Sticky draft


(A Oosthuizen (C)) #1

I have found a few threads about this topic going back to 2015 but could not find any real answer or resolution to it. What is the deal with this? Especially going uphill. Nowhere in real life have I ever been stuck behind a rider if I’m going faster on an incline so why would this even be in the drafting algorithm? 


(Mathew W.) #2

While the drafting in Zwift was tweaked a few times here and there a couple years ago, it is working as intended. The “sticky” drafting is an intentional part of the drafting mechanic and is not meant to reflect exactly how drafting works in real life. 

Zwift Insider has a great article from a couple years ago on Zwift drafting that goes over the mechanics in more depth. While the article is a couple years old, the information is still relevant to how Zwift drafting works. 


(Steve Ellis) #3

Mathew:

While the drafting in Zwift is working as intended, it is not working as desired. As the Zwift Insider article states, being in the draft you want, or not being stuck behind a slower rider,  “becomes a matter of putting down the right amount of power at the right time, plus a little bit of luck.”

Several posts on the Zwift forum here have asked for this to be fixed, so that, for instance, when you’re drafting with a group, you don’t get split off and stuck onto a slower rider the group is passing. After that happens 2, 3, 4 times in a race, hard group ride, etc., you don’t have the power left anymore to power around the slower rider to catch back up to the group you were part of. 


(David K) #4

It seems like this question would better fit in our Feature Requests forum where it could be considered by our other members and our development team.


(A Oosthuizen (C)) #5

The issue is not that I do not that I don’t understand how zwift implemented drafting but rather that after years of comments regarding the illogical way it works, getting stuck behind a rider who is producing less power that you on an _ uphill _ seems to make sense in somebody’s mind. Like the article says, on flats it’s a matter of finesse and luck and with some practise I’ve managed make it work, but the _ uphill _ thing? Really! Not even in my virtual mind does that make sense.

But I raised it since I saw the age old posts and was just wondering if I missed how this was addressed somewhere. Obviously not so I guess it is what it is then. Thanks for the responses. 


(Steve Ellis) #6

David:

Just in 2018, a couple of Feature Request forum posts have asked for the Zwift drafting to work like cycling drafting. Perhaps more drafting comments have been in in other sections of support.zwift.com because sticky drafting seems to need fixing, not featuring.


(Andre Hufschmid) #7

I’ve noticed the same thing several times while riding. If you don’t put out quite a lot more power you get stuck behind people that are going way slower than you. This is so annoying. The best thing is to ride a time trail bike, at least there drafting is not in place.

I have also noticed, people drafting behind you, when its only you and another rider get way to much benefits. I was doing 4.6w/kg a few days ago and this dude was keeping up doing only 2.5-2.8w/kg. This is just not normal.

I’ve also had the case where this dude actually closed the gap again only doing 2.5w/kg after he lost 2-3 seconds to me. Something is strange here.


(Tomas S) #8

Andre: 

The dude your talking about is probably a lot heavier than you,because of that he has a lower w/kg number,


(Matt von Wahlde 7403) #9

I’ve done a lot of Zwift races (over 150, i think) and I don’t have a huge problem with the drafting implementation, but I do agree its not much like real life. 

I wish they would publish the pseudocode (english description of what the code does), for the drafting so that people could understand it, accept it or not. I think the black box nature of it annoys/offends some people.

I assume that there a lot of good reasons why they can’t improve it.  They clearly have talented developers and I just don’t believe they intentionally cripple it. I just wish Zwift could be more transparent about it, or at least establish a roadmap…


(Gerrie Delport) #10

Matt I tend to agree with you, I also don’t have a issue with it and got used to it by now. 

 

What i told my self is that to have a perfect draft you need to know exactly where every rider is at all times, but with small delays over the internet they have to have some kind of buffer to account for that. This make sense for the sticky draft, so you have to make a noticeable change/effort before Zwift will know you are trying to overtake. 

 

 


(A Oosthuizen (C)) #11

Keep in mind that the issue is not with drafting in general. Awkward as it may be it works if you understand the whole ‘finesse’ and ‘luck’ concepts. I agree that to simulate the perfect draft is not really feasible. The real issue is with the way the sticky bubble has been implemented. For me particularly going uphill and for others when overhauling a slower group. If we take just take the uphill draft for instance. Pseudo code would be as simple as ‘If gradient >x% then drafting bubble = false’.


(Matt von Wahlde 7403) #12

I think a “draft meter” on the HUD would be a great assistance to see visually what everyone’s draft benefit was. I never noticed much draft effect on uphills, I assumed that it did shut off after speed decreased below a certain point (like real life, of course, I have had strong downdrafts down mountain rides where drafting made a big difference, but that’s another story…)


(David K) #13

These are all awesome ideas, and I agree our draft system definitely could use improvement, but it’s not entirely likely our developers will see this post when considering new features to add to Zwift since they’re in our feedback section and not the feature request section.

I am aware of a couple of active related feature requests that have already been submitted internally, so I’ve added a few more votes to each of them. If you guy do want to see drafting improved, adding posts to existing feature requests would help.


(Danny Boyd) #14

Andre Hufschmid, how do you know the other guy isnt a 100+kg rider…