Questions about training with power

  • Creator
  • #46570

    I’m looking for some suggestions and guidance about training with pace and power. Living in a city in the northeast USA, I’ve been running less than I’d like to due to COVID-19. (It’s like being in a Pac-Man game… 🙂 ) Becoming a better rower / sculler is something I’d like to do. I’ve focused more on ergometer training lately, and anticipate doing even more over this coming winter.

    Training on a rowing erg gives me both heart rate and power data. I’m trying to figure out how to plan and measure my training. I’m a recreational athlete, trained about 225 hours last year and am on track for a similar or slightly higher number of hours this year. So my primary metric that I’m focused on increasing is hours spent training, especially >85% of those hours below aerobic threshold (AeT).

    I’d be grateful for any suggestions for how to think about my training, focusing more on erging than on running. The reduced impact / load-bearing aspects of rowing suggest that I should be able to row for more maybe 50% hours per week than I would otherwise be able to run (comparing the ~800 hours per year for elite road runners vs. 1,200 hours per year for elite rowers).

    There is lots more data that I can dig into with a power meter. I’ve done a lactate test on the erg, and have wattage numbers at 2mmol/L (AeT) and 4mmol/L (anaerobic threshold, AnT). Training and Racing with a Power Meter by Allen, Coggan, and McGregor has been interesting, although I’m not sure I’ve understood yet too much of the data analysis they present.

    One of the specific questions I have is: how might I look at and think about a workout where the hrTSS that TrainingPeaks calculates is substantially higher than the power-based TSS that the software calculates? For instance, I did a 2x20min workout last weekend (link), and the hrTSS was 41 vs. the power-based TSS of 28. I set the my rowing power and heart rate zones in TrainingPeaks based on the results of the lactate test I mentioned.

    PS: looks like my post got caught in a spam filter – hopefully I didn’t cause duplicate posts.

  • Inactive
    Anonymous on #46613


    I’m surprised that the hrTSS and power TSS are so different. I’ve never used the power TSS but in theory it should be the most accurate. I suggest sticking to one metric and since you have Watts I’d do that.

    It might be interesting to do the HR drift test and compare the AeT it generates to the 2mMol power you got.

    I would do the base work below AeT as normal but I would recommend using a time trial effort for 30+ min to see what your AnT power is. For endurance training you will want to be training at or above that wattage.

    I long for the simplicity that is possible when you have pace or power to control your training. A luxury we don’t often get here.


    Reed on #46652

    Thank you, Scott – I’ll do some experiments and see what I learn. I haven’t done either a heart rate drift test nor a “field” AnT test on the erg yet. It’ll be interesting to see how those data points match up, or don’t, with my 2mmol / 4mmol numbers.

    Reed on #46691

    I did a heart rate drift test, to help assess the validity of the numbers I got via the blood lactate test I did a couple of weeks ago. (That lactate test suggested that my aerobic threshold for rowing is at ~148 watts, 165bpm heart rate.) After warming up for 15 minutes, I did a 60-minute workout targeting 150 watts at a 20 stroke-per-minute rate (cadence). I kept the wattage flat, but my heart rate rose and I also found myself upping the rate to 22spm or 24spm by the end.

    Here’s the TrainingPeaks workout link.

    My observations:

    • TrainingPeaks calculates Pw:Hr to be 7.13%.
    • Pulling the erg data into Excel, I found slightly different numbers. Average power for first half 151.1 watts, second half 150.5 watts. Average heart rate 168.4 and 179.7 respectively. So I calculate a heart rate drift of 6.73%. I’m not sure why TP’s math is different from mine. Maybe some numbers were changed slightly during data import (TrainingPeaks won’t accept the data from the RP3 rower, but does and syncs it to TrainingPeaks…) Or more likely, I’m taking an average of the per-stroke data points, and TrainingPeaks is looking at the timestamps and interpolating the area under the curve.
    • TrainingPeaks calculates a power-based TSS of 57 for that total 1h18m workout, but a 99 hrTSS. I’m not really concerned about the numbers, but am curious why there’s such a difference. I think I’ve configured the TrainingPeaks rowing heart rate zones & rowing power zones appropriately. But, the 4mmol/L heart rate that I recorded the other day was ~180bpm, which was the average for the last half of this workout.

    Seems like my 151 watt target for AeT was a little bit high, although in the right ballpark. As I mentioned earlier, I’m training 4-5 hours a week, so this test was a decently hard workout for me. Not sure if that would affect the interpretation of the test.


    Anonymous on #46722

    Reed; I think you nailed in your last paragraph. This test was 20-25% of your weekly volume.

    Also you might be seeing a muscular endurance factor kicking in during the test that distorts the drift test.


    Garret on #46736

    Hi Reed !

    I was thinking …. if 180bpm is AnT then the test has been distorted because once you got to 180 there was no further increase possible.

    I’ve seen Scott Semple comment, from time to time, on AeT test questions where he observers that if a test is at or close to AnT HR there’s low drift because AnT HR is a ceiling.

    I wonder if in this case hrTSS is actually a better reflection of the actual stress.

    The power based TSS for the hour following your warm up is around 45 and doesn’t vary much between the first and second 30 mins.

    But by HR measurement you were close to 180bpm (AnT) for the final 30 mins which by definition should be a TSS of around 50 for those 30 mins.

    If true then a significant difference between hrTSS and power TSS could be a really handy insight / indicator.

    I’m experimenting with power based work on the turbo trainer and I am looking to see where my power at Hr AeT lies relative to FTP and the traditional % of FTP zones that are in the literature.

    – Garret.

    Reed on #46739

    Hey Garret! Yes, I remember reading Scott’s comment related to that. I’m not sure that I understood it, though.

    There are a couple of other data points that I have:

    • 4mmol/L lactate test indicated an AnT of 179bpm at 213w.
    • I did a 2K test (2,000m row) in ~7:20. Average heart rate 191bpm, max 202 bpm, average power 253w.
    • I have not been rowing much at all lately. And, I haven’t been training very high volumes in general. So movement economy might be a factor, and as Scott J. mentioned cardiac drift from fatigue (from being untrained relative to the demands of the workout).

    It’ll be interesting to re-do this test in a few months. My AeT watts per kilo is around 1.8, AnT ~2.7w/kg. That puts me, based data from Nolte’s book Rowing Faster, at somewhere around a competitive high school rower or maybe a bottom-tier university rower. Increasing hours, >80% below AeT and 5-15% above AnT, remains my guiding set of metrics.

    By the way, in Coggan’s book, there was a line that caught my eye: “Level 3: Tempo. Many a coach has referred to this level as a cyclist’s no-man’s-land. […] If all you have is 3 hours a week to ride, drill it in the upper range of Level 3 and get ini a great workout.” (3rd edition pg. 61) I find the FTP, CPV, AEPF, quadrant analysis, and so on interesting. But am left wondering if it’s like a cyclist with 22% body fat hyper focusing on the weight savings of a titanium derailleur. 🙂

    Garret on #46745

    Hi Reed

    I think Scott Semple’s saying that at an intensity between AeT and AnT one sees HR drift >= 5% but at AnT intensity on doesn’t see significant (> 5%) drift.

    In a recent turbo trainer ride at AnT (based on estimated FTP) for 50 mins my HR drift was 3.5% which I think is what Scott S. is getting at. If I didn’t know better I could interpret that as having been at an intensity below AeT 🙂

    – Garret

    Anonymous on #47472

    Disclaimer: I’m not familiar with rowing training at all…

    It reminds me of cycling, where HRs are normally lower for a given intensity due to less muscle mass being recruited. So cycling-hrTSS is usually less than TSS. Perhaps with rowing, it’s the opposite. If more muscle mass is recruited, then HRs would be higher for a given intensity so that rowing-hrTSS would score higher than TSS.

    Either way, I would work off of power rather than HR. Power is the dog; HR is the tail.

    Other than that, I think all of the usual caveats apply.

    Reed on #47630

    Looks like I ended up creating multiple threads – mea culpa.

    I am curious about what you’re describing, Scott. I’m sure that technique and efficiently play a role as well.

    A few weeks in, I have enough workout data to make a few observations. If anyone else is experimenting with indoor power-based equipment, maybe it’s useful to you. The discrepancy between TrainingPeaks hrTSS and power TSS is much greater at easier paces, and I think it’s because TrainingPeaks is calculating TSS base on my AnT wattage. It seems like it might be ignoring power zones for the purposes of calculating TSS. Whereas for heart rate zones, time in Z1 vs time in Z3 will be counted as 0.6 intensity factor vs. 0.9 intensity factor.

    Garret on #47648

    Hi Reed

    I’ve been comparing Training Peaks hrTSS and TSS for power based cycling on a turbo trainer.

    The data I have so far is for a number of AnT intensity efforts of about 45 minutes duration.
    My Training peaks hrTSS is within 2% of power TSS with the values being in the 75-80 range. HR drift also is less than 2%.

    I don’t have lactate data but I have a good confidence in my threshold power and AnT HR estimates.

    At the moment I suspect that because TSS’s definition is based on power whereas hrTSS is an approximation this may be a source of differences between hrTSS and power TSSat lower effort levels.

    I’ll share results from AeT and lower intensity efforts when I get data for those.

    – Garret

    Reed on #46740

    Looks like the forum’s spam filter doesn’t like links. I’d found a link to Scott Semple’s comment [1] and had a link to the 2k test I mentioned. [2]


    Anonymous on #47665

    AeT/AnT DRIFT:

    I should clarify. AnT HR is not a ceiling. But the highest sustainable effort for X minutes is a duration-dependent ceiling of sorts. Any time trial effort will have a maximum sustainable pace. As I understand it, that’s the basis of the critical power concept.

    So that’s the dog. Heart rate is the tail. Whatever the duration, the highest sustainable output will elicit a certain heart rate. For a 30m time trial, that HR will be darn close to AnT HR.

    Many times in this forum, have tried to do an AeT drift test and have gone too hard, often close to an AnT effort. So the HR that they see has little drift, and they wrongly assume that they are within their aerobic capacity.

    In that case, I often say:

    * AeT HR could drift higher but doesn’t; while
    * AnT HR doesn’t drift higher, because it can’t.

    Now I see how that could be confusing.

    Bottom line: AnT HR is not a ceiling. The highest sustainable effort over X minutes is self-limiting and if the output is stable, so too will the heart rate.

    Anonymous on #47666


    My comment on recruitment was just a guess.

    When I’ve done lactate tests on a bike (as an infrequent and untrained cyclist), my threshold HRs are at least 20 beats lower than during a weight-bearing test.

    This is likely due to less recruitment, and with less recruitment, there’s less central stress, so HR is lower. This is a common phenomenon; it’s not just me. I suspect the lower recruitment is because of both less activation (for everyone) and lower efficiency (for untrained cyclists). I wouldn’t be surprised if triathletes don’t see this effect as much.

    So that made me wonder if rowing will show the opposite. Perhaps with greater recruitment (by using both upper and lower body) the central demand is greater, so HRs will be higher than purely weight-bearing activities. That’s just a guess though.

    XC skiing would perhaps show a similar effect. And perhaps that’s why XC skiers show much higher VO2s than other sports (because their training is quadripedal and more demanding on the heart and lungs).

    Maybe Scott J. has more thoughts on this.

    Anonymous on #47688

    Great discussion guys. I’m not sure i have much to add. I think Reed might have done the AeT test at too high a power and HR because he ended up near his AnT. I agree with Semple that power is really what you want measure as it is an indication of performance whereas HR is a quit imperfect proxy for effort, intensity, speed or performance. Since you have the possibility to use power I’d stick with that.


    Reed on #48306

    I took a few more measurements to try and get a better handle on what easy & hard wattages are for me. No substantially different info, but some helpful added data points.

    • Baseline blood lactate, taken just now during an afternoon at my standing desk: 0.8mmol/L
    • After a good 20-30 minute warmup: 1.5mmol/L
    • After 15 minutes @ 150 watts: 2.8mmol/L
    • After 90 seconds @ 300 watts: 5.9mmol/L

    TrainingPeaks link:

    Looks to me like (as discussed above) I should keep my easy work well under 150 watts. 30-30s are pretty fun on the erg, and absent more info via a AnT time trial aiming for 300+ watts is probably a good target for those.

Viewing 15 replies - 1 through 15 (of 16 total)
  • The forum ‘General Training Discussion’ is closed to new topics and replies.