeroes out of the AP but leaving them in the NP? I would have thought the "eliminate zeroes" setting would eliminate from both but maybe not part of the algorithym.

Thanks. Makes sense if that's the case.

You paid money for a powermeter, get the best data from it, make sure your record zero and also turn off auto-pause. I do not know why they even give you that option.

Thanks. I'll try that out on the flat routes. I live in a very hilly area where we often literally climb for 90 minutes and then barely touch the pedals coming back down. So keeping the zeroes in there made the data pretty meaningless. But for the flat courses, might make sense. Why would you say that gives more value? Maybe I just eliminate AP from my screen and focus on NP?

If your goal is to have the best data available to train with and get the most out of your training time, include zeros and turn off auto pause.

If your goal is to produce the highest number you can regardless that it is not based on any reality, do the opposite

Why auto-pause?

well it is also how you apply your power; normalized calculation squares the power ... so if you for example pedalled for a long time at 200, ie onyl 10 up from your avg 190 and then effectively coasted at zero for 10 mins home ... then NP will be lower

auto pause screws up the numbers. while siting at that stop light your legs are getting rest, but your computer isn't taking data(lowering average/normalized power). At the end of the ride you are happy with that 250w AVG then go do race using that number. In the race you never stop and with no stops and can only manage to sustain 220w, that is after you blew up trying to maintain too much power at the start. Then after the race you buy a new bike, because it must be the bike.

Lets say i ride hard for 20 minutes, then lay on the side of the road for 10 minutes, then ride hard for 20 minutes, then lay down for 10 minutes, then rider hard for 20 more minutes. That could result in a 60 minute power that is much higher than I could do without the rest. It is an extreme example, but it can show you how having some rest can result in higher numbers.

A better question is why have auto pause on? All it does it make your numbers higher than they actually are.

No, keeping zeros is not meaningless! It is showing what you ACTUALLY did. Stop letting vanity get in the way of good data.

You need to change your frame of mind from thinking about a ride as a indivisible block. Include zeros. The climb section gets it own analysis, the downhill is its own effort and together they form the entire ride. If you need to break it up then use laps, which will allow you to separate out NP and AP while you are riding if that helps.

It is a basic error to attempt the climb NP for the climb+descent time. Zeros do disrupt the NP calculation, and that is because your body is getting rest and this the physiological stress on the body is lowered.

Remember the definition of NP (from training peaks):

• By taking these factors into account, normalized power provides a better measure of the true physiological demands of a given training session - in essence, it is an estimate of the power that you could have maintained for the same physiological "cost" if your power output had been perfectly constant (e.g., as on a stationary cycle ergometer), rather than variable.

Taking the zeros out is what makes the data meaningless.

Remove zeros and you are no longer calculating average power, but something else. And that something else makes no physiological sense.

I mean, why not remove all readings below say 50W, or 100W? That would be just as meaningless as removing zeros.

---

http://www.cyclecoach.com
http://www.aerocoach.com.au

With a handle of R2, I can see why you might think that. :) However, the algorithm actually applies a 4th-order weighting...but only after first taking a 30 s rolling average.

While we're on the topic: so-called "Smart Recording" also screws up the calculation of normalized power (and hence TSS), although IIRC a few years ago Garmin changed their devices to always record at 1 s intervals when paired w/ a powermeter.

Excessive down-sampling of the data stream (e.g., older PowerTap and Polar head units) also screws things up...

I'm surprised that Garmin would include zeros for NP but not for AP. Doesn't really make sense. Nonetheless, you should definitely include zeros. For that matter, you should also turn off auto-pause, although I generally hit stop if I have to put my foot down for more than ~10 sec. More ego than anything else....makes my average speed look bad.

Mathematically speaking NP should be strictly >= AP, since NP = (sum(P^4))^1/4, it cannot be less than AP.

If you do a 20min max effort and lay down for 10min in the middle of it with auto pause off, do you think the number you get will give a better or worse reflection of what you could do for 20min than if you'd do the same with auto pause on?

Not that it matters that much when analyzing the files but during the ride I think auto pause can be useful in some situations.

---

---

BA coaching http://www.bjornandersson.se

i still ride with auto pause on set to go on "pause" at speeds of less then 2mph, i get the point that it might skew the number domes but if you don't have it on i would just be hitting stop at stop lights or when i go to the store to get a snickers. Without it on auto pause my rides would end up looking longer then they actually are. I am also basing all my racing numbers off a FTP test not on long rides. My 2 cents.

---

2024: Bevoman, Galveston, Alcatraz, Marble Falls, Santa Cruz

The other side of it though is did you ride for 5 hours or 4 hours and 20 minutes if you were stuck at lights for 20 minutes? Either way you end up with data that is off because total ride time is important too and the time spent stuck was not training time.

I don't think that is what he was saying. He said do 20 min hard, THEN lay down for 10, rinse, repeat 2 more times. With auto-pause your data would show you rode for 60 minutes, which is technically true, but your AVG power for those 60 minutes would be a number that most people could not actually maintain for an hour.

Also there is always the LAP button. turn off auto-lap and manually lap segments of the ride and know exactly how much power over that 20 minutes without screwing up the rest of the data.

Use the lap function so you will get good power numbers for the way up long climbs and can isolate that in the overall ride file if followed by long decent.

I was just giving an extreme example of where auto pause off will give a less accurate reflection of what you can actually do than auto pause on. Even in his example auto pause on will give a more accurate value than auto pause off even if not 100%. In the end it doesn't really matter that much when you sit down and analyze it as you evaluate the different segments and take all factors into account. Fooling yourself that you can do a certain number has nothing to do with auto pause.

In those examples given above using the lap button really makes no practical difference except it's slightly easier to analyze afterwards.

---

---

BA coaching http://www.bjornandersson.se

I don't turn auto pause off either, but there is a good argument that rest, is rest, so any time stopped should be included.

I guess I'm to lazy to turn it on and off and I rely on it for races, as I'll forget to hit start/stop in T1 and T2 otherwise. The downside is my run in transition gets included.

---

TrainingBible Coaching
http://www.trainingbible.com

Meh..i disagree.

I think get a more accurate reflection of your ride with the auto pause turned on. One or two minute stoplight/pee breaks in the ride aren't game changers in terms of trying to get a picture of your effort.

---

"Good genes are not a requirement, just the obsession to beat ones brains out daily"...the Griz

Include zeros (I agree with everyone else).

Auto-pause is a matter of preference.

For me I generally turn it on because I normally can ride in a fairly uninterrupted fashion, so therefore when I'm stopping it is likely I'm stopping for a prolonged period of time (coffee, etc). I used to disable autopause but so many rides got screwed up by not stopping the timer while grabbing a coffee, or worse, forgetting to restart it when I get back on the bike.

If you're riding somewhere with lots of forced stops (e.g. traffic lights) then I'd not have auto-pause as they're more like mini-recovery between intervals.

True, funny and hits a little close to home......I love it.

So I'm getting the sense that maybe I should start including the zeroes? :-)

Thanks all. Will do. Very helpful!!

I do not know if I understand your question. Are you asking if you take a break of 10 minutes during a 20 minute effort? So 10 min hard, 10 min rest, 10 min hard? Without autopause my best 20 minute power during that would be much lower than with autopause turned off.

If that was not the question and more a question about my scenario, my scenario was not really about the 20 minute power, but the 60 minute power that would result with auto-pause being turned on.

I will admit that many times I have forgotten to stop recording when I enter T2. So that is a downside, but you can always edit your ride files afterwards to chop off the ride after you have stopped moving.

If they are not a big deal, then why do you have auto-pause turned on?

Yea, longer breaks like a long coffee stop on a group ride I am unsure on stopping. For example if I ride my bike to the ferry and then take the ferry (~30 minutes plus any wait for the ferry), I would not have my computer running the whole time. I would either just have seperate rides or pause during the ferry ride. But people having multiple 5 minute breaks during a 100 mile training ride to simulate an ironman are really setting unrealistic numbers if they want to simulate a race where they would not be stopping.

See bolded above.

---

"Good genes are not a requirement, just the obsession to beat ones brains out daily"...the Griz

Don't forget the 30 s rolling average.

Lower yes. My question is which number is closer to what you can actually produce for 20min in my scenario(or if we did the same thing in your scenario for that matter)?

None of which is perfect but auto pause on will always give you a more realistic number if you look at average power during a ride except for in the most extreme scenarios. Like I said earlier you always have to consider the circumstances when you evaluate the ride afterwards anyway so I don't really understand why including non moving data in the averages is better.

---

---

BA coaching http://www.bjornandersson.se

I'm going to do a minor hijack of this thread with a quick somewhat related question.

From what I understand V.I. is the ratio of NP to AV. Am I correct in thinking that a training ride with several stops or slowing down for traffic will have a higher V.I. than say a race that should be a relatively uninterrupted effort and more focused steady ride? Also, will a hilly route cause a higher V.I. typically? I'm somewhat new to power training and just trying to slowly understand. On a normal steady state training ride my V.I. is typically > 1.1.

That's my take on it. I want to see what my average power was while riding and I don't think a few short stops here and there change the ride much. Plus, Trainingpeaks shows the total time in parenthesis so I can see how much time was spent stopped and I take that into account and minimize it on key sessions.

If I'm going to get so anal about my data that I'm worrying about a few minutes at stop lights, I'll just do all my riding indoors.

Yes VI is the relationship between AVG power and NP. One extreme example of this is a hilly group ride I did a few weeks ago, with a lot of waiting to regroup at the top of the hills. no pause, no zeros, my AVG power was 162, but my NP was 231, so my VI was 1.43. On that same ride last week by myself, and I didn't hit many traffic lights, my AP was 218 and my NP was 256 for a VI of 1.17. During a race on a flat course with very few turns your VI should be close to 1.00.
To me I think of VI as the number of "matches" I can burn on the bike during a race. What that number is however will be different for every person and different race distances

Rather than starting a new thread I'm going to hijack with my mostly related question. When displaying average power on a head unit what smoothing is typically used to ride along with? The case I'm looking for is when I'm doing a 5 minute interval and trying to hold 250 watts power, do I want 3 second, 5 second, or 10 second average to try to maintain the interval?

I go with 3s and 30s averaging. 3s gives you a smooth but recent view of power and then 30s gives you wider range. So if you're trying to hold 250w and you look down and your 3s is 280w, but your 30s is 260w then you know you just spiked recently. If you're doing 280s on 3s and also on 30s then you know you've been going quite hard for a longer stint.

I never look at direct power while riding. I always leave it on 3sec power so the numbers don't jump all over the place.

Regarding auto-pause...since I base most of my training around ride time, I keep auto-pause on. Otherwise that 4 hour ride would really only be 3 and half hours with a break in the middle (for instance).

If auto pause is screwing up your ride so much that you can't figure out your pace for a race you need a new to go find a new route. A 1-2 min rest is going to have a near zero effect on your ride and power output. Yes if you're doing an FTP test you shouldn't have stop lights in the middle of the test, if stop lights are screwing up your training during a regular ride you are way over thinking this.

Because sometimes you have to stop? Why would I want it averaging zeros every time I have to stop at a sign or red light?

Is that really how you ride?

Numbers except for, you know, actual time riding.

To get the most accurate numbers. You want those zeros, because you are not producing any watts and are resting.

It is an extreme example, but it is to illustrate how not using zeros will lead to inaccurate data.

Well I think neither will give the best 20 minute better test, but having auto-pause can change your MMP and make it inaccurate. Doing this every ride and it can add up.

Average power is the total mechanical energy of the ride divided by the total duration of the ride (i.e. finish time less start time).

Anything else is not average power.

---

http://www.cyclecoach.com
http://www.aerocoach.com.au

This. Same thing with people saying they averaged 20 mph on their ride while taking the stops out of their average. The effort is not the same, and if nothing else you are inflating your FTP just standing there. Believe whatever you want to believe about your power, but if you are using auto pause you are doing yourself a disservice. Your inflated power number won't make you any faster, and taking out the zeros hurts your training by incorrectly reporting back to formulas that expect them to be there.

I could ride 150 watts for 2 hours and average just that.

I could ride 150 watts for an hour and then 4 hours later ride 150 watts for an hour again and my average power is 50 watts.

Does that hurt my training? It's just an extreme example of what you are saying.

jaretj

I'm going to assume 150 watts is nowhere near your threshold, but I really don't know. Try this differently.

What if your threshold is 275 watts. Do you think these are the same workout?

Workout 1
7AM 250 watts for 1 hr
10AM 250 watts for 1 hr

Workout 2
7AM 250 watts for 2 hours

Do you think you could even complete workout #2?

If you ran 8x800 @2:30/800 with 3 mins recovery do you record the workout as a 20 minute 4 miler?

Recovery matters to your body, so it should matter in your record of the workout.

Sure it matters, which is why you don't do a route with a traffic light in the middle of a planned interval. Your biking example is ridiculous. We're not talking about 2 hours between workouts. We're talking about 1-2 mins waiting for a stop light.

That is an exaggerated case of course, but I think it plays in the micro version just as well. Take a workout where you are doing cyclocross starts. Typically this workout involves generating 700+ watts from a standing start, waiting 2 minutes and doing it again. you might do 8-10 of these, where the AP in the start is 400+ watts for a minute, WAY over threshold. If I had auto pause on while I was recovering between starts, I record a workout of 10 minutes at 150% of FTP. In reality I had a workout with a duration of 30 minutes with an average output of 133 watts.

And yes I do this workout from time to time and it looks like this:

30 minutes sustained riding at 80% FTP followed by 10 standing starts with 2 minutes recovery between efforts.

which is 30 minutes at 192 watts and 30 minutes at 133 watts for an average of 162.5 watts. With auto pause this is a 40 minute workout with an average of 244 watts.

Someone want to convert those two to TSS?

I'm gonna guess about 70-80 TSS. You should do the workout with auto pause on and see what the difference is.

I still don't think that's the same as your computer auto pausing during a stop light though. TSS is calculated off normalized power, not average power, and it's a rolling 30-second average so short intervals with rest breaks and starting and stopping your computer has more potential to throw that off than a short break at a stop light every now and then.

If the concern is TSS, I don't see how pausing or not pausing at stop lights is going to throw off TSS enough to worry about. Most people don't even have an accurate FTP in the system, partly because they do tests to estimate FTP rather than doing a 40K TT and because people don't test often enough to be able to say their FTP is 100% correct all the time. That's going to throw off TSS more than a stoplight.

There's a difference between understanding what average power is, and what training is effective. Just because people stop and/or incorrectly calculate average power doesn't mean their training is ineffective.

Where it can go off the rails is when people misinterpret the numbers to indicate something about their capability, and use that misunderstanding about their capability when making key training/racing decisions.

If they are using good analysis software that correctly calculates such values, such misunderstandings should be reduced.

---

http://www.cyclecoach.com
http://www.aerocoach.com.au

Which completely doesn't matter a bit because in races you...coast. At least in bike races.

Interval sessions? I don't do those on stretches of road where I have to stop. Normal riding? I'm for sure not counting time when I'm not moving. Totally jacks up ride time and skews speed and power numbers. I want to know what I'm doing when I'm pedaling, not stopping.

This seems to be much ado about nothing. I will never understand wanting to have numbers for when I'm not actually moving. It's a bike ride. Makes zero sense to have numbers from times when I'm not riding my bike. .

What actual data do you find so necessary as to include information from times when you're not riding?

I can understand interval or workouts (again, I wouldn't ever do intervals when I knew I'd have to stop, and when simply reset if I did have to stop) but I can't understand full rides. You're inflating ride time at best and giving yourself an inaccurate picture of what you're actually doing riding.

Think about an actual race situation where you're forced to stop. Do you genuinely think that rest is going to have a positive effect? No race where I've ever had to stop was for the positive (free laps in crits, wheel changes in road races, etc). It was a highly anxious time after which I then had to put in a significantly harder effort to get back to the speed/position I was at. "Rest" actually hurt had negative consequences.

That's not in any way relevant to a training ride, though.

I don't get your position. You are riding at 20 mph. If you consistently ride at 20 mph, that is your speed. Why include time when you're not actually moving?

Do you think stopping and restarting and still maintaining a 20 mph average is actually easier than just riding a single, non-stop session at 20 mph? It's most certainly not! It's much, much more difficult.

Someone saying they averaged 20 mph while going in and out of town with redlights and stops and all that had to go significantly harder than someone who rolled out their front door and averaged 20 mph without ever slowing down or stopping.

But you'd suggest the first one was "cheating" and doing themselves a disservice? That's really strange.

My ftp isn't affected a bit by my computer numbers including stops or not. If you think it is, you have some very odd notions of physiology.

Your post reminds me of a guy I met about 15 y ago who told me he could average well over 400 W during criteriums. Yes, he was a cat. 1, but not all that big, so I knew instantly that the number was incorrect. Turns out he was a Mac user, and therefore never downloaded his SRM PCIV (which excluded zero values from averages) as at the time the SRM software only ran under Windows.

Bottom line: time not pedaling = time resting, which increases the power you can generate while actually pedaling. Excluding zero values therefore provides a distorted picture of what you did/what you can do/the physiological demands.

Lets say you go do a 2 hour hilly ride on your own. Your avg power for the 2 hours is 300 watts (not including zero's). Your effort level was moderately hard.

Now, hypothetically speaking, go do another 2 hour ride (maybe a week later) that is completely flat with zero stopping so the zero's or no zero's are irrelevant here. I can guarantee you that you won't be anywhere close to holding the whole 2 hours at 300 watts. I'd bet you would average maybe around 200 watts for the 2 hours yet you gave the same amount of effort compared to the hilly ride.

The point of including zero's is ensure all your rides are measured and reported from the same viewpoint. Without including zero's, you cannot compare ride to ride (unless you ride the same exact route every single time).

---

blog

As Andy said, time spent producing no (or low) power impacts the power you can produce. I consider that useful information.

Aside from that basic principle, in some events it's also important to understand how much non-work you can get away with. It takes craft and skill to produce the least amount of power so that you can put the hammer down when it really matters.

But as I pointed out earlier, incorrectly calculating average power doesn't mean one's training is poor, it just means average power has been incorrectly calculated, or that you are calculating something else.

---

http://www.cyclecoach.com
http://www.aerocoach.com.au

You're responding to the wrong person.

I'm not talking about coasting. I'm talking about not moving, ie, at a stoplight. Your example has nothing to do with autopause.

If you could point me to my post where I ever mentioned not averaging zeros, I'd appreciate it.

I'm talking about stopping, as in not moving. There seems to be a communication error on your end.

Yes, I'm calculating the average power I produced when my bike was actually moving forward.

To be sure, that is different.

It's also different to calculating an average of all non-zero power.

---

http://www.cyclecoach.com
http://www.aerocoach.com.au

Yes, we said the same thing, I think you misunderstood my point to be the opposite of what you said. It is much harder to average 20mph while stopping, but if someone auto pauses their ride, and then says they averaged 20, I'm saying they cheated, not the other way around.

Whenever I do a TT type ride, my normalized power is lower than my average power. Crit/RR are always much higher normalized. And, yes, I'm including zeros.

40k example.

They cheated because they worked harder? Personally my rides with a lot of stop/starting tend to have a lower average speed (with autopause on) than longer uninterrupted rides.

Unless you're stopping for 5 minutes at a time I doubt the recovery is going to offset the repeated accelerations plus the faster cruising speed necessary to offset the time spent accelerating.

---

---------------------------------------------------------
All we have to decide is what to do with the time that is given to us. ~Gandalf

Here is the disconnect. The average number you are posting is irrelevant. Why would you care what your average power is over intervals including zeros? If you're doing the intervals what matters is the interval power not the interval plus rest power. Its the same thing as why stops lights are irrelevant. I'm not looking at my average power with a 4 min break for a stoplight. I'm looking at my average power, on the part of the ride that doesn't include stoplights which is most of it. Also, based on your analysis your ride was 60 mins, where you really only rode 40 mins. I would much rather know the latter number not the former since I want to know actual ride time, which is way more important to me.

Whether auto pause for relatively short periods has an effect on the validity of average power numbers for assessing your training is surely dependent upon the intensity of the ride? On a base ride at 65% FTP for 5 hours, I don't think the training effect would be significantly influenced by a few stops for a couple of minutes. However on a tempo ride at 90% FTP, that few minutes break would allow you to achieve the 90% level with much less effort and hence less training effect.

I was thinking of the example of the run/walk strategy for the IM marathon, where you are going at race pace, but the run segments are at a higher speed than you would achieve by a straight run with no stops. If you took the average of your running speed, then this would not be a useful figure for planning a IM marathon.

Not because they worked harder, because they got to 20 by taking out rest periods. Their garmin says 20 when they really only went 18.

Whatever software you're using isn't calculating normalized ("weighted") and/or average power correctly.

As a result, any TSS ("training load") values are also going to be incorrect.

No, I didn't.

Rest is rest, regardless of whether you're on or off the bike.

Nope.

I'm saying someone averaging 20, WITH autostop, who has to stop and slow down and accelerate, is working SIGNIFICANTLY harder than someone who averages 20 but never has to stop at a stop sign or stoplight and just maintains steady power/speed.

So what I'm taking from this is you think the first guy is "masking" their ride and actually has it easier because he's not including the stops. But that's absolutely not the case because slowing down and accelerating multiple times and still maintaining 20 is obviously harder to do.

And if that's not what you're saying, then I apologize.

Yeah, you did. Which is why you gave a completely irrelevant example about excluding zeros. The exact opposite of what I was discussing. .

No point in arguing about it. Your example is proof-perfect.

Yes, but I slow down at the stop signs and DONT have auto pause turned on, and still average 20, I worked way harder than they guy who had it on, because I would have had a higher average, let's say 22, if I had turned it on. That's my point. I still think we agree.

Exactly this.

And now he's "training" 50% longer. Absurd.

We don't. Your "logic" is ridiculous and makes no practical sense whatsoever.

You're simply twisting my example to fit your completely irrelevant one.

You talked about two people averaging 20 having different efforts and the one with autopause putting out less effort. When in fact, in the example I gave, it'd take someone with autopause who is constantly slowing and stopping considerably more effort to average 20 than someone who doesn't have autopause on but has open roads with little to no stopping.

So my example is exactly counter to your initial example and illustrates how NP would be much higher than a steady state to average the same 20. Now you turn that around with an example about averaging 22. So you averaged 22 with your slowing and stopping. Still significantly harder than a steady-state 22. Still totally counter to your notion of two equal speeds with autopause equating an easier effort. Not the case in this example.

Hold on, you and I both go on a ride, and you have auto pause on and I don't, and we both hit the same stop signs/lights, etc and I have auto pause on, and you don't, who has the higher average speed when we are done?

You right? And we did the exact same ride. Same effort, same everything. Assuming we are identical riders. But your speed was 20 and mine 19.

What I'm saying is that if I go back out and hit the same lights and don't have on auto pause and I average 20 this time, I worked harder than either of us the first time.

How is this example related to my example again? My example countered your one about autopause being easier. If you're not stopping at all, the autopause is redundant so your comparison to someone else that has autopause but slows and stops and all is incorrect as they are the ones working harder and have a higher NP. It's two different examples.

In any case, your notion of autopause doing a disservice to a rider is just b.s. You can't qualify something like that as it's of no practical importance. Any one with an accessible road and any sense is not doing long interval workouts through multiple intersections with lights and signs and all. And that's the only part that would actually be of significance, but even then, not so much in certain races, like...a bike race. Like I said before, in a bike race you coast anyway. That's the direction I was heading before I used an example that clearly didn't come across the way I would have liked. So apologies for the regression.

I don't really know what to say except this is the same point I made from the beginning, albeit worded very differently. I frequently encounter 50 year old men with bowling balls where their abs should be who tell me they can average 20 mph on the bike for x distance. I'm almost always shocked to hear this, until it occurs to me they are auto pausers. Here in Chicagoland there is no route you could take within 50 miles of the city where stoplights don't have an impact of roughly 3-5% of total ride time, unless you choose to ignore them altogether, as many do. My lunch loop crosses 2 major roads and I spend at least 5 minutes of my hour stopped. Just a fact of life here, so I maybe pay more attention to rest time than others.