Why Speed Should Be the Main Metric in Zwift and not Watts over KG used in categorizing in Zwift
I believe that when it comes to categorizing cyclists, especially in a virtual setting like Zwift, we should focus on speed as the primary metric. Here’s why I think this approach makes a lot of sense:
The Benefits of a Speed-Based System
Clear and Direct Measurement:
Speed is a straightforward way to measure performance. It tells us exactly how fast a rider can go without needing to dive into complicated calculations like watts per kilogram.
Reflects Real Cycling Goals:
Ultimately, in cycling, we all want to ride as fast as we can. A system based on speed aligns perfectly with this goal and feels more intuitive.
Easier Comparisons:
Riders can easily compare their speeds with others, which adds to the competitive spirit without the confusion of extra metrics.
How It Works on Different Terrains
Climbing (Hills)
Speed on Climbs: When it comes to hills, average speed can effectively show how well a rider handles elevation changes. We could categorize riders based on their performance on specific climbs.
Setting Benchmarks: By establishing benchmarks for different climbs, we can create clear categories based on how quickly riders can ascend.
Flat Terrain
Speed on Flat Courses: On flat terrain, speed is a great indicator of a rider’s ability to maintain high velocities. We could categorize riders based on their average speed over flat segments.
Time Trials: Incorporating time trials on flat and rolling courses could help us assess and categorize riders based on their speed performance.
Challenges to Consider
While I see the advantages of a speed-based system, there are a few challenges we need to think about:
Drafting Effects:
In group rides, drafting can significantly boost speed, which might not accurately reflect an individual rider’s power.
Equipment Differences: Variations in bike setups can also impact speed, adding another layer of complexity.
Implementation Ideas
Segmented Courses :
We could create specific segments for climbing and flat courses where riders can compete for speed records. This would help establish benchmarks for categorization.
Dynamic Categorization:
A system that adjusts categories based on recent speed performances across different terrains could ensure riders are placed according to their current abilities.
Combining Metrics:
We might also consider combining speed with other metrics, like average power output during specific segments, to get a fuller picture of a rider’s performance.
Conclusion
In my view, a speed-based categorization system in Zwift could provide a clearer and more effective way to assess rider performance across various terrains. By focusing on how fast riders can go—whether climbing hills or racing on flat courses—we can create a more engaging and competitive environment.
Despite the challenges, particularly with drafting and environmental factors, I really believe that a well-implemented speed-based system could lead to a fair and exciting competitive landscape. After all, at the heart of cycling is the desire to ride faster and challenge ourselves against others!
Just replied to a thread about this last week; and I came up with this, a list of average speeds for some of the rides I am beacon on, ranging from D to B pace.
You try and decide which are the D, C, and B paced rides, and then extrapolate how or why speed should be useful metric in Zwift.
39kph
38.1kph
40kph
43.2kph
43.2kph
43.2kph
It should of course be noted that the rides were not on the same routes specifically, but, comparable enough for the sake of argument. To lock down more of this; 4 of those rides are forced bike frames of two different types, and all contained roughly groups of 30-60 riders on average throughout the event.
With this in mind, let me remind the whole world here, that some of us use miles per hour, and the difference between 38kph and 43kph is only a whole 3 mph…
So knowing that that list of 6 rides could either be D pace, C pace, or B pace, and they all fall within a 3mph range for those of us not using the metric system; hopefully you can see how quickly all of this falls apart.
Zwift does not do speeds correctly, period.
So it cannot be used as a system within its environment.
If you go down this path you’d have to use a theoretical speed rather than any measured segments, the draft effect is far too big for that.
Zwift should be able to calculate a theoretical speed for everyone based on their height, weight, and FTP or zFTP. I’d agree that doing so would make for much better categories than the FTP w/kg we’ve had.
The riders nearby list should also show speed rather than w/kg. It would immediately stop the questions of why someone is faster with a lower w/kg.
I think it will be very difficult to come up with a useable formula due to the challenges you’ve described in terms of drafting and bike choice. However one positive thing is this would be the only system that takes rider height and therefore cda into account. A 160cm rider and 190cm rider of the same weight, same watts, same w/kg, and same bike will have different speeds.
Calculating speed from biodata is the only way to do it, agreed, and would create a closer spread if all biodata is genuine, accurate and a true reflection of a rider’s capability - calculated speed at 130% ftp (approx pull power) is the principle I use for managing TTT teams.
However it would do nothing to prevent sandbagging, it’s easy to manipulate your metrics to stay at the top of a category. Results based is the only real option to reduce that aspect.
There are 3 main variables that determine individual rider speed on Zwift: Power, height and weight. Both height and weight affect CdA which is important for calculating speed.
Zwift knows your power, height and weight and they also know the formula they use to calculate speed. Zwift could calculate a theoretical speed (let’s say, based on 20 minute power) for each and every single rider over a number of different terrain types (flat, rolling, mountain).
Bike choice and drafting should be normalized when calculating a theoretical speed. Terrain will lead to different results for different riders (light riders with a high w/kg will rank higher for hills than on flat terrain for example).
This would be a more accurate way of determining potential performance than W/kg.
It would also mean Zwift never actually needs to measure the speed of riders. They already know the 3 factors that affect speed so they would only need to calculate a theoretical speed based on those factors.
The problem is you would have to explain the system to simple cyclists. W/kg may be imperfect but it’s easy to understand.
All this discussion is moot though as Zwift is in the testing phase of the new Racing Score which gets rid of W/kg.
From what I have read the seed score used in Racing Score factors in that lighter riders lose out on flat terrain. The W/kg system that Zwift has used up to know has used a blunt rule (the minimum Watts limits for each category) to account for lighter riders losing out on flat ground.
To make it easier for a layperson to understand you could convert it to a percentile, and make it a star system. Someone could be a 5 star flat rider and a 3 star climber. Racing Score will sort out categories for racing, but it could still be improved for group rides, pace partners etc.
Yeah, that’s a good idea. Categorisation by a riders theoretical speed would certainly help with group rides and pace partners. But there’s 9 pace partners at the moment and they are all ranked by w/kg so you’d need some system that was consistent across group rides and pace partners which is the advantage of W/kg (despite it’s limitations).
The easiest methodology would be to get rid of drafting; and get rid of height and weight by giving all riders the same attributes, and just do the whole platform by power. eezy peazy
When two riders are both traveling at 25 km/h, they are moving at the same speed, regardless of their power output.
One rider might be producing 300 watts while the other is producing 200 watts due to drafting or differences in body weight.
Their positions in the race are determined by their speed, not by the power they generate. In a real-world race, cyclists can observe and react to the speed of other riders, not their power output.
By not displaying the speeds of other riders, Zwift omits crucial information that would be available in a real race scenario.