First of all, the Wahoo app is made to control the Kickr trainer. Yes, it can be controlled by other trainers, but
only by ANT + FE-C. You can use the controls, at least when it comes to resistance modes and manual power input. The incline mode either didn't work at all, or somehow reacted, but still rather useless, I guess it's related to the Kickr trainer specific mode: incline and wind speed and bike type - probably only useful with a Kickr trainer (including Core) so is the LVL (power curve selection) mode.
The only plus point compared to the Elite app is that the app is completely free and you can very easily use the manual mode of the trainer power setting.
As for the ERG mode from the Wahoo counters, eg Bolt, they may not work well with other trainers than Kickr, for which they were optimized. The google Wahoo group has confirmed cases of damage to the Elite trainers, which were controlled by Wahoo counters. Besides, Wahoo behaved differently with Tacx trainers (eg Neo).
In theory, using ANT + FE-C steering from Wahoo Bolt should work properly, as long as Wahoo is sending the correct steering commands.
Hence, the descriptions of the Elite trainer control from the Wahoo computer may be (though not necessarily) a problem of the Wahoo computer itself (eg Bolt).
My experience with the ERG mode controlled from Garmin Edge 1030 (ANT FE-C) or Zwift does not confirm the cases described by the Bolt user. On the contrary, the control is good, the level of power maintenance is also acceptable.
I did a lot of testing with the XR. In short, the test with smoothing "1":
I marked the change 36-> 52 with arrows and vice versa.
Overall, you can see that the Elite developers have sought to limit high power spikes at the cost of slower reaching target power. It can be seen in both repetitions made on 36/17, as well as on table 52. In the case of 36, reaching the set braking power took approx. 6-7 seconds, including the initial lag (i.e. from the interval change sound), in the second repetition it was longer (longer lag ). On 52/17 you can see a problem with stabilizing the low power of 100W due to the relatively high rotational speed of the flywheel, which is a normal phenomenon and you should be aware of this when changing gears at low power. At 52 it appears that the target power was reached faster, but so were the power fluctuations.
Change the alignment settings to 5:
The effect becomes apparent, especially in reducing power hold fluctuations at the expense of a slight shift to stabilization: 9-10 seconds including initial lag.
Value 10:
Most stable values and maintenance of power, at the expense of shifting to full stabilization. First repetition: from the interval change signal to the reduction of the rapid increase in braking: approx. 10 seconds, from 20 seconds after the signal nice and stable power maintenance. The second is similar: the first 10 s. Initial relatively rapid increase, the next 10 s. Until completely stabilized. 3s lag included.
It looks even better for 52/17: stabilization is achieved in approx. 12s (taking into account approx. 3s of lag - i.e. stabilization of approx. 9s), slightly greater fluctuations in braking power.
The descent with power in the above case is 10s. in the previous - about 5 seconds, and in the first - about 3-4 seconds.
Averaging is what it actually is, which is averaging. The trainer counts the power readings, averages them to the given value and makes resistance adjustments based on that.
Overlapping this with the crank power meter reading (without smoothing):
You can see that the resistance of the trainer increases and decreases much more than it results from the power curve transmitted from the trainer meter (because this is averaged over the last 10 seconds).
As you can see the real increase in the resistance of the trainer is much faster than previously estimated according to the averaged curve and amounts to about 3-4 seconds (5-6 with lag). The decline is also 3-4 seconds.
The trainer doesn't adjust the braking power so slowly, it just pulls up quite hard and then adjusts slower. In the case of dial 36, it was pulled up to 175W in 2 seconds, but from the beginning of the repetition it is 15 seconds, i.e. 13 seconds of tightening. Slow, but very smooth, without power surges (maximum 203W).
The spin chart on the 52/17 counter looks similar:
also here you can see a rapid increase in braking power, although the reported power is averaged, hence it grows slower. The actual power recorded by the external meter shows slightly larger power fluctuations than in the previous case, but this is understandable due to the greater speed of rotation of the flywheel.
Here, the 192W pull-up lasts (according to Power2Max) 4-5s, another 6s. it is a tightening to the set power (actually around, because it reports 197W). Maximum jumps of braking power a bit larger, but not exceeding 7W (on the crank 9W).
The side effect will be lower power recorded with the Direto XR meter by the loss of value during the increasing averaging. Not much, but always.
They could introduce the possibility of entering the average power separately, and the possibility of adjusting the braking power according to the average power. So, separately, the power could be transmitted without averaging (or with averaging specified by the user), and the braking power corrected by averaging. So it would be like in the above graphs, but for the external meter (although P2M averages 3s internally anyway).
I did a test of 100W / 200W intervals (gear 36/17) with power equalization 10. The legs feel different than what you can see on the power numbers taken from the trainer, so I verified it with the power measurement in the crank. Power2Max also has internal averaging (probably 2-3s), hence the measurement is already a bit even, but it is interesting to compare the pure P2M measurement with the average 10 measurement in the Direto XR:
Apart from the lag from the start of the LAP signal (approx. 3s), it is clearly visible that the resistance on the crank is changed normally, only the reported is smoothed. Let's look at a single interval (1):
In the P2M measurement, you can clearly see that the resistance increases in 2 seconds. It should be remembered that P2M has averaging, so the target resistance can be obtained even faster. The Direto measurement report is stretched because it is averaged to 10s. hence the longer rise time (approx. 10 seconds) as well as the fall of the reported power.
Comparing the interval with the power compensation parameter "1" and the data recorded by Power2Max:
You can see that the measurements with their mileage practically do not differ (I did not calibrate P2M, hence a slight downward difference, when I calibrate P2M, the measurements are very similar), and if we compare the last two graphs, you can see a slight difference in stabilization in favor of the first (second) plot, where the braking power is stabilized based on a large power averaging, but this is not such a difference as can be seen from the power data itself reported from the Direto XR.
Elite should introduce a choice of: power averaging, which is a parameter of equalization and stabilization of the operation of the motor moving the magnets, but the possibility of transmitting power not averaged or averaged according to the user's wishes. This would make it possible to set such a good stabilization, as well as the lack of distorted power data for analytical and training purposes (as it is now when you assume power equalization 10). You have to ask Elite to add such an option, although looking at the speed of introducing changes to the Elite, I can see it black ...
nice if you'd like show off but from training register is disturbing a bit. 
let's try