MattTheHat 

I've been using a power meter for a couple of months now. I have a fairly good idea of my FTP because I happen to have a couple of Strava segments that are very close to 20 minutes in length that I can ride at near 100% effort. As I do more research into power-based training it seems like I need to do a better FTP test. So, I figure I'll do the 20-minute version. I have an area in mind where I know I can ride without the need to adjust effort for traffic. This area is just far enough from the house to provide a good warm up with a couple of hard, short efforts.

I've never done this, so I thought when I'm ready to begin the test, I'll hit the "Start Ride" button on my ride computer and then "End Ride" when I finish the 20 minutes, and upload that to Strava.

I have several questions:

1. Does it matter if the area I ride for the test has rollers as long as I put in reasonable effort on the downhill portions? It seems ideally I'd have flat or a slight uphill grade for the whole 20 minutes. It's fairly flat in my area, but I can't think of anywhere that won't have at least some rollers. (Say four or five little roller of maybe 50' elevation gain, up to maybe 5% grade.)

2. I normally do most of the uphill part of the rollers out of the saddle, so power there is going to be twice my FTP. Would it make for a more accurate test to stay in the saddle to keep the power more...average?

3. When I review the segment data, do I use the Average Power number for the Strava ride or the "Weighted Average Power" number?

MattTheHat 

And from looking at my power curve data, it looks like my FTP will probably be closer to 85% of my 20 minute number, not 95%.

redlude97

first, an ftp test is not simply 95% of your best 20 min effort, you need to do the anaerobic clearing efforts

https://www.hunterallenpowerblog.com...-ftp-test.html

Second you use avg power. If you have strava summit just look at your power duration curve for the test

gregf83 

For an FTP test you should be keeping your power steady. This means gearing down on the hills so you maintain constant power, don't stand up and keep the power up when going down the hill. If you get going too fast just apply the brakes while pedaling. I used to have a 4 mile loop I would use for steady intervals but some of the pavement was rough on the corners and I needed to slow down so I would just drag my brakes while continuing to pedal at constant power. It worked quite well on my loop.

If you're just using the 20min test to gauge progress then doing an all-out 20min effort is fine. If you're trying to predict what you could do for an hour (i.e. something close to FTP) then the recommended protocol (by Hunter Allen) is to do a 5min all-out effort prior to the 20min effort. Most people don't seem to bother with this which is likely why their actual ftp is less than 95% of their 20min power.

edit: I see redlude posted the detailed protocol above.

MattTheHat 

Yep, I’d noted the 3 1 minute intervals but not the 5 minute all out. Between all of that and keeping up with the different percentages of FTP, of which I don’t even know for sure, the 60 minute test just might be simpler. I’m just going to have to find an acceptable piece of road.

MattTheHat 

Steady power. That makes sense, especially if I’ll be using the average power result.

My aim is to get an accurate basis for keeping up with exertion levels vs recovery. If I use an FTP that’s too high I’m likely to over train. If I use an FTP that’s too low I’ll under train. Neither is the end of the world, but I’d like to get it right, especially for predicting where I’m likely to over train and need more recovery. I ride every day because I enjoy it. Some days I need to ride with less effort to aid recovery. I’m not currently getting that right and several times a month I’m getting really tired. With a correct FTP to plug into Strava, well, that seems like a good place to start.

MattTheHat 

I found an interesting article which explains several things I’ve noticed on my power curves in Strava. https://science4performance.com/2018...a-power-curve/

And it answers something I’ve wondered about. The article states that if your ride data includes several hard efforts, the 60 minute maximum power would match FTP closely if it showed weighted average power. Alas, Strava does not offer that metric. The power curve is showing average power.

Very interesting nonetheless. And a good explanation of what type of information can be learned from the curve.

asgelle

I don't quite follow. FTP is useful for quantifying the volume of training, and yes, if your FTP is off, your calculated training volume will be as well. However, the other side of the equation is how much load you can handle at any time. So if your FTP estimate is too high, the load you can tolerate based on it will be lower than a more accurate value would predict and vice versa for a low FTP estimate, but unless you apply some arbitrary training limit not based on your individual abilities, it shouldn't matter (in terms of under/over training).

MattTheHat 

Yes, that’s basically the problem. I have a very limited understanding of what I should be basing my training levels on. With lack of understanding, I’m trying to rely solely on Strava’s weekly Relative Effort recommendations. I’m finding them too low. So I’ll step it up way too much the following week. And then I get really tired and just struggle for a few days with low energy. Other factors are obviously in play such as diet, sleep habits, stress levels, etc. I do pretty good keeping those things pretty static.

The long and short of this is that I probably need to stick more closely to the Relative Effort recommendations and if I feel I need to increase weekly effort, do it in smaller increments.

asgelle

No. Don’t follow arbitrary recommendations. Find out what your personal limits are. You might consider a book on training and racing with a power meter.

MattTheHat 

Well, I know my limit lies between Strava’s Relative Effort of 700 and 1,100 so I’m narrowing it down. 😁 I do need to find a good book that’s not too technical.

Drew Eckhardt 

FWIW, you can use Critical Power calculated off short (2-3 minutes) and long (15-20 minutes) maximal efforts instead.

Basically, you have some Critical Power you can sustain plus an Anaerobic Work Capacity or W' you can spend over a long or short period.

Dr Coggan said
https://staminist.com/read.php?1,434
Dr Skiba comments on Slowtwitch
https://forum.slowtwitch.com/forum/S..._Test_P2652621

DrIsotope

You pour enough data into Strava over a long enough period and it's algorithm can get pretty good. My last FTP test put me at 290W (which is what I have input to determine my training load, etc,) and Strava's estimated FTP via the Power Curve graph is 287W.

It's at the very least accurate enough to keep me from blowing myself up endurance-wise every week.

OBoile

Use Chrome and the Elevate add on... you'll get a whole bunch of new statistics.

MattTheHat 

Ooooh, interesting. I'll check it out.

burnthesheep

I'll take 1 and 2 together. It should be a steady/consistent all-out effort. You shouldn't have anaerobic spikes to double ftp during the effort, that means you aren't pacing it well. Double ftp seems silly high and poorly paced. Which begs the question, is it a suitable route? If you're that far over and then under at other times, it's not so good. Generally I stay in the saddle unless in the last 1/3 of the time I need a little 30 second stretch to then sit back down.

#3 , use the "analyze ride" button. Then click and drag over the 20 minutes of the ride you need. It will give you an average. Use that.

Lastly, your assumption of 85% is probably more realistic than a lot of people. So kudos there. Most people claim 95%. If you had their kids and put their johnson in a cigar cutter they still couldn't sniff 95% for an hour. I'd actually say 90 to 92% for most folks works. I've done some SS workouts for an hour sometime after an ftp test and usually can manage 90% without any fanfare or real difficulty.

If you can't make a good 20min test happen, don't worry about not having the common "brag factor" test duration of 20min or an hour. Some plans base it off of a combo of best 8min efforts. All you need is a reliable result with which you can setup your training.

If you're dead set on 20min, do it on a trainer OR research to find old local triathlon bike courses. Generally flattish and have good flow.

MattTheHat 

No, apparently it's not the best route. I'll have to reconsider and find something flatter and without traffic issues. That's going to be a challenge in this area.

I have no interest in brag factor. I just want something as accurate as it can be. Strava's estimated FTP might even be adequate, but it's estimating higher than what I think it should be.

MattTheHat 

That's interesting. With enough data, it really *should* be fairly accurate.

MattTheHat 

Thanks for that, I'll go read.

caloso

One thing to do is just take the estimate you have, whether it's from Strava or testing, and then just plug those into a training program. If you consistently struggle to finish interval sets, then just revise down 5-10 watts. Do the intervals based on the revision, and then after a few weeks, tweak again.

jadocs

Based on my experience and something to keep in mind.... if you are not used to going all out for an hour that your 20 minute FTP estimate will be higher (sometimes quite a bit depending on your level of fitness) than what it actually is. A 20 min test for someone who is not used to pushing all out for an hour will probably over estimate your FTP by about 20-40 watts depending on your level of fitness. Do it both ways, a 20 min test and next week try an hour and see what the difference is.

MattTheHat 

Yep, that's another reason I'd prefer the 20 minute test. A couple of months before I got my power meter I did a 60 minute effort at 95% maximum calculated heart rate for some reason. I don't even remember why. It was not pleasant.

I need to just suck it up and find somewhere to do the 60 minute test. Seems like the simplest for the way my brain works.

jadocs

The easiest way to get used to it is to do it often.

I do it every week as part of my structured training. I just finished doing one about an hour ago. It will pay dividends believe me.

Don't worry about what your power is initially. Just go as hard as you can sustain yourself for one hour every week. Throw in some tempo and sprint days. Everything else is gravy and for fun (ie group rides).

Carbonfiberboy 

If I understand you correctly, you're saying at least at first train hard by following your perceptions, i.e. pain, heart rate, power production, those signs of both recovery and overreaching, and let the numbers do what they will. After doing this for a while, the numbers will become meaningful enough to be able to base training off them.

That's what I did when I started using training numbers. It took maybe 3 months for the numbers to settle down to something I could rely on to some extent, and a couple years to see what my best peak might be and how to reach it. That was before I started using power, so maybe it'd be quicker with power, I don't know. Kinda doubt it.

asgelle

That's kind of what I'm saying. Certainly you have to have enough training data for it to settle down and be meaningful, but what I wanted to emphasize was that once the numbers have settled down, they should only be interpreted in terms of the individual's unique capabilities. What might be an excessive training load for one person, might not be even close to the limit for another. To use the example of CTL, one person might be able to maintain a CTL of 85 for quite some time, while that might lead to overtraining for someone else. On the other hand, a third person could possibly handle a CTL of 115. Everyone needs to find those limits for themself.