Can I get a decent workout on an eBike?

I've been following a conversation about eBikes occurring over at https://theconversation.edu.au/electric-bikes-at-250-watts-the-view-has-opened-up-nicely-10465 and seeing a pretty interesting stream of comments rolling in, with the full spectrum of inspirational stories through to trenchant trolling and cynicism. I was curious to see the cynics observing that there was no way an ebike would provide an observable benefit to fitness, nor make much of a demand upon a rider for effort. Having a background in exercise science meant that it was time to gather some data and bring some evidence into The Conversation. I know just how hard I work riding off-road, so thought it was time to back this up with a little rigour.

I've done three tests now under reproduced conditions - standardising clothing (smart business clothes, flat shoes), weight of extras (bike lock, helmet), terrain (5km local off-road loop with a steady sustained climb, noting that 5-10km is an ideal car-replacement ride distance) and varying only the bike with the following changes:
Test 1: Unpowered mountain bike (PACE RC200 - my old race bike converted to commuter duties) 
Test 2: Stealth Fighter @200w limit, with a 40km/h limited hi-torque motor
Test 3:  Stealth Fighter @3kw, with the same motor

For each test, I used Strava (http://www.strava.com) to track my HeartRate, speed, elevation and overall ride length, with a Wahoo Fitness BlueHR heart monitor hooked up (yes, we sell them). For each test, Strava generates a nice plot of elevation, speed and heartrate, which I've included in the result sets. Across all results, I've highlighted the highest values in red.

Note: I'm 44, so have a theoretical maximum heart-rate of 176bpm (using the least objectionable formula of HRmax = 205.8 − (0.685 × age) listed at http://en.wikipedia.org/wiki/Heart_rate). For an indication of the intensity of riding effort used based on average and maximum HeartRates, I've included the following guide:

---

Exercise work zones (Fox and Haskell formula between 20 and 70-year-old): red zone (VO2Max), anaerobic, aerobic, weight control and warming up. From http://en.wikipedia.org/wiki/File:Exercise_zones.png

Standard MTB 5km Loop

Strava Plot for Test 1, based on unpowered mtb on 5km test circuit.

Results
Heart Rate: Avg=167 Max=182

AvgHR % of (theoretical)Max:

Time: 12m:53s

Max Speed: 55.8km/h
Avg Speed: 23.1km/h

---

Strava Plot for Test 1, based on unpowered mtb on 5km test circuit.

Stealth Fighter @200w 5km loop

Strava Plot for Test 2, based on 200w limited Stealth Fighter eBike on 5km test circuit.

Results
Heart Rate: Avg=158 Max=171
AvgHR % of (theoretical)MaxHR: 90%

Time: 11m:40s 
(1m13s quicker over 5km than Test1)

Max Speed: 47.9km/h
Avg Speed: 25.2km/h

---

StealthFighter @3kW 5km loop

Strava Plot for Test 3, based on Stealth Fighter eBike running at 3kw on 5km test circuit.

Results
Heart Rate: Avg=157 Max=172
AvgHR % of (theoretical)MaxHR: 89%

Time: 8m:24s
4m29s quicker over 5km than Test1
3m16s quicker over 5km than Test2

Max Speed: 49.8km/h
Avg Speed: 36.2km/h

---

Conclusions
Yes, I can get a great workout on an eBike - in Tests 2 and 3, I was within 1bpm average sustained at a fairly level output throughout the ride, working at 85%MaxHR or higher, but not as high as I have to when riding unpowered (see Test 1). 

Test 2 and 3 showed negligible difference in peak speeds on a downhill section of this 5km loop, but a marked difference in average speed. This is a clear demonstration that the average speed is most impacted by the use of power when climbing, which is the sort of terrain where the benefits of an eBike are most noticable and where maximum assistance is provided. To see the magnitude of difference, look at average speeds in the 2-3.5km stage of the ride. 

Interestingly, I was fastest on the test loop downhill on my old MTB. 

Also interesting was that I was only marginally faster in terms of average speed on a 200w limited eBike over 5km than I would have been if I'd stayed unpowered on my old bike. However, I'd have taken 4min20s longer to arrive at my destination than if I'd been allowed to use a speed limited eBike of higher output. By the time you extrapolate that out to a longer trip of 5-15km, clearly the gap in elapsed commute time would start to stretch out further. 

Also consider that these tests were done unladen - in commuting practice, I usually need to also be carrying a backpack with food, laptop - spare shoes/clothing if I'm unpowered, as I'd ride in cleats - to a range of different destinations. As a consultant, my workplaces are a little more diverse and transient than a standard 9-5 role. The role of power-assisted riding when dealing with a load, especially when climbing, should be considered in that.