Can I improve this ERG control or should I temper my expectations?

I’ve been using a fluid trainer forever and have only recently picked up a smart trainer - I’ve had no real concerns with it so far but today was the first workout with since it arriving that included ramping out of the saddle efforts (basin), and the performance was a bit… meh.
I tried a few things - slowing my cadence a la climbing, speeding up my cadence a la attacking, keeping the cadence the same as when seated, smoothing the pedal stroke etc., but it was as if the response time was lagging so far behind that I could never sit on target during the ramp. This isn’t really a problem when the ramp ends with a nice steady state interval because it all works itself out, but the sets in basin end with a short and steep ramp up to 150%, and I wasn’t getting anywhere near that figure because my trainer just wasn’t giving me the resistance.

Is that normal? Do I need to temper my expectations? Any tips as to how to get better performance out of the trainer? Might I be better runnin an offset rather than Powermatch?

I’m running a Saris H3, power matching to a Left only 4iiii Ultegra PM. I would experiment with running with powermatch disabled, but there’s a pretty big difference between the trainer and the PM.

Today’s Basin workout is here:

You will have to make your account publically viewable for us to see the link you posted. I don’t have experience with powermatch but hopefully if they can see the workout others can help you


Do you use the desktop app for TR or the mobile app? As far as I know, powermatch has been improved in the desktop app (can’t test it because I can’t run it).

1 Like

Check out this, so you really understand what ERG is meant to do (and not to do):

My most important tip for all new ERG users,

    • Cadence is King! Meaning, set and hold your desired cadence for any given effort. Keep it as steady as is practical.
    • Do NOT watch power or attempt to “fix” a high or low power value. You will only make things worse.
    • It is ok to adjust cadence up or down, but keep in mind that every change you add makes the entire system work to “solve the power problem” by adjusting the trainer in response to cadence changes. So, make your changes as needed, but be stable every other moment when you can.

Also, what platform are you using for TR? The latest Win/Mac app has an improved PowerMatch for ERG. It will come to the Apple/Android mobile whenever that update is release. The new W/M version is supposed to be much improved.


Will do, as soon as I’ve eaten dinner!

1 Like

Thanks @mwglow15 , @splash , @mcneese.chad

I’m on the Android app. I could test it on PC, but I’ll need to move my desktop to the pain cave. I might end up doing that, actually.

I know how to ride ERG (or at least I think I do), as I’ve read/watched plenty of “How To” articles on it - I haven’t had a problem with the more steady state efforts because I keep my cadence pretty tight and gradually adjust it when needed, but this workout requires riding out of saddle at (I think) a lower than normal cadence. I’ll watch that video shortly @mcneese.chad to see if theres anything I’m doing wrong - as soon as the wife’s TV program is over!

I’m in 36/24 or 36/21, cadence of 93rpm on seated efforts, and I tried that cadence and above/below it on the standing efforts with the initial cadence change starting a few seconds before the ramping interval.

1 Like

The ramp test is the wrong workout to test ERG in. Try something like Antelope first.

Also I find the iOS App much better than the desktop for it.

thanks @alen but this wasn’t the ramp test - it was a workout with a lot of ramps in it. I’ve ridden quite a few workouts in ERG mode since getting the trainer, but this one (Basin) is the first one with with lots of power ramps in it.

Now that I can access your ride data:

You have coupled issues here:

  1. First, the workout is trying to increase the trainer resistance to match the increase to follow the ramp. This is a “gentle” ramp and assuming a constant cadence, the H3 could match that pretty easily. Saying that, it will also lag just a bit behind on ramps like this and be a bit off until you stabilize on the next flat step. The flywheel size and resistance unit lag lead to that result.

  2. Second, per the ride instructions, you are standing, which is great in essence. But with the standing, you are slowing cadence. Again, on it’s own would be relatively fine. Assuming a flat step, the trainer could follow the cadence decay and give close to the target. But it would suffer a similar lag as mentioned in #1 above.

  3. The point? You are stacking two lagging inputs (ramp up effort + slowing cadence), both of which require the resistance unit to add resistance. The trainer is just not able to take both of those alterations in a dynamic way to keep “on target”.

  4. Add in the typical delay we see in Power Match and you get some less than perfect results.

Personal gripe, but I dislike the “slow your cadence” approach to standing that is so often recommended, and is done by TR. When I started on ERG, I tried it and really disliked the lag to drop from 90 to 60 rpm for standing and then reverse the trend to get back to seated position and cadence. Notably, I formed my initial impression from a CycleOps PowerBeam Pro trainer, that is REALLY slow to adjust. But even when I apply the slow change approach on my better trainers, it is still not something I like.

As such, I just started to employ the same tactic that we adopt when riding outside. Shift to harder gears for standing efforts. I did this for years and found it much better for performing the workouts and keeping ERG closer to target.

Here is the article I wrote on the proces:

Keep in mind that you need to use proper shifting technique and a slight over-power, under-power when perform a shift under a decent load. When you do that, it works quite well.

Check out my Basin ride where I use that shifting techique. Notably, I am just using Kickr power (with ERG mode power smoothing turned OFF), so I don’t get the additional delays from Power Match. But it shows close follow of the trainer to the workout when shifting and holding a stable cadence vs the slow drag up or down as they tell you to do.


Personally I think that if the ramp is above a particular steepness a lot of trainers just won’t keep up and it will always look like you’re chasing it. TBH I don’t believe there’s a lot you could do about it. Seeing your session curves will help.

I also ride a Saris H3, however I do not use powermatch.

Give some similar workout a go without using powermatch and see if your happy with the performance of the H3.

I had a look at another one of your rides ‘Sierra +1’ assuming this was on the H3, your power is certainly really up and down.


Also, do you have any power smoothing? I can tell but a 3-5s smoothing is helpful

For clarity, power smoothing in the TR app only affects the digital / numeric display for the live power data. It does not affect the line graph during or after the workout.

Assuming the rider is not watching and reacting to power data, there is no likely impact on the results here from any power smoothing setting.


Yes and yes.

Erg mode for all its usefulness is not capable of quickly replicating the large and sudden cadence change one would normally use when moving from seated to standing. There are workarounds and techniques but in that situation (and a few others) erg mode is just not going to replicate “real” cycling and will take a bit of time to settle in at the new cadence/power combo. Basically, anything where normally you shift 3 or 4 gears fast and/or make a big cadence change is going to catch up erg mode a bit.

This is no reason to ditch erg mode but you do need to learn to play the erg mode game and some situations are just always going to be better in resistance mode.


Would single sided PM also introduce lag?

1 Like

Sure, since it’s possible that you will be increasing or decreasing power, and it only catches it half the time, it’s possible. Considering that power is simple doubled, it might matter. But I don’t know that it would be noticed on it’s own. Adding all the factors together just complicates the process overall compared to a single data point like my trainer only example.

Completely agree with all of your points - l just thought that a big expensive trainer like the H3 would be able to handle that sort of demand. The demanded rate of change on these ramps is only about 1W/s, and I believe that the H3 is supposed to be a bit of a beast in ERG mode (that’s why I went for it over the Kickr, actually).

Regarding changing gear for standing/sitting: thats actually how I did my first few “trial” workouts when I set the trainer up, and like you I thought it was much better. But then EVERY article on ERG mode tells you that you must never change gear as it’ll make your trainer work harder and result in sloppier control, so I stopped doing it and now I’m here! I think the first step for me here is going to be going to to the big ring for out of saddle efforts, then dropping back to the little for seated. My only concern with that is if the ramp down demand is faster than the flywheel deceleration is couod get a bit messy, but then I’d rather it be messy on the way down than way up.
Thanks for the article link!

@Gary_Gumnut I use powermatch because the difference between the H3 power reading and the power meter is pretty big - big enough to cause me to bail out of a couple of tough workouts that I otherwise would have finished!
The power you’re seeing on Sierra +1 is from my power meter so it will be much jumper than your H3 power. Thats because the H3 power is mechanically damped by the flywheel and the power meter data isn’t - the H3 data is smoother, but the actual power smoothness will be pretty similar. The power recorded on that workout is actually pretty good!

  • Hence the reason I wrote my article on shifting and my personal objective to retire that old methodology (myth) to the dust bin.
  • Heading into another direction, there is discussion about ERG gearing related to flywheel speed. It’s even messier, but worth consideration as the “pick a gear and break the shifter” ideology is outdated at best. But that is a discussion for a different thread. :stuck_out_tongue:
  • For “sprint” efforts at speed, yes, the H3 will leave you spinning a bit easier for longer than the prescribed ERG setup in many cases. It’s really not a “problem” in the grand scheme of things. That variance is in the least important part of the workout and will have little or no impact on the training effect you take away from the workout as a whole.
1 Like

This is one of those workouts where I would ask myself…could I do better than this if I was outside? Many will disagree with that logic feeling that they spent the money on a machine to make them more machine like, but I see the trainer as a tool, not an end solution.

Don’t get me wrong. I’m not nuts enough to do these workouts on manual mode or worse yet, rollers, but I’m willing to say…yeah…this is about what I’d do if I was outside. So, if outside only training was good enough for all those old guys that fly past me every weekend, I can live with a realistic simulation of outside efforts while inside.


Yup, at some point “close is close enough” and sweating single watt variances is counterproductive. This may not be one of those cases yet, and I think there is room to improve without crazy effort.

But I do think a step back to recognize the overall goal of stress on the body is the goal, not making pretty graphs :stuck_out_tongue: I say this as a reformed over-analyzer that got to a point of realizing the mental worry about those variances (or the possibility of hitting the “perfect” workout) or fixing the “problem” were not solutions in and of themselves. We get some great stress even in these “messy” workouts and should learn to take them for what they are.