Eric F 

The server is managing to generate a virtual image that includes the relative position of other users in real time, and pedal feel that is unique for each user, at that moment in time. I would think that adding another layer for drafting effect, felt in the pedals, would be possible. I know very little nothing about this kind of programming, so maybe it's way more complex than I'm imagining.

Eric F 

Hmmm...interesting that I'm having a different experience. I wonder if it's effected by my higher weight, which means a higher power output for the same w/kg. Generally, on flat terrain, I find that I can ride at the same w/kg, and occasionally .1-.2 w/kg less than the bot, and stay with the pack.

PeteHski

AFAIK pedal feel (resistance) is only a function of your trainer responding to the variable slope value Zwift is broadcasting to it. Drafting affects your avatar speed, but not your actual trainer resistance.

For Zwift to vary your trainer resistance with drafting would require another trainer control variable. I don’t know if this is possible with the current Smart trainer control protocol. I doubt it otherwise Zwift and/or other apps would probably have done this by now. I’m not aware of any other apps that can do this. They can simulate drafting on the screen and you can reduce power to maintain speed in the draft, but your trainer resistance is driven only by the slope variable.

If I’ve missed something then it would be very interesting to know.

zacster

The thing with drafting IRL is that effect of the wind disappearing and it would be hard to translate that into a resistance change, which the trainer could support. I was thinking a while back about wind in Zwift, before this thread came up, and I realized how hard it would be to simulate that. The power curve of wind is different than grade and of course it is invisible. Even on my best days when I turn into a wind it takes everything out of me. I did a ride outdoors the other day and I was feeling great, then I turned around. Not only was the wind in my face, it immediately felt 10 degrees colder.

It is always a wind free day in Watopia.

Eric F 

Again, I’m pretty ignorant on the programming - and hardware interface requirements - that would be required. I guess it would be pretty complex, affected not only by the distance to the rider ahead, but also the number of riders, and the speed, as an additional layer on top of slope…but it still seems possible.

PeteHski

It is certainly possible to calculate a parameter for “draft power reduction” and that’s what Zwift and the other simulation apps do to control avatar speed. For example, when drafting in Rouvy it displays your following distance and power saving in real time. What is not clear to me is whether or not this parameter can be broadcast to your trainer to modify resistance. I guess that depends on how flexible the trainer control protocol is. That part I don’t know.

PeteHski

It is actually relatively easy to add wind into the virtual power/speed model and some sims do that already, but I’m not sure if the Smart trainer control protocol can use that parameter to actively control resistance. I get the impression the App just sends the slope value to the trainer and relies on the trainer to provide resistance based on its own power model for the given slope.

This is how my Wahoo Kickr Bike appears to work anyway. If I run Zwift in parallel to the Wahoo App, with Zwift controlling the trainer, I can see the slope value changing in the Wahoo App and resistance changes in respect to the Wahoo calculated power/speed model rather than the Zwift avatar virtual model. The only parameter common to both is the slope value.

unterhausen

I think I have seen discussions of drafting on zwift, and the conclusion was they do it the way they do it for ease of implementation and also because they didn't want to give the full effect of irl drafting. It would be cool to be in one of those large groups that looks like a surging hoard of overlapping rodents and be flying along with no effort, but in my case, that would be counterproductive. Since I'm only there to train. I don't think it's impossible to do a better job, and maybe one of their competitors will someday.

It occurs to me that IRL, you can feel that the wind is being blocked. I know the required power output is significantly less while drafting, but they are never going to be able to block the wind. Although this is another great product idea, they could make a fan that simulates the effect of riding behind another person.

RChung

If only there were some way to convert aerodynamic drag into an equivalent slope drag.

himespau 

indieVelo adds cornering brake feel and wind feel to the resistance you can feel. You can tune how much of it you get, just like hill gradient (trainer difficulty in Zwift). Not sure if wind feel includes draft or just wind. I turned it all the way off because it was creating some times when there was too little resistance and I was running out of gears when I wanted to attack. Haven't turned it back on since they got virtual gearing. It does appear to be possible to make "draft resistance" something you can feel.

himespau 

Not sure how it's changed with the most recent update (complications due to surgery have kept me off the bike for a while), but it used to be that the range you could be in front of (or behind) the pace partner depended on the size of the group. Also on the world. You could be further infront or behind the pace partner in Watopia than in Makuri (or those few times where they put pace partners in other worlds for special events like TdF).

PeteHski

I thought about fudging the slope value to combine drafting aero, but then it would mess up trainers like mine with a physical climb function driven by the slope value.

What I want to know is whether or not trainer control can support multiple resistance parameters. So you could potentially have slope, aero resistance and rolling resistance as independent variables.

PeteHski

This is interesting. I think I will go back and try this out for myself. It will be obvious if they are messing with the slope resistance to achieve this and if not then it suggests another trainer control parameter is possible.

RChung

Slope and rolling resistance scale exaclly alike. Aero doesn't scale like slope and RR, but you can translate aero drag into a slope (or rolling resistance) equivalent.

zacster

I'm sure the protocol could handle it, it would be the trainers being able to understand it. But even the trainers are software controlled, I get updates from time to time for my Kickr Core. Wahoo could add it, working with Zwift, and then open the code up so other manufacturers could implement it. Or they could all do it together. It could be one of those things that only works on Bluetooth, or ANT+, but most systems handle both. In any case it would take some programming on all sides and I have no idea how hard or easy that would be. If it were easy it probably would already be done.

PeteHski

I understand this, but if the course slope value was modified to include aero then my Kickr climb function would be messed up and no longer reflect the actual course slope profile.

Aero and rolling resistance appear to be controlled entirely on the trainer side using only the slope value broadcast by Zwift as the input. But maybe Zwift or other Apps are able to broadcast other resistance control parameters. I had a very quick look, but didn’t find anything obvious.

This is something I’m curious about in understanding how resistance is actually generated from Zwift et al. I have already noted that my speed and virtual gearing on the trainer is derived from the trainer power model (Wahoo) rather than the Zwift power model, which implies that Zwift is only controlling trainer resistance via the slope parameter. AFAIK I have no way of modifying my trainer resistance curve (eg with a lower drag coefficient) for any given slope. This has been discussed a few times on the Wahoo forums.

unterhausen

I don't think we know that for sure. Only spegler has looked at the data, and he forgot. They still could mess with resistance to simulate aero, they're doing it already somehow. It's just that they don't choose to lower resistance as much as drafting would irl. There are various things in the Fe-C data that allow a trainer to simulate real-world conditions, but also just plain old resistance.

I keep meaning to set up wireshark to look at what they are sending.

PeteHski

They can do whatever they like in the Zwift virtual avatar model. The question I have is what is actually controlling trainer resistance other than the broadcast slope parameter? It would be great if you could find out what Zwift is actually broadcasting to the trainer, other than the slope.

RChung

Zwift already does this by allowing users to change the slope effect. And, since rolling resistance scales like slope, including rolling resistance already messes with the "actual" course slope profile. I examine differences between the actual course profile and the virtual course profile all the time.

Basically, although we differentiate between sources of drag (rolling, aero, slope, acceleration, mechanical losses) really what matters to our legs is just the total drag. Similarly, although the aero drag component is separable into Cd and A, we almost always only care about their product. "Slope" may be the only thing that Zwift transmits to FE-C trainers, but if you think of total drag and the way that it changes every second in Zwift, there's no problem with piggybacking other drag onto the "slope."

ingo

The drafting feel has come up in other online discussions in Zwift forums. I disagree that you cannot feel the draft through the pedals in Zwift, I certainly can. When going from no draft to drafting another rider (or riders) pedal resistance drops whether your speed increases or not. You can see the power number in the upper left decrease as you enter the draft, cadence is determined by avatar speed and gear. So cadence does not drop when you enter the draft. If power goes down, and cadence stays the same, torque must decrease. Lower torque means less resistance at the pedals.

When slowly closing in on a group from behind, match their speed and don't pass. Speed stays the same and you are not pedalling as hard due to the draft, cadence will not change. This is easiest to feel in a robopacer group. Ride at the front of the pack and go from 2nd row to the front and pedal resistance increases. Drop back into the group and pedal resistance decreases.

The drafting algorithm in Zwift is about half the aero effect in real life, and much less when in a very large drafting line. The maximum draft effect in Zwift is 30% power reduction even in a pack of hundreds moving at 40 kph. In real life, being in the middle of a pack of 30+ riders will decrease power by more than 50% at that speed.

spelger

the FE-C protocol supports exactly this. i recall that in my app i only change the slope and leave the others at their defaults.

spelger

i think what PeteHski wants is for the grade at 10 percent to make the climb part of his kicker bike to rise to that 10% grade. that is a certain amount of resistance as well. then when the wind blows at 30mph he wants the simulated grade to not change the position of the climb but to add the resistance, or lessen it, depending on which way the wind is blowing.

PeteHski

The only problem with modifying the slope parameter to add in other sources of drag is when you are using the slope value to drive a Kickr Climb or similar bike tilting device. So if I’m riding along a flat road and moving in and out of the draft, I don’t want my Kickr Climb physically tilting my bike up and down in the process of simulating changes in aero resistance. I only want the bike to tilt according to the slope profile.

PeteHski

Exactly this thanks!

PeteHski

This sounds promising then. I just haven’t seen much evidence of Zwift making use of these additional parameters.