tetonrider

Over the years I've fielded numerous questions about power on the mountain bike vs power on the road.

Some of these come from athletes who train with power on the road and are inquiring about whether they should have a power meter on their MTB (or any bike used off-road) so that their coach (me) can have a more complete picture of training; others come from people who want to buy a power meter but need to know how it it might help their training. Disclosure: I am personally a big fan of power meters for training and I sell/install/service power meters. Plus, I am just a geek who likes data.

I happen to ride with power meters on all my bikes, and I figured that a thread on this might be useful.

Power off-road is difficult to use to guide structured training and is certainly not for everyone, but it can give more insight into someone's overall training load, as well as allow an athlete or coach to better understand the actual demands of their off-road events so they can structure more effective training around it.

I have found hrTSS to be a sub-par proxy for actual TSS as measured by a direct-force power meter, though the utility of hrTSS varies by the ride (can be region-dependent). I've done these comparisons for MANY rides -- using actual power from a ride and comparing that file (with power stripped) run through the algorithms that produce hrTSS.

I was on a fun ride the other night and decided to go reasonably hard up a 5-ish minute climb. Many of us here know about how hard a 5-minute interval is and what power is repeatable vs unsustainable.

In the chart below, I drew a dotted line @ 300w. AP was 314w (NP was 311). If you ask me to ride 300w/5' on the road, this is easily repeatable, and I can hold power + or - 10w. However, on a MTB, I'm almost NEVER in that 290-310 band. Instead, I'm as often at 450+ as I am at <150.

This climb was on what I'd call fairly smooth dirt. I was riding a single-speed on it, so that muddies the water a bit (sometimes I have to push an unnaturally low cadence and other moments I am spun out). There are a couple spots where you have to pay attention to avoid a pedal strike, and while there are some switchbacks one can pedal through all of them. The chart would look even more stochastic if I was pedaling through rockier terrain.

[Power = yellow; cadence = green; light blue lines @ 290 & 310w w/ dotted line @ 300.]
Screen Shot 2017-09-30 at 10.56.54 PM.jpg

Thought this might spark some discussion, or fade onto page 2...

globecanvas

I was thinking about this yesterday while getting my ass kicked on a MTB ride with an ex-pro.

I used to have a PM on my MTB and I do have one on my cross bike. Like anyone who does this I was surprised at first that off-road rides almost always feel harder, sometimes a lot harder, than the power numbers. For a while I'd calculate TSS from HR for these rides but that's not ideal either because there aren't enough zones, so it doesn't really capture all those 30 second absolute maximum torque type efforts.

So now I just make up a number based on how hard it felt. Like yesterday I gave myself a 2 hour NP about 20 watts lower than all-time max.

The basic question is whether the training value is proportional to how hard it feels. To me this is more a question of specificity than physiology.

rubiksoval

Four or five of the hardest crits I did this year are all just like that; AP and NP in z2 range, but barely able to turn the pedals over at the end. It's the neuromuscular demands that simply aren't accounted for in those metrics. Quadrant analysis can help a bit, and maybe some "matches burned" charting, but it's really interesting how trashed your legs can be when the typically-used power numbers are showing like that.

Takes a bit of readjusting to figure out what you need to be able to do and how you need to train in order to do that, because the typical "aerobic fitness" demands aren't always adequate.

Racer Ex 

I think the difference between road and mountain bike "power" is in the applied torque. You are riding over uneven terrain with obstacles that require an application of torque to get over and through.

If you consider that those microbursts are unlikely to be adequately recorded on a device that is recording at (best setting) one-second intervals, then you are going to miss the actuality of the effort.

It would be interesting to set up a science / track SRM (8 strain gauges) at a half or quarter second recording fed into appropriate software and see how that would look.

tetonrider

IME experience, I haven't found sub-second recording to be necessary; 1s intervals are enough to see what is going on with force over time. Maybe for a track start it is more relevant. The data would be overwhelming in a longer ride.

I would argue that accurate numbers are, and we know some meters are particularly poor when cadence is changing frequently, as it does in MTB.

As @rubiksoval said, quandrant analysis is a helpful tool here.

With these files it is necessary to look at how the power was generated. Power on the MTB is much better than HR (/hrTSS). I have found that hrTSS often substantially overstates TSS with rolling terrain as HR remains high while no work is being performed.

The point of using power on the MTB is to quantify the actual work. Much like lifting weights in a gym when one is not used to it -- the actual weights and reps measure the work done; accounting for something like DOMS (because one hasn't done squats in a year) lacks objectivity. "It feels harder" doesn't make something a better workout--it may just mean one is less prepared and needs adaptation.

We generally use power to account for day-to-day variations in "what feels like 300w". That changes based on whether we slept poorly, are dehydrated, have super-high motivation, etc., and yet the training benefit of 300w is unchanged.

There may be value in capturing other subjective data (like DOMS or RPE), but I think it should be separate/supplemental to actual power data.

Racer Ex 

If you're using the power meter to quantify workload, I don't think it's apples to apples on a road bike versus a mountain bike or a cross bike for sure.

Almost without exception the same AP produced on a road bike will result in a higher heart rate on a mountain bike or cross bike for that same AP. Of course on a cross bike you're getting off and running which is not reflected on the power meter. On a mountain bike you could make the case that the higher heart rate is the result of using more upper-body to control the bike. The rougher the trail or course the higher the heart rate the tends to be produced for the same average wattage.

One second can tend to be a fairly long time on a mountain bike on a rough course. Anybody who's landed in a bush can tell you that this is the case.

My hypothesis is that the faster recording rate might better reflect the actual effort needed to ride over a rougher trail. Even then the meter is not going to reflect some of that torque load that goes into the muscles, upper body and core fatigue Etc.

Because as cyclists we use power meters as a training tool and not as a purely scientific measuring device, if you're training across multiple disciplines then it's probably wise to add additional training stress for the same AP on a mountain or cross bike. At least that's been my personal experience and my experience with folks who do multiple disciplines with multiple meters.

This thread is actually a good reminder that I need to do get a couple of my cross racers to give me better course descriptions and PE on the races.

tetonrider

If all else is equal, when one's ability to produce power on a road bike does not result in the same reactions as the same power made on a MTB, then I take that as a cue to dig deeper into the athlete's preparation. (I'm specifically NOT saying "AP" here because, of course, AP from 2 disciplines is not going to be achieved the same way.)

I find it valuable to look into torque, not just power IMO.

When second-by-second review of a file and/or quadrant analysis reveals that the MTB ride required high torque to achieve a given power and the same torque is applied out on the road, I expect to see the same reaction in the rider.

(I'm saying "all else equal" to eliminate rough terrain... at least for a moment.)

Also, if lack of 'upper body strength' is causing a rider to fatigue early, I believe the fatigue will show as declining ability to hit high numbers during a ride or inability to sustain power. Either way, the result is less power to the pedals = less stress = fewer kJ expended. That rider with the weak core loses power on a smooth road, too. We may see it in a maximal sprint or when they have to climb a 15% grade.

Some people are inefficient while time trialing, but I don't adjust for their inability to hold position.

IMO, if there is such an issue with a rider, I think this should be captured separately. TSS was never intended to mix cycling with the effect of a run, a weight session, or to add incremental value to a MTB ride because 'it feels harder.' (I think the concept of TSS gets a bit silly -- again, in my opinion -- because all TSS is not created equally. I can ride achieve 100TSS from an endurance ride or a sprint session, and they have VERY different effects on the body.)

Something like a weak core may make it hard for a rider to produce power under high-torque situations (or to have trouble sustaining power); l this will be reflected in the data we already have for the ride, but we must look beyond TSS.

One can use HR decoupling to get some insight here, too.

So far, I haven't heard anything that would convince me that direct-force power measurement + existing analytics are insufficient to assess the load on the rider, but I'm still listening.

I've proposed (and use) "life stress" as a metric, but my purpose here is measuring readiness to train. A weight session, a stressful day at work, a crash on the MTB might all affect recovery and readiness to train (or add to 'life stress'), but they do not add to TSS. At least that is the hypothesis I've been testing with myself and others.

Your mileage may and probably does vary.

Good discussion.

tommyrod74

As a fellow coach who races road and XC, and coaches plenty of both, and has power meters on all bikes...

I really only look at NP from XC rides. There is SO much "noise" in the power data from off-road rides and races that I find it's really hard to look at anything else.

I only really use it descriptively vs. prescriptively (on that bike), unless doing intervals on pavement.

HR/RPE are so far off of the AP numbers that they usually coincide with on the road. I've come to the conclusion that there's so much else going on that's affecting HR besides pedaling (adrenaline, upper body/lower body use for handling the bike, etc) that goes into this that a power meter can't begin to quantify.

The closer the course is to "road riding" (AKA consistent pedaling, minimal tech features, rolling terrain vs. crazy steep up/down) the more useful the power numbers. On super-tech terrain all that goes right out the window.

All that said - in a race, usually I hit a NP very similar to what I'd expect for a tough crit/RR of the same duration.

tetonrider

One thing that is interesting is I often find NP to be pretty close to AP in many rides. This is a signal to do a deeper dive to me.

Absolutely! Descriptive is still valuable, for sure.

Sometimes you get a fire road or other consistent area, but it's the exception rather than the rule for training. I think recording the effort during a race is highly valuable for developing structured training to prepare for that event (or similar events) in the future.

Ttoc6

Id like to point out you are also a pretty skilled mountain biker. Less skilled riders are going to be putting their rear wheel into rocks and having to spin up more often. Hitting roots that cause their rear wheel to spin free for a revolution.. Etc

tetonrider

Skill aside, my observations are not based only on my data, though not as many riders have power on the MTB, let alone accurate power.

My comment was more to say that despite having rides that are highly variable, mathematically AP and NP happen to work out to be nearly the same on numerous rides.

tommyrod74

Yeah - I like to have the data for TSS purposes, of course, but only use it during a race for a long fire road or other sustained climb (Pisgah Stage Race was a good example). NP and AP are usually similar until I start hitting race pace, then they diverge hugely, as I'm focused on riding very efficiently (nailing climbs to build gaps, preserving momentum everywhere else) and not focused on pedaling everywhere. On more sedate they are often similar. Still, even in race situations, the two are closer than, say, a really hard crit (without a breakaway effort).

I totally agree about the usefulness of the data to build workouts.

It's very interesting how, no matter the VI (variability index), the NP I can realistically achieve in any sort of race is very consistent.

To add: Interestingly, though I'm faster on the full suspension bike than the hardtail (and have power on both), AP and NP are also higher on the FS bike. I first thought that it was due to just trying harder - as that bike's also considerably more fun - but I soon realized from file analysis that it's because the FS allows me to pedal more and coast less through rough terrain...

globecanvas

This is a bit off topic, but I just put a much bigger front tire on my FS MTB (2.6 Minion DHF) and had to go for a headlamp ride last night to try it out and my goodness what a difference a big sticky front tire makes. I did a 45' single track loop that I've done probably 100 times that is technical enough that even at JRA speeds it takes a tempo type of NP effort just to stay on the bike and I would say the tire saved me 20 watts.

topflightpro

Slightly off topic, I just bought an SRM for my track bike. I plan to install this weekend. Track power is kind of a weird thing too. One can very easily track and measure progress and performance by cadence, speed and time, as there are not really many other variables on the track, but having power gives me one more metric to analyze and will hopefully help me identify a few areas to improve.

tetonrider

I'd like to be sure I'm reading this right: A DHF made it *easier* for you? Sticky ...are you running the 3C version?

The DHF is a great tire, but it's not known to be a fast tire. It's really grippy, which makes it great for enduro/DH.

What were you using before?

It's sure fun to ride high volume, aggressive tires, but generally not fast for "trail" riding.

tetonrider

*slightly* off-topic?

And, unlike other disciplines, since you are always pedaling on the track you could compute power based on gear, speed and estimates of CdA & Crr with a high degree of accuracy, esp for the individual events.

Still, power is useful and I would think it might be particularly interesting in your group/sprint events.

Enjoy, and keep us posted with how your training on the track w/ power goes.

globecanvas

Way easier! This is technical single track so rolling fast is not really a factor. Easier = staying on the bike with less effort.