Help Me Better Understand These SMART Power Intervals

[ulty (Brent)]

I just did the workout SMART - The Beauty of Suffering - 120 which uses dynamic power intervals. I did the workout using the EBC app on Android, connected to the Xert remote web player. The entire workout was done in “auto” mode setting, which means ERG mode for this workout.

It’s the first workout where I really noticed variance in and paid attention to the SMART interval metrics, and saw that they were very small and never met the declared MPA targets. When looking at the charts of the completed workout the dynamic target power shows it could or should have been greater, and didn’t match the power values during the workout.

Here are screenshots of the last dynamic power interval from my workout, which was defined in the workout file as “Intervals under Fatigue - 93% Reserve MPA - 05:00 minutes.” The first screenshot shows instance values from the middle of the interval with “baseline” values before the dynamic power changes kick in, and second screenshot shows the values at the very end of the interval, where the dynamic power increases were at maximum.

During the workout the ERG power for the interval was very consistent with a target power of around 181-182 W, with the actual ERG power fluctuating between 179-181 W, which is a typical range for ERG mode. This power remained consistent for about 90% of the interval, after which in the final 10-15% the target power began increasing until it maxed out at 186 W at the very end. The target MPA for the interval was 605 W, but MPA only got reduced to 614 W.

I thought there would be a more significant change in target power during the workout, especially since the reported MPA never reached the target MPA. You can see in the charts that the dynamic target power line increases both much earlier and much higher than what was actually given during the workout.

So what is this discrepancy and why does it exist?

From the support article Advanced Workout Design using SMART Intervals, this interval’s power should be defined as “Target MPA %Reserve – The interval power is calculated such that MPA is the TP + the number as %(PP-TP) by the end of the interval.” However the terminology of the definition is a bit confusing to me, so I’m having trouble applying it as a formula to calculate power values for points of time within the interval. What does “the number” refer to? And what does “as %(PP-TP)” mean?

I’m also coming off a several week break and being sick, so my fitness signature probably isn’t 100% accurate, which can account for how difficult I perceived the workout to be but doesn’t explain the chart data.

[ulty (Brent)]

I found a couple other references with examples that clarify the MPA %Reserve calculation.
A, B, C.

It is (x% * (PP - TP)).

However this formula still doesn’t make sense of how my workout power was determined.

Again using the “93% Reserve MPA” from the workout file, if I plug in the numbers from the two data points in the screenshots I get the following.

Point 1:
MPA = 644 W
PP = 648 W
TP = 168 W
93% x (648 PP- 168 TP) = 446.4 W
93% x (644 MPA - 168 TP) = 442.7 W

Point 2:
MPA = 615 W
PP = 648 W
TP = 168 W
93% x (648 PP - 168 TP) = 446.4 W
93% x (615 MPA - 168 TP) = 415.7 W

These mid 400’s W numbers are nowhere near either ERG power of the interval or the target power in charts. If TP in this instance somehow refers to target power and not threshold power, the difference between those number is only ~20 W in this scenario, and subtracting 20 from those numbers is still nowhere close.

The formula of PP - TP doesn’t make sense for SMART because the value would never change during a workout. Using MPA makes sense because it would change based upon the effort so far, but there must be something else going on in the calculation because these numbers just don’t match up with either the actual workout of the target power in the charts.

[ManofSteele (Scott Steele)]

@ulty don’t forget to add back your threshold power after calculating the difference between PP & TP.

Most basic way I can explain it: %Reserve is basically an amount between PP & TP. A percentage closer to 0% brings MPA nearly equal to your TP while a percentage near 100% would keep MPA near Peak Power.

These Target MPA intervals can be used as hard intervals - where MPA is pulled down a specified amount by the end of the interval - or as recovery intervals, where MPA is allowed to recover so much before starting the next interval.

Perhaps the easiest way to visualize this is likely to look at the SMART - Elastic Heart workouts series… Rather than using rigid %FTP for the 5x5 min intervals, each 5min interval is based on reached specific %MPA reserves by the end of the interval (so the target will adjust up/down interactively if you’re above or below the target power).

The 3.0 difficulty version targets 85% MPA reserve by the end of each interval:

The 4.0 difficulty version targets 40% reserve - notice how much deeper you need to pull MPA in this version of the workout. You’d need to pull MPA down over halfway to Threshold on each interval… Ouch! This workout would likely be very difficult to complete:

An example of MPA reserve used as a recovery interval is the SMART - Dirty Little Secret. This is a Mixed Mode workout, where basically you hammer each 20s sprint as hard as you can and the recovery interval afterwards will adjust such that your MPA recovers enough for the next hard 20s sprint:

If you really overcook the previous sprint interval, the recovery will be lower to allow you to recover more. If you ease off and don’t push, then the interval will adjust up, since your MPA will already be ready to go for the next sprint. IMO, Mixed Mode workouts is the best way that we can utilize our Target MPA intervals, since they’re based on your actual efforts. If you’re following a workout entirely in ERG mode, the target MPA feature likely isn’t as important since most trainers are quite good at holding you at the target power.

HTH!

[ulty (Brent)]

OK. Now I realize I was thinking about it backwards. I was thinking the formula was used to define the power target of the interval. But instead it’s used to define the rider’s unused available power (MPA).

So if I go back to the formulas and correct by adding back in TP:

Both Points 1 & 2
PP = 648 W
TP = 168 W
93% x (648 PP - 168 TP) + 168 TP = 614.4 MPA

So this makes sense since the final MPA of the interval was 614-615 W, and the interval was targeting to hit that specific MPA.

My remaining questions are:

Where does the data for the target power line in the chart come from, such that was different than the power EBC was actually giving me?

Why was the target MPA for the interval showing as 605 W in EBC instead of 614 W according to the formula?

(It makes sense that these two questions are related - i.e. more power output would be required to reduce MPA an additional 10 W, but why the data difference? Shouldn’t the target power #’s just be the same as ERG power +/- variance from cadence?)

[ridgerider2]

What model trainer?
Looks like power smoothing is enabled.
If so, I suggest you disable that to get a true picture of your power output.

[ulty (Brent)]

Hmm… It is a Kickr Core and I’ll check and try changing the setting for power smoothing.

I don’t think it explains the target power line on the chart though. Starting around halfway through the interval, the target power line steadily increases until it plateaus with around 30-45s left @ 30 W higher than the ERG power. Even with power smoothing I would think a consistent 30 W difference over 30s should have been recorded / reported as such.

I have another SMART-power workout scheduled today with %Reserve MPA intervals. The intervals are only 60s in this one but only a little above TP; I’ll try to watch the #’s to see what happens. I’m doing another SMART - The Beauty of Suffering - 120 a couple days later (with the longer 5m intervals), so can really compare the data between the two duplicate workouts with the different power smoothing settings.

[ulty (Brent)]

@ridgerider2 @ManofSteele

Here’s a snippet of my workout from today with power smoothing disabled on the trainer. This is SMART - Repeated Attacks - 3.5. The chart is zoomed in to the first regular MMP interval + the first set of SMART-power intervals. The workout spec for these is:

First Wind	5 min MMP	1:30 min	92% TP	3:00 min	1x
Attack and Hold!	83% Reserve MPA	1:00 min	93% Reserve MPA	6:00 min	4x

The situation is pretty much the same as before. The target power line in the post-workout analysis graph doesn’t reflect what was presented by EBC during the workout for the SMART-power intervals. The regular non-SMART intervals having matching power & target power. MPA is also different than what was shown by EBC during the workout. During the workout, EBC (via the remote web player) was showing my MPA as never going below about 614 and fully recovering about halfway through each “rest between power” section. Actually I’m pretty sure that “MPA” in this chart is actually showing from a baseline of the Target MPA for each interval from the workout as those numbers seem to match up to what I remember EBC reporting - that doesn’t seem to be the case with the previous workout though.

So still pretty confused by this.

BTW, still finding value in the workouts, but unsure if the full potential is not being met, or if there’s just some kind of reporting discrepancy in the charts.

[ridgerider2]

Before discussing your issue here’s my simple explanation of dynamic SMART intervals without the math.

Type 1 is dynamic DURATION.
Many of the original SMART workouts with dynamic intervals were based on this type.

• If you consistently undershoot the target power the duration of the interval increases.
• If you overshoot the target the interval duration decreases.

Duration is elastic, stretching out or shrinking based on your performance.
That freaked out a number of users because you might start a 2-minute interval that crept upwards in time. If you are struggling to finish and see seconds slowing down that could be concerning. It also meant the 60-minute workout you picked could run longer than the time you allocated for your training session.
Due to user feedback Xert decided to convert most SMART DURATION workouts in the Library to Type 2 dynamic POWER.

• If you consistently undershoot the target power the target increases in watts.
• If you overshoot the target the power target decreases.

Any changes are especially noticeable as you reach the end of the interval.
The goal is to accumulate the targeted strain value by modulating duration or power.
SMART dynamic intervals also work in Slope mode, but you need to monitor the rainbow gauge and adhere to the target if it changes.

As for your issue…
For some reason your trainer isn’t responding to target changes under ERG control.
Here’s my last two Beauty of Suffering workouts on file.
Halcyon (March 2021) –

Hammer (March 2025) –

For kicks I rode Beauty today split into two sessions.
The first half I used iOS EBC. Here’s the first interval segment –

Second half of workout was on Android EBC. Here’s the last long interval –

I deliberately switched to Slope mode because my back started bothering me and the only relief is to sit up straight (no hands) and ride easy for a bit. Then I switched back to AUTO.
Notice target watts rising into orange level as I completed what should have been a single 5-minute interval at 90% Reserve MPA.

I think you should file a support ticket to further investigate your issue.
In the meantime, you could ride any dynamic intervals in Slope mode to match targets.
I routinely do that nowadays for some portion of most HIT workouts.

[ulty (Brent)]

@ridgerider2

For some reason your trainer isn’t responding to target changes under ERG control.

I really don’t think this is the case. Regular interval power gets set correctly, linear ramps like warmup and cooldown work fine, and curved ramps both going up and down work as well.

There also is power change in the dynamic power intervals, it’s just rather small. I think the biggest difference was maybe 5 Watts.

This one goes from ~205 to ~210:

@ridgerider2

I think you should file a support ticket to further investigate your issue.

I know Scott, et al, are looking into a few things offline and I’m sure will report back with some definitive knowledge after some time to dig in further.

[ridgerider2]

Your screenshot shows a target of 246 while your trainer generated 210 at that moment.
That’s ~15% lower than target for the duration of the interval.
On your Repeated Attacks chart above the first interval shown is compliant while all the following intervals (green and orange) are below target.
My ERG controlled power is within a couple watts of targets when I’m in AUTO mode.
Here’s one of my 5-minute intervals today under AUTO control –

There will always be discrepancies on short spikey reds due to trainer response during transitions, but on long steady intervals your trainer should be push you close to target.

[ulty (Brent)]

The target power on the first interval matches because it is a static power interval, defined as
5 min MMP for 1:30 min followed by 92% TP 3:00 min. Then the rest of the intervals are dynamic SMART power intervals, which is where the target power line on chart diverges.

The target power #’s in the chart only diverge in the post-workout analysis, but during the actual workout itself the target power #’s in EBC matched the ERG power.

So for that example:

EBC base target power: 206 W
EBC dynamic target power: 210 W
Power output: 210 W
Post-workout analysis chart target power: 246 W

For reference here’s the chart from the prescribed workout followed by the entire post workout chart:

[ridgerider2]

So it appears your target power chart is inflated post workout.
Mine are spot on with discrepancies limited to an occasional short blip at end of interval. Or during my time spent in Slope mode where I deliberately ignore targets.
Does same thing happen if you ride a SMART workout in Slope mode during the dynamic power intervals?

[ulty (Brent)]

I can’t / haven’t successfully ridden anything in slope mode because I have a single cog on my trainer, although I do have an idea to try and get it to work with virtual shifting. Haven’t tried it yet though. Beauty of Suffering might be a good one to try it on because the intervals are fairly long so there is plenty of time to adjust slope. Once I get motivated to jump on the trainer, my mood is usually just focused on the workout and not experimenting, but I’ll get around to it some time.

If I end up trying it, I’ll update.