I was thinking about this issue last week when I did a 3x20min workout. The workout suggested 252W for each 20 min block (Fun!), so it ended up rounding down to 250W. I will be doing it at 255W this week by bumping up the difficulty, and am pretty sure it will feel worse 
To give the developers some hint of where to look inside the code, itāll propably look something like this:
while (workout == active), do
{
calculate trainertargetwatts (myftpvalue)
if trainertargetwattslastdigit >=0 and <=3,
trainertargetwattslastdigit == 0;
else
trainertargetwattslastdigit == 5;
return trainertargetwatts;
}
while (workout == active), do
push trainertargetwatts (15); #minimum push interval 15 seconds
so to correct this:
while (workout == active), do
{
calculate trainertargetwatts (myftpvalue)
return trainertargetwatts;
}
while (workout == active), do
push trainertargetwatts (1); #push every second
2 Likes
I donāt think itās like that. They probably started storing the target watts in a uint8, which can represent only 256 values. Later they found out that some dopers could produce more than 256 watts, so they started scaling the values by 5 to accommodate for up to some 1280 monster watts. You can have resolution or range, not both (unless you buy more memory, but no one wants another price hike). There is no free lunch.
1 Like
Then why did it work prior to January 2022?
wuuuut??
Back in those days, I remember, that Zwift and e.g. Wahoo for the trainer side were blaming each other for unstable control algorithms. Maybe Zwift geniuses found āa solutionā for robust power prescription using 5W stepping in combination with other parameters - and are frightened in changing anything here. 
However, all we know, from our admins we are paying for, is that the Manhattan project was trivial, comparatively!? 
1 Like
What formatting style is that? My poor eyes! 
1 Like
Form follows function. And I wanted to comply with Zwifts own code (which propably seems to be a mess)
But jokes aside, weāre still getting absolutely zero response or any kind of communication. Is it on their agenda? Is there any kind of progress? Is it even worked on?
Itās three years since it got changed, definitly not to the better.
Btw, Happy Birthdayā¦ sigh
2 Likes
IĀ“m sick and tired of being ignored.
The problem has been proved beyond any doubt in several ways!
The problem has existed for three++ years!
The zwift game has a lot of good features and physics and mathematics seem to work good in most cases.
The 5W increments are just ridiculous and meaningless.
While a 5W increase might seem small for a user with a 500W FTP, the same increase will be huge for a user with 100W FTP.
The 500W user gets a minimum increase of 1% while the 100W user gets a minimum increase of 5%.
The Zwift programmers are surely competent at programming, but I have some doubts about their cycling and fitness understanding. It is no use telling zwift how to implement this, IĀ“m sure they know how to code the fix. It all depends on their willingness to look into it. They simply donĀ“t want to change and they donĀ“t want to explain why.
At the moment IĀ“m training 205W constant load for 60 minutes which is at my limit. I would like to test 206W, 207W etc to find out what my capacity is, but thatĀ“s impossible with todayĀ“s solution. The next step would be 210W which is too high for me.
3 Likes
While you wait for the solution just do 210 for, say, 15 minutes and then knock it back to 205 for the remaining 45. Next time, try 30/30. Then 45/15. Then 60/0.
Or ignore what I just said. Just have a bit more rest, eat loads of carbs and then do the hour at 210 because I bet you can. 
What trainer do you have whereby you can hold exactly 205 watts but not exactly 206? Iād be genuinely surprised if that degree of accuracy and consistency is even humanly possible.
1 Like
Itās still listed as a feature request, although it is clearly a bug, but even more a known issue.
Go look at your Strava file you will see that erg is not holding 205w it is jumping all over the place.
Thank you for all your good advice, David!
2 Likes
Dear Colin and Gerrie.
Thank you for the info. IĀ“m well aware of this. Read the whole thread and maybe youĀ“ll get a better understanding of the problem. The thread heading is Ā«Bring back the exact watt goals for workouts. Stop the roundingĀ». Now I just regret that I wrote the two last sentences because I was not clear enough.
Zwift will only set the resistance of the trainer in 5W increments. That is even if I use the +/- buttons or if I make a custom workout with values that donĀ“t match a 5W increment.
I never said exactly 205W, but when I make a custom workout with 205W resistance for one hour, the average will be about 205W. For values of 206W and 207W zwift is also outputting 205W because of the rounding. For values of 208W and 209W zwift is rounding up and outputting 210W. There is no way of getting a 208W output and hence an average of about 208W. This behaviour is meaningless and it does not take advantage of the trainerĀ“s accuracy.
1 Like
Hi Colin, please read my answer to Gerrie.
Fully ACK, you summarized the truely cultivated Zwift bug! 
It is annoying and frustrating that this bug still exists!
I wonder if the ignorant comment of someone even with the moderator logo in his name on jumping watts would have been given to Niki_T_NikiT (one of the initial complaints) or to Nils van der Poel (see his training schedule above)ā¦
Hi Dreg.
Thanks a lot for the ACK!
I must say that Zwiftās behavior makes me wonder. What is their motivation for staying completely silent? Are they confident that customers wonāt leave them and that they can do exactly what they want without consequences? Customers are tied to Zwift because they do not want to lose their game history. Iām now 70 years old, with a lifelong experience in software and hardware development in the tech industry, and we always tried to make products with the best possible performance concerning speed, sampling frequency, range, and resolution. We needed to stay ahead of our competitors. Zwift does not seem to worry about this. With todayās technology, our demands should be a piece of cake to implement. The resolution should, in my opinion, be even less than 1W, and the output frequency should be at least 1Hz .
Ah, cool age! Zwift not only discriminates lightweight women but also older peopleā¦
I think we can be quite sure that there is some kind of technical reason why they truely fear to change the code. However, in three years time a solution could have been implemented, with the exact same resources if a few buildings wouldnāt have been rendered or some kind of gaming features wouldnāt exist.
Concering precision: the trolls in this thread depict that the power scatters, over pedal stroke, over time, on different days. Yes, obviously, the power values depend on the condition of the sensors, on their calibration with respect to temperature, on the condition of the trainers belt drive, etc. Thus, there is a reasonable limit for the wattage prescription and the frequency. However, those 5W are crap!